Volume 43 No. 1 June 2011

1

SABRAO Journal of Breeding and Genetics 43 (1) 1-14, 2011

SELECTION FOR LOW ΔΔ9-TETRAHYDROCANNABINOL

CONTENT IN THAI HEMP CULTIVARS

W. KUNKAEW1*, S. JULSRIGIVAL1, P. TIPPARAT2 and S. PINMANEE3

SUMMARY

Selection for reduced Δ9-tetrahydrocannabinol (THC) content in four Thai hemp cultivars (including V50, Mae Sa Mai, Huay Hoi and Pang Ung) was carried out in highland areas in the northern Thailand. Research work was conducted for two consecutive growing seasons during 2008 to 2009 at Pangda Royal Agricultural Station, Samoeng district, Chiang Mai province, Thailand. Results of selection indicated that after selecting for two successive generations, the average THC content of four Thai hemp cultivars reduced to 18.0-55.0% and cannabidiol (CBD) content increased to 20.0-127.0%. The results of selection also indicated that chemotype classification could be grouped by using the ratio of CBD/THC content as follows: non-drug type (CBD/THC>10.0), intermediate type (1.0≤CBD/THC≤10.0) and drug type (CBD/THC<1.0). Thus, selection for reduced THC content, high ratio of CBD/THC content could be used as alternative criteria for improving low THC content in hemp cultivars. As well, mass selection method was considered as an effective and suitable method for improving these THC and CBD traits.

Key words: chemotype, hemp, THC, CBD

INTRODUCTION Hemp (Cannabis sativa L.) has been a source of fiber crops for a long time. Hemp has been cultivated by Hmong, the hill tribe people who live in the highland areas in northern Thailand. In Thailand, hemp is classified as a narcotic crop which farmers are prohibited to grow. However, this crop is still planted in some outreach areas where it cannot be noticed by government officers. Her Majesty the Queen is very much interested in hemp crop cultivation for fiber products and kindly recommended to the relavant government sectors to promote hemp cultivation in the highland areas since 2004. 1 Royal Project Foundation, 65 Suthep Road, Chiang Mai 50200, Thailand.

2 Regional Medical Sciences Center, 191 M. 8, T. Don Kaew, A. Mae Rim, Chiang Mai 50180, Thailand.

3 Highland Research and Development Institute (Public Organization), 65 Suthep Road, Chiang Mai 50200, Thailand.

* Corresponding author: [email protected]

2

Research work on cultivation and development of hemp in Thailand were first studied and reported by Queen Sirikit Botanic Garden in Chiang Mai, Thailand. Later, the Economic and Social Institute of Thailand granted financial funds to the Royal Project Foundation and to the Highland Research and Development Institute (Public Organization) to conduct research works on hemp production and varietal improvement for low Δ9-tetrahydrocannabinol (THC) content.

In some European countries, monoecious fiber hemp is considered as an economic crop which can be grown legally by farmers if THC content is lower than 0.2 percent (Mechtler et al., 2004; Callaway, 2008). Presently, it is found that demand of hemp products which is produced widely for industrial uses such as fibers, seeds, oils, and medicines are increasing (Ranalli, 1999; Johnson, 2010). Thus, hemp is widely and commonly planted for fiber and/or seed production in many countries in Europe and some countries in Asia where climatic and soil factors are suitable for planting.

In Australia, studies on adaptation of hemp varieties indicated that varieties introduced from sub-tropical climates were able to adapt in Australia better than varieties brought from temperate areas. In addition, analysis of THC content of these introduced varieties showed evidently higher values than their places of introduction. Some varieties gave THC content as high as 1.0%. Hot weather which is a climatic factor to inhibit plant growth and development as well as increasing of THC content was reported by Jobling and Warner (2001). Furthermore, it is also found that local hemp varieties possess THC content more than 0.5% which is prohibited to be grown by farmers. Therefore a varietal improvement program for decreasing THC content in hemp was first initiated in Thailand by Kunkaew et al. (2009, 2010). Their studies indicated that THC content was different among varieties and varied quantitatively within populations of each variety. The variation of THC content could be classified into three chemotype groups, including drug, intermediate and non-drug type groups. Kaveeta et al. (2006) and Sengloung et al. (2009) studied the morphology, growth and development of hemp crop growing on the highland areas in Thailand.

Chemical components in hemp plant were studied and reported by many researchers. An important chemotype of hemp is cannabinoid which is one of the terpenophenolic compounds and is a unique character to cannabis plants. These terpenophenolic compounds are produced by glandular trichromes that occur on most aerial surfaces of the plants. Two important compounds of cannabinoid are Δ9-tetrahydrocannabinol content or THC and cannabidiol or CBD. Since THC is classified as one kind of narcotic substances, thus, the amount of THC content in plant is an important data for dividing hemp plant into drug and non-drug type. Thus, information on THC and CBD content in hemp are very useful in guiding low THC content improvement in hemp variety (Hillig and Mahlberg, 2004). Genetic control for THC and CBD content in hemp have been previously studied and reported. De Meijer et al. (2003) and Mandolino et al. (2003) found that THC and CBD traits were controlled by a single locus gene (B) with two co-dominant alleles (BT and BD). Therefore, pure CBD plant or fiber type has a BD/BD, while pure THC plant or drug type has a BT/BT genotype at B locus and intermediate type has the heterozygous of BD/BT genotype which two different alleles BD and BT having as co-dominant control. The ratio of CBD/THC indicates the qualitative character of these

3

chemotypes. It was found that hybrid F1 derived from crossing between pure CBD and THC plants would give all heterozygous (CBD/THC) plants. The segregation of the CBD, CBD/THC and THC chemotypes was observed in accordance with a 1:2:1 ratio, which suggests a single Mendelian locus determining chemotype, with two co-dominant alleles, one for CBD and another for THC. It was further investigated that BD/BD genotype had a lethal effect for seed fertility and viability. BT/BT genotype had more vigor for these two traits than BD/BD genotype. These inferior problems of BD/BD genotype may probably be due to semi-lethal effect which is caused by co-dominant pairing of two alleles (de Meijer et al., 2003).

Hennink (1994) studied the inheritance of THC and CBD traits in hemp and reported that narrow-sense heritability (h2

n) of THC and CBD traits was rather low, estimated about 15.0% and 8.0%, respectively, in contrast with results of Kunkaew et al. (2010) who reported that heritability of THC trait of local hemp cultivars estimated under highland growing condition in Thailand was about 71-87%. Population improvement in crop plant could be made by simple mass selection method, especially improving of qualitative traits which are controlled by a few genes (Allard, 1960). De Meijer and van Soest (1992) and Hennink (1994) were successful in developing low THC content varieties in hemp crops by using mass selection method.

The objectives of this research work are to develop low THC content in local hemp varieties for growing as an economic fiber and oil seed crop on the highland areas in the northern part of Thailand.

MATERIALS AND METHODS

In 2008 growing season (during June – December), original or first generation population (M0) of four local hemp cultivars collected in Thailand which included V50, Mae Sa Mai, Huay Hoi and Pang Ung were prepared for THC and CBD identification. The experiment was conducted in plastic shelter growing condition at Pangda Royal Agricultural station (elevation 720 m above mean sea level), Samoeng District, Chiang Mai Province, Thailand. The average temperature at this station was 18.8 – 29.6 °C (mean was 23.4 °C), average air humidity was 53.6 – 95% (mean was 74.3%) and rainfall was 1,075.5 mm. About 80 days and 95-120 days after sowing which were early stages of flowering of male and female plants, respectively, 50 male plants and 100 female plants were randomly sampled. Leaves on the top parts of stem (about 1/3 of stem) of both male and female plants were collected separately for THC and CBD content analysis. Gas chromatography method was used for analyzing these two chemotypes content (United Nations, 1987). The analysis was conducted at the laboratory of Regional Medical Sciences Center in Chiang Mai Province. After the analysis results were obtained, THC content data of both male and female plants of each cultivar were examined for plant which its THC content was lower than 0.3% and then individual plant was selected. These low THC content plants of both sexes were still kept in the plots for producing advanced generation hybrid seeds while the undesirable plants were rogued. At maturity, female plants were harvested and seeds were bulked together in order to develop first selection cycle (M1) generation.

4

In 2009 growing season, seeds of M1 generation were planted and collection of leaves in both male and female plants of each hemp cultivar, THC and CBD analysis were carried out the same as M0 generation.

THC and CBD data of each hemp cultivar of both M0 and M1 generations were classified into three chemotype groups which included (1) non-drug type (THC<0.3%), (2) intermediate type (THC>0.3% and CBD>0.5%) and drug type (THC>0.3% and CBD<0.5%) which was proposed by de Meijer et al. (1992). Relationship between THC and CBD traits for distinguishing among the chemotype groups was examined by simple regression method (Steel and Torrie, 1960).

RESULTS

Analysis of chemotype contents and classifications of 4 hemp cultivars are presented in Tables 1-4 and Figures 1-4. Results of each cultivar are described as follow:

1. V50 cultivar.

A. Chemotype composition:

(a) THC content. Analysis of chemotype content of this hemp cultivar showed that average THC content obtained from male and female plants of original population (M0-generation) were 0.62% and 0.40%, respectively, average of both sexes was 0.47%. After one generation of selection (M1-generation), average THC content of male and female plants decreased to 0.24% and 0.20%, respectively. Average of both sexes decreased to 0.21% or decreased about 55.0% from M0 generation.

(b) CBD content. It was found that average CBD content of M0 generation for male and female plants were 1.01% and 0.72%, respectively, average of both sexes was 0.82%. In M1-generation, CBD content of male plants reduced to 0.85% but female plants increased to 1.04%, averaged of both sexes increased to 0.98% or increased from M0 generation about 20.0%.

B. Chemotype classification:

(a) Non-drug type. It was found that non-drug type of M0 generation consisted of 42.0% and 48.0% of male and female plants, respectively, average of both sexes was 45.0%. In M1 generation, average of male and female plants increased to 60.5% and 74.7%, respectively, average of both sexes increased to 67.5% or increased about 50.0% from M0 generation. For association analysis between THC and CBD plant of non-drug type, it indicated that there was high positive significant relationship between THC and CBD content of this chemotype group with R2=0.86**.

5

(b) Intermediate type. It was found that M0 generation consisted of 26.0% of male and 32.0% of female plants, averaged 29.0% of both sexes. In M1 generation, average of male and female plants increased to 34.9% and 23.2%, respectively, average of both sexes increased slightly to 29.1%. There was high positive significant relationship between THC and CBD content of this chemotype group with R2=0.91**.

Figure 1. THC and CBD contents of individual plant belongs to M0 and M1 of V50

hemp cultivar and relationship between THC and CBD of each chemotype group. The experiment was conducted at Pangda Agricultural Station in 2008 and 2009 growing seasons.

CBD/THC=19.21 (R2=0.86**)

CBD/THC=2.18 (R2=0.91**) CBD/THC=0.22 (R2=0.83**)

M0-female

M1-male

M0-male M1-female

6

Table 1. THC content, CBD content and chemotype group of M0 and M1 of V50 hemp cultivar.

Chemotype M0 M1 content/group Male Female Average Male Female Average* A) Chemotype content

a) THC content (%) b) CBD content (%)

0.62 1.01

0.40 0.72

0.47 0.82

0.24 0.85

0.20 1.04

0.21 (-55%) 0.98 (+20%)

B) Chemotype group a) Non-drug type (%) b) Intermediate type (%) c) Drug type (%)

42.0 26.0 32.0

48.0 32.0 20.0

45.0 29.0 26.0

60.5 34.9 4.6

74.7 23.2 2.1

67.5 (+50%) 29.1 (+0%) 3.4 (-87%)

* Values in brackets are percentages of increasing or decreasing chemotype content and chemotype groups. (c) Drug type. It was found that M0 generation consisted of 32.0% and 20.0% of male and female plants, respectively, averaged 26.0% of both sexes. In M1 generation, average of male and female plants reduced to 4.6% and 2.1%, respectively, average of both sexes decreased slightly to 3.4% or decreased about 87.0% from M0 generation. There was a high positive significant relationship between THC and CBD content for this chemotype group with R2=0.83**.

2. Mae Sa Mai cultivar.


(a) THC content. The average THC contents of male and female plants of M0 generation were 0.79% and 0.62%, respectively, and the average of both sexes was 0.68%. In M1 generation, average THC content of male and female plants decreased to 0.45% and 0.38%, respectively, average of both sexes decreased to 0.40% or decreased about 41.0% from M0 generation.

(b) CBD content. It was found that average content of CBD of male and female plants of M0 generation were 0.41% and 0.34%, respectively, averaged 0.36% of both sexes. In M1 generation, CBD content of male and female plants increased to 0.42% and 0.73%, respectively, average of both sexes increased to 0.63% or increased about 75.0% from M0 generation.


(a) Non-drug type. Results indicated that M0 generation consisted of 14.0% of male and 14.0% of female plants, averaged 14.0% of both sexes. In M1 generation, average of male and female plants increased to 45.8% and 42.0%, respectively, average of both sexes increased to 43.9% or about 214.0% increased from M0 generation. There was high positive significant relationship between THC and CBD content for this chemotype group with R2=0.91**.

7

(b) Intermediate type. Results indicated that M0 generation consisted of 26.0% of male and 15.0% of female plants, averaged 20.5% of both sexes. In M1 generation, male plants decreased to 20.8% but female plants increased to 37.0%, average of both sexes increased to 28.9% or about 41.0% increased from M0 generation. There was high positive significant relationship between THC and CBD content in this chemotype as well with R2=0.90**.

(c) Drug type. It was found that M0 generation consisted of 60.0% of male and 71.0% of female plants, averaged 65.5% of both sexes. In M1 generation, male and female plants decreased to 33.4% and 21.0%, respectively, average of both sexes decreased to 27.2% or about 58.0% decreased from M0 generation. There was high positive significant relationship between THC and CBD content of this chemotype as well with R2=0.72**.

Figure 2. THC and CBD contents of individual plant belongs to M0 and M1 of Mae

Sa Mai hemp cultivar and relationship between THC and CBD of each chemotype group. The experiment was conducted at Pangda Agricultural Station in 2008 and 2009 growing seasons.

CBD/THC=0.29 (R2=0.72**) CBD/THC=1.92 (R2=0.90**)

CBD/THC=14.38 (R2=0.91**)

M0-female

M1-male

M0-male M1-female

8

Table 2. THC content, CBD content and chemotype group of M0 and M1 of Mae Sa Mai hemp cultivar.



0.79 0.41

0.62 0.34

0.68 0.36

0.45 0.42

0.38 0.73

0.40 (-41%) 0.63 (+75%)


14.0 26.0 60.0

14.0 15.0 71.0

14.0 20.5 65.5

45.8 20.8 33.4

42.0 37.0 21.0

43.9 (+214%) 28.9 (+41%) 27.2 (-58%)

* Values in brackets are percentages of increasing or decreasing chemotype content and chemotype groups. 3. Huay Hoi cultivar.


(a) THC content. Results indicated that average THC content of male and female plants of M0 generation were 0.54% and 0.47%, respectively, averaged 0.49% of both sexes. In M1 generation, average THC content of male and female plants decreased to 0.45% and 0.37%, respectively, average of both sexes decreased to 0.40% or about 18.0% decreased from M0 generation.

(b) CBD content. It was found that average CBD content of M0 generation of male and female plants were 0.45% and 0.34%, respectively, average of both sexes was 0.38%. In M1 generation, CBD content of male plants decreased to 0.43% but female plants increased to 0.60%, average value of both sexes increased to 0.55% or increased about 45.0% from M0 generation.


(a) Non-drug type. Results indicated that average non-drug type of M0 generation consisted of 32.0% of male and 38.0% of female plants, averaged 35.0% of both sexes. In M1 generation, average of male and female plants increased to 43.6% and 50.0%, respectively, average of both sexes increased to 46.8% or increased about 34.0% from M0 generation. There was high positive significant relationship between THC and CBD content in this chemotype group with R2=0.84**.

(b) Intermediate type. Results indicated that average intermediate type of M0 generation consisted of 24.0% of male plants and 6.0% of female plants; average of both sexes was 15.0%. In M1 generation, average of male and female plants increased to 30.8% and 21.1%, respectively, average of both sexes increased slightly to 26.0% or about 73.0% increased from M0 generation. There was a high positive significant relationship between THC and CBD content of this chemotype group as well with R2=0.92**.

9

(c) Drug type. It was found that drug type of M0 generation consisted of 44.0% of male plants and 56.0% of female plants, averaged 50.0% of both sexes. In M1 generation, average of male and female plants decreased to 25.6% and 28.9%, respectively, average of both sexes decreased to 27.2% or about 46.0% decreased from M0 generation. There was high positive significant relationship between THC and CBD content for this chemotype group with R2=0.77**.

Figure 3. THC and CBD contents of individual plant belongs to M0 and M1 of

Huay Hoi hemp cultivar and relationship between THC and CBD of each chemotype group. The experiment was conducted at Pangda Agricultural Station in 2008 and 2009 growing seasons.

CBD/THC=0.25 (R2=0.77**) CBD/THC=2.00 (R2=0.92**)

CBD/THC=15.05 (R2=0.84**)

M0-female

M1-male

M0-male M1-female

10

Table 3. THC content, CBD content and chemotype group of M0 and M1 of Huay Hoi hemp cultivar.



0.54 0.45

0.47 0.34

0.49 0.38

0.45 0.43

0.37 0.60

0.40 (-18%) 0.55 (+45%)


32.0 24.0 44.0

38.0 6.0 56.0

35.0 15.0 50.0

43.6 30.8 25.6

50.0 21.1 28.9

46.8 (+34%) 26.0 (+73%) 27.2 (-46%)

* Values in brackets are percentages of increasing or decreasing chemotype content and chemotype groups. 4. Pang Ung cultivar.


(a) THC content. The results indicated that average THC content of male and female plants of M0 generation was 0.70% and 0.71%, respectively, averaged 0.71% of both sexes. In M1 generation, average THC content of male and female plants decreased to 0.65% and 0.51%, respectively, average of both sexes decreased to 0.55% or decreased about 23.0% from M0 generation.

(b) CBD content. It was found that average CBD content of M0 generation of male and female plants were 0.16% and 0.15%, respectively, average of both sexes was 0.15%. In M1 generation, average CBD content of male and female plants increased to 0.35% and 0.34%, respectively, average value of both sexes increased to 0.34% or increased about 127.0% from M0 generation.


(a) Non-drug type. Results indicated that average non-drug type of M0 generation consisted of 18.0% of male plants and 18.0% of female plants, averaged 18.0% of both sexes. In M1 generation, average of male and female of non-drug type increased to 25.5% and 26.5%, respectively, average of both sexes increased to 26.0% or increased about 44.0% from M0 generation. For this chemotype group, association between THC and CBD content could not be identified.

(b) Intermediate type. Results indicated that average intermediate type of M0 generation consisted of 4.0% of male plants and 1.0% of female plants, averaged 2.5% of both sexes. In M1 generation, average of male and female plants increased to 29.8% and 18.4%, respectively. Average of both sexes increased to 24.1% or increased about 864.0% from M0 generation. There was high positive significant relationship between THC and CBD content of this chemotype group with R2=0.88**.

11

(c) Drug type. It was found that drug type of M0 generation consisted of 78.0% of male plants and 81.0% of female plants, averaged 79.5% of both sexes. In M1 generation, average of male and female plants decreased to 44.7% and 55.1%, respectively, average of both sexes decreased to 49.9% or decreased about 37.0% from M0 generation. There was high positive significant relationship between THC and CBD content of this chemotype group as well with R2=0.79**.

Figure 4. THC and CBD contents of individual plant belongs to M0 and M1 of

Pang Ung hemp cultivar and relationship between THC and CBD of each chemotype group. The experiment was conducted at Pangda Agricultural Station in 2008 and 2009 growing seasons.

Table 4. THC content, CBD content and chemotype group of M0 and M1 of Pang

Ung hemp cultivar. Chemotype M0 M1 content/group Male Female Average Male Female Average* A) Chemotype content


0.70 0.16

0.71 0.15

0.71 0.15

0.65 0.35

0.51 0.34

0.55 (-23%) 0.34 (+127%)


18.0 4.0

78.0

18.0 1.0 81.0

18.0 2.5 79.5

25.5 29.8 44.7

26.5 18.4 55.1

26.0 (+44%)

24.1 (+864%) 49.9 (-37%)

* Values in brackets are percentages of increasing or decreasing chemotype content and chemotype groups.

CBD/THC=0.19 (R2=0.79**) CBD/THC=1.59 (R2=0.88**)

M0-female

M1-male

M0-male M1-female

12

DISCUSSION

Improvement of hemp crop for reduced THC content by using mass selection method was carried out during 2008-2009 cropping seasons. After one cycle of selection in both male and female plants, results indicated that local hemp cultivars which were grown on the highland areas in Thailand had a wide range of variation in THC and CBD contents. The variation in content of these two chemotypes could be classified clearly into three chemotype groups which are: (1) drug type (CBD/THC <1.0); (2) intermediate type (1.0 ≤CBD/THC ≤10.0); and (3) non-drug type (CBD/THC >10.0). These results were similar to those reported by de Meijer et al. (1992, 2003), Hillig and Mahlberg (2004), Mandolino and Carboni (2004) and Mechtler et al. (2004).

Reduction of THC and increasing CBD content of each hemp varietal population was also examined after one generation of selection. These results were obtained since the number of plants belonging to non-drug and intermediate types were increased and vice versa for drug type plants. In addition, using of mass selection method is an effective and appropriate means for selecting qualitative traits which are controlled by few genes (Allard, 1960). Progress in selection for low THC content in local hemp cultivars is mainly due to genetic factors which THC and CBD traits are controlled by single locus of gene with two co-dominant alleles (de Meijer et al., 2003; Mandolino et al., 2003). As well, heritability of THC content is rather high (Kunkaew et al., 2010).

This study evidently revealed that there were high positive significant relationships between THC and CBD content of three chemotype groups (drug, intermediate and non-drug) of four hemp cultivars with R2 ranging from 72.0-92.0%; these similar results were reported by Hennink (1994). Thus, selection for reduced THC content, high ratio of CBD/THC content could be used as an alternative criterion for improving low THC content in hemp cultivars. It is anticipated that low THC content hemp cultivars with good agronomic characters will be obtained from this research project and are suitable for agricultural production by farmers who live on the tropical highland areas in the northern part of Thailand.

ACKNOWLEDGEMENTS

The authors would like to thank Royal Project Foundation and Highland Research and Development Institute (Public Organization) for supporting to conduct this experiment.

13

REFERENCES

Allard, R.W. 1960. Principles of plant breeding. John Wiley & Sons, Inc. London, 485 p. Amaducci, S., M. Colauzzi, A. Zatta and G. Venturi. 2008. Flowering dynamics in

monoecious and dioecious hemp genotypes. J. Ind. Hemp 13: 5-19. Callaway, J.C. 2008. A more reliable evaluation of hemp THC levels is necessary

and possible. J. Ind. Hemp 13: 117-144. De Meijer, E.P.M. and L.J.M. van Soest. 1992. The CPRO Cannabis germplasm

collection. Euphytica 62: 201-211. De Meijer, E.P.M., H.J. van der Kamp and F.A. van Eeuwijk. 1992. Characterisation

of Cannabis accessions with regard to cannabinoid content in relation to other plant characters. Euphytica 62: 187-200.

De Meijer, E.P.M., M. Bagatta, A. Carboni, P. Crucitti, V.M.C. Moliterni, P. Ranalli and G. Mandolino. 2003. The inheritance of chemical phenotype in Cannabis sativa L. Genetics 163: 335-346.

Hennink, S. 1994. Optimisation of breeding for agronomic traits in fiber hemp (Cannabis sativa L.) by study of parent-offspring relationships. Euphytica 78: 69-76.

Hillig, K.W. and P.G. Mahlberg. 2004. A chemotaxonomic analysis of cannabinoid variation in Cannabis (Cannabaceae). American Journal of Botany 91: 966-975.

Jobling, T. and P. Warner. 2001. New tropical industrial hemp. Proceeding of the 10th Australian Agronomy Conference 2001, Australian Society of Agronomy. http://www.regional.org.au/au/asa/2001/5/c/jobling.htm#TopOfPage

Johnson, R. 2010. Hemp as an agricultural commodity. Congressional Research Service. Report for Congress (Code No. RL32725). 22 p.

Kaveeta, L., Y. Paopun, K. Promratrak, M. Nanakorn, S. Suwanwong, R. Sengloung, W. Nanakorn and S. Thativiwat. 2006. Morphological and anatomical characteristics of hemp (Cannabis sativa L) in Thailand. (in Thai, with English abstract) Agricultural Science Journal 37(4): 293-302.

Kunkaew, W., S. Julsrigival, P. Tipparat, A. Punyalue, V. Punsupa, M. Srihawong and P. Auntin. 2009. Classification of hemp (Cannabis indica L.) in Thailand by THC and CBD content. (in Thai, with English abstract) Agricultural Science Journal 40: 307-314.

Kunkaew, W., S. Julsrigival, P. Tipparat, V. Punsupa and P. Auntin. 2010.Heritability of delta-9-tetrahydrocannabinol contents in hemp (Cannabis indica L.) grown on highland area in Thailand. (in Thai, with English abstract) Agricultural Science Journal 41: 75-80.

Mandolino, G. and A. Carboni. 2004. Potential of marker-assisted selection in hemp genetic improvement. Euphytica 140: 107-120.

Mandolino, G., M. Bagatta, A. Carboni, P. Ranalli and E. de Meijer. 2003. Qualitative and quantitative aspects of the inheritance of chemical phenotype in Cannabis. J. Ind. Hemp 8: 51-72.

Mechtler, K., J. Bailer and K. de Hueber. 2004. Variation of Δ9-THC content in single plants of hemp varieties. Industrial Crops and Products 19: 19-24.

Ranalli, P. 1999. Advances in hemp research. Food Products Press (Haworth Press), London, 272 p.

14

Sengloung, T., L. Kaveeta and W. Nanakorn. 2009. Effect of sowing date on growth and development of Thai hemp (Cannabis sativa L.). Kasetsart J. (Nat. Sci.) 43: 423-431.

Steel, R.G.D. and G.H. Torrie. 1960. Principles and procedures of statistics. Mc. Graw Hill Book Comp. Inc. New York, 481 p.

United Nations. 1987. Reccommended methods for testing cannabis. New York.

15

SABRAO Journal of Breeding and Genetics 43 (1) 15-26, 2011

CHARACTERIZATION OF AROMATIC RICE CULTIVARS FROM IRAN AND SURROUNDING REGIONS FOR

AROMA AND AGRONOMIC TRAITS

MEHRAN VAZIRZANJANI1 , WAKIL AHMAD SARHADI1, JIN JIN NWE1, MOZHGAN KHALAJ AMIRHOSSEINI2, ROEURN SIRANET2, NGUYEN

QUOC TRUNG2, SHINYA KAWAI3, YUTAKA HIRATA1

SUMMARY

Many aromatic rice cultivars are cultivated in Asian countries. The genetic diversity of modern rice cultivars has been reduced due to intensive breeding efforts and more diverse germplasm would enhance the selection efficiency of desirable varieties in the rice breeding programs. In this study, rice cultivars from Iran, Afghanistan, Uzbekistan, check cultivars (Nipponbare, Koshihikari, Jasmine 85, Basmati 370), and cross combinations (Jasmine 85 x Nipponbare, LTH x Pashadi Konar) were compared for aroma, morphological and agronomic characters. In the present research 1.7% KOH sensory test and PCR were used for aroma analysis. We evaluated leaf number per plant, tiller number per plant, and plant height. Iranian cultivars exhibited more tillers and more leaves compare to Uzbek, Afghan, and check cultivars. These are the desirable characters in rice plant for breeding program. Analysis of aroma performed by using 1.7 % KOH sensory test and PCR analysis. Fajr and Dorfak from Iran were comparable to Basmati 370 and Jasmine 85. Genetic analysis for aroma in the F2 generation from the cross between (Jasmine 85 x Nipponbare) exhibited a segregation ratio of 3 non-aromatic: 1 aromatic. Segregation in the F3 generation was also evaluated and was consistent with the segregation of the F2 generation. The F2 generation derived from the cross between LTH from China, and Pashadi Konar from Afghanistan exhibited a segregation ratio of 3:1 non-aromatic, and aromatic, respectively. These results indicate that aroma is controlled by a single recessive gene.

Key words: Aromatic rice, 2- acetyl- 1-pyroline, Agronomical traits, Molecular analysis 1 United Graduate School of Agricultural Science, Tokyo University of Agriculture and Technology, 3-5-8 Saiwai-cho, Fuchu, Tokyo 183-8509, Japan 2 Laboratory of Plant Genetics and Biotechnology, Graduate School of Agriculture, Tokyo University of Agriculture and Technology, 3-5-8 Saiwai-cho, Fuchu, Tokyo 183-8509, Japan * Corresponding author: [email protected]

16

INTRODUCTION Rice is the second major crop after wheat in Iran and has important role in Iranian food habits. In the past rice was considered as an expensive food for poor people and mostly used for celebrations, but today with changing the food culture, rice plays an important role in food habit of Iran. Aromatic long grain rice, represented by Jasmine and Basmati has a heady, perfumed fragrance and flavor that is suitable for Thai and Indian cuisine. Its flavor has been thought to develop fully after a post-harvest year on the shelf. Most popular aromatic rice can be relied to cook up dry, light and fluffy like other long grain. In past decades, hybrid varieties couldn't be sold in the market due to their undesirable quality. On the other hand local Iranian varieties which had good quality became rare; and it showed that quality of rice should be attended beside high yield, and the prerequisite was considered to be the identification of compounds and other factors which effect rice quality. Local rice varieties could be sold 2 to 2.5 times more than hybrid varieties, therefore definition of quality specification of rice was necessary for rice producers, customers, and most importantly for rice breeders. In Iran, the most important factor of using local landraces for breeding is quality.

More than 100 volatile compounds have been identified in rice (Buttery et al., 1988; Mahindru et al., 1995; Buttery et al., 1999; Grimm et al., 2001.). Among these compounds, (Buttery et al., 2005) have identified 2-acetyl-1-pyrroline (2-AP) as the principal aroma compound. Later many studies have confirmed the presence of this compound in all the scented rice varieties. Molecular markers such as simple sequence repeat (SSR) and single nucleotide polymorphism (SNP) markers that are genetically linked to aroma have been developed for the quick selection of aromatic rice (Lang and Buu, 2000; Coderio et al., 2002; Jin et al., 2003). However, these markers are not tightly linked with aroma gene and therefore are not reliable (Garland et al., 2000; Hien et al., 2005; Sarhadi et al., 2009). More recently, in traditional Basmati and Jasmine-like rice, (Bradbury et al. 2005) researchers further restricted the aroma region and identified a single recessive gene for aroma; which encodes betaine aldehyde dehydrogenase BADH2. The deletion observed in exon 7 of this (BADH2) gene generates a premature stop codon and presumably results in loss of activity. It was hypothesized that loss of BADH2 activity causes 2-AP accumulation (Bradbury et al. 2005). Thus the purpose of this study was to characterize morphological, agronomic and aromatic traits of local aromatic rice germplasm from Iran and surrounding regions and genetic analysis of aroma in Iranian aromatic rice landraces.

MATERIAL AND METHODS

Plant material This study included sixteen Iranian cultivars, including Fajr, Shafagh, Pouya, Shiroudi, Tabesh, Nemat, Neda (Mazandaran province), Kadus, Saleh, Dorfak, Sepidroud, Khazar (Gilan province), Zayandehroud, Sazandegi (Isfahan province), Doroudzan, and Qasredashti (Fars province), in addition to four landraces from surrounding regions viz., Lawangi, and Pashadi Konar from Afghanistan, Gulnar,

17

and Shortanby from Uzbekistan, and Nipponbare and Koshihikari from Japan, Basmati 370 from India and Jasmine 85 from Thailand were included as reference genotypes (Table 2). The Iranian cultivars were obtained from Rice Research Institute Iran (RRII). For genetic analysis, F2 individuals derived from the cross between LTH (non-aromatic), and Pashadi Konar (aromatic) were used. A population of F2 and F3 individuals derived from the cross between Jasmine 85 (aromatic), and Nipponbare (non-aromatic) were also used for genetic analysis.

Agronomical characterization The rice plants were grown in the Honmachi farm of Tokyo University of

Agriculture and Technology in spring and summer 2007 and 2008. The seeds were sown in plastic trays by using soil and fungicide with assured water supply. The field experiment was performed in randomized block design with three replications. Each replication consisted of twenty two plots, and sixteen plants were grown in the plots. One seedling per hill (20 x 25 cm) at three weeks after sowing was randomly transplanted. The plant height was measured by measuring from ground level to extended panicle on the main stem (SES, IRRI 1996). Yield components including panicles number per plant, grains number per panicle, 1000-grain weight, panicle weight, were also measured. For 1000-grain weight, an electronic scale was used to weigh the random samples of 1000 grains.

Sensory test For the sensory test, 100 mg of the young leaves from each cultivar was

weighed at heading stage and cut to small pieces, then put into Petri dishes. 10 ml of 1.7% KOH solution (Sood and Siddiq 1978) was also added into Petri dishes and left for an hour at room temperature. The samples were evaluated by five analysts and data were recorded. The results were scored from 0 to 0.5 as non-aromatic (-), and 0.5 to 1 as aromatic (+).

Evaluation of aroma by using molecular marker Total DNA was extracted from young leaves by the rapid DNA extraction

method using cetyl -trimethyl ammonium bromide (CTAB) (Doyle and Doyle, 1987). PCR (ASTEC, Gene Amp PC system 320, Japan) was performed in 25 μl reactions, including 15.4 μl of double distilled water (DDW), 1 unit of Taq DNA Polymerase, 20 ng of genomic DNA, 2.5 μl of 10X buffer, 2 μl of 50 mM MgCl2, 2 μl of 5mM DNTPs, 10 pM of each primer, external sense primer (ESP) 5′-TTGTTTGGAGCTTGCTGATG-3′, internal fragrant antisense primer (IFAP) 5′-CATAGGAGCAGCTGAAATATATACC-3′, internal non-fragrant sense primer (INSP) 5′-CTGGTAAAAAGATTATGGCTTCA-3′ and external antisense primer (EAP) 5′-AGTGCTTTACAAAGTCCCGC-3′ (Bradbury et al. 2005). The PCR conditions were as follows: denaturation at 94oC for 2 min followed by 30 cycles at 94oC for 30 s, 58oC for 30 s, 72oC for 30 s, and a final elongation step at 72oC for 5 min. PCR products were separated by electrophoresis on a 1.5 % agarose gel. One kb MassRullerTM (Fermentas Inc®) was used to estimate the fragment size. The gel was stained in ethidium bromide for 20 min, and then photographed under UV light.

18

RESULTS

Agronomical characterization Understanding relationships among rice (Oryza sativa L.) in yield, yield components (panicle density, filled grains per panicle, unfilled grains per panicle, and grain weight) is critical for breeders to utilize these relationships effectively. In this study, yield components, including panicles number per plant, panicle weight, panicle length, seeds number per panicle, and 1000-grain weight were characterized. Among all cultivars, Pashadi Konar from Afghanistan expressed the longest panicle and heavier grain weight. However the ability to produce panicles in Iranian cultivars was higher than the others. The number of panicles per plant ranged from 9 to 21, and no correlation between panicles number per plant and grain number per panicle was found. For instance, Shafagh, Shiroudi, Saleh, and Sepidroud cultivars showed the larger number of panicles per plant, but did not possess the larger number of seeds per panicle. Doroudzan exhibited the heaviest panicle (5.1 g), but did not have the heaviest 1000-grain weight. Panicle length in Iranian cultivars ranged from 24.2 to 32.5 cm. The number of seeds per panicle among Iranian cultivars ranged from 46 to 138, and there was a coincidence with panicle shape, as an example in Qasredashti, which possessed the lowest seeds per panicle and exhibited an open panicle. Khazar which produced the largest seeds per panicle, possesses an intermediate panicle shape. Based on the results, the cultivars which possessed compact or intermediate panicles had a higher seed number per panicle compared to those cultivars which possessed an open panicle shape. Iranian cultivars have open and intermediate panicles. The 1000-grain weight in Iranian cultivars ranged from 18.6 to 29.6 g (Table 1). Plant height among cultivars was measured and classified into three groups (short, intermediate, and tall group) (Table 2).

The seed length in Iranian cultivars (Fajr, Tabesh, Nemat, and Dorfak) ranged from 11.1 to 11.5 mm, whereas the length was 10.1 mm in Basmati and 10.2 mm in Jasmine. This result indicated that most Iranian cultivars have longer grains than Basmati 370, and Jasmine 85, and this is a desirable character for aromatic rice in breeding programmes. Molecular analysis of aroma Two methods were applied to classify aromatic and non-aromatic rice cultivars in this study: 1.7% KOH sensory test and PCR analysis. Methods for smelling leaf tissue, grains after heating in water, and reacting with solutions of 1.7% KOH are routinely used (Sood and Siddiq 1978). In this research, leaf tissue after the reaction with 1.7% KOH solution was smelled. Basmati 370 was used as check cultivar for aroma, and Nipponbare was used as check cultivar for non-aroma. The results showed that six Iranian cultivars, Fajr, Shafagh, Dorfak, Zayandehroud, Sazandegi, and Ghasredashti, compared with Lawangi and Pashadi Konar from Afghanistan were aromaic, and the other Iranian cultivars also Gulnar and Shortanby from Uzbekistan were non-aromatic (Figure 2).

PCR analysis was applied to evaluate the cultivars, and confirmed that non-aromatic cultivars produced a 355 bp fragment and aromatic cultivars produced a 257 bp fragment. We found six of sixteen Iranian cultivars (Fajr, Shafagh, Dorfak, Zayandehroud, Sazandegi, Ghasredashti, and Lawangi and Pashadi Konar from

19

Table 1. Comparison of yield components among rice cultivars Variety Origin Panicle

number / plant

Panicle weight (g)

Panicle length (cm)

Seed number / panicle

1000 seeds weight (g)

Fajr Iran 18 ± 0.7 3.3±0.3 27.6±1.6 77 ± 13.3 18.6±0.8 Shafagh Iran 20 ± 2.4 4.5±0.3 27.5±1.9 74 ± 19.7 23.3±2.4 Pouya Iran 19 ± 2.5 3.7±0.6 30± 2.0 85 ± 35.5 24.0±2.4 Shiroudi Iran 20 ± 4.3 3.2±0.7 30.5±1.0 81 ± 30.0 22.8±1.3 Tabesh Iran 17 ± 0.4 3.6±0.4 31.5±3.8 73± 22.8 29.5±2.0 Nemat Iran 19 ± 1.9 4.3±1.0 32.5±1.3 112± 37.1 29.6±1.2 Neda Iran 18 ± 3.1 3.3±0.1 24.2±1.8 84 ± 9.5 24.3±0.1 Kadus Iran 18 ± 1.5 3.9±0.3 27.7±2.4 107 ± 41.2 24.7±1.9 Saleh Iran 20 ± 3.7 2.3±0.4 29.5±1.1 86 ± 18.2 22.6±2.8 Dorfak Iran 17 ± 2.0 3.3±1.0 28.5±2.2 122 ± 40.7 23.9±0.9 Sepidroud Iran 21 ± 3.0 3.3±0.4 27.5±1.5 79 ± 28.0 23.2±2.6 Khazar Iran 12 ± 0.6 3.9±0.5 29.8±1.2 138 ± 41.3 22.6±0.6 Zayandehroud Iran 17 ± 1.2 2.6±0.3 26.9±1.4 57 ± 21.9 21.3±1.0 Sazandegi Iran 16 ± 1.6 3.7±2.4 31.1±1.2 98 ± 35.0 22.9±1.5 Doroudzan Iran 16 ± 1.6 5.1±0.4 26.3±1.3 103 ± 15.7 25.5±2.1 Ghasredashti Iran 17 ± 3.3 3.1±0.8 30.7±0.5 46 ± 14.6 21.9±0.5 Gulnar Uzbekistan 9 ± 1.0 4.7±0.7 19±2.6 205±14.5 29.2±0.4 Shortanby Uzbekistan 11±1.5 7.1±0.5 18.6±0.5 118±5.8 28.3±4.5 Lawangi Afghanistan 12±04 3±0.5 29.0±6.0 123±28.0 27.5±1.1 Pashadikonar Afghanistan 10±0.3 4.9±0.6 33.0±0.9 148±21.0 32.4±3.5 Basmati 370 India 9±1.5 2.4±0.6 25.7±0.9 153±16.4 24.7±1.5 Jasmine 85 Thailand 10±2.0 4.2±0.6 23.7±0.5 149±32.0 26.8±1.3 Nipponbare Japan 10±1.5 3.8±0.7 23.4±0.2 125±8.6 25.4±2.0 Koshihikari Japan 10±1.2 4.1±0.2 24.2±0.4 128±7.5 26.0±1.9 ANOVA * * * 8 8 *: Significant difference at 5% level

20

Table 2. Comparison of plant height between aromatic and non-aromatic rice cultivars Plant height Genotypes Origin Type

Neda

Iran

Non-aromatic Shafagh Aromatic Kadus Non-aromatic Dorfak Aromatic Fajr Aromatic Shiroudi Non-aromatic Khazar Non-aromatic Nemat Non-aromatic Doroudzan Non-aromatic Sepidroud Non-aromatic Nipponbare Japan Non-aromatic Koshihikari Non-aromatic

Intermediate group (115.3 cm to 129.9 cm)

Saleh Iran

Non-aromatic Tabesh Non-aromatic Pouya Non-aromatic Gulnar Uzbekistan Non-aromatic Shortanby Non-aromatic Jasmine 85 Thailand Aromatic

Tall group (131.5 cm to 174.8 cm)

Zayandehroud Iran

Aromatic Sazandegi Non-aromatic Ghasredashti Aromatic Lawangi Afghanistan Aromatic Pashadi Konar Aromatic Basmati 370 India Aromatic

21

Figure 1. Relative positions of PCR primers used to distinguish between aromatic

and non-aromatic rice cultivars. External Sense Primer (ESP) and External Antisense Primer (EAP) generate a fragment of approximately 580 bp as a positive control for each sample. Internal Non-fragrant Sense Primer (INSP) and corresponding External Antisense Primer (EAP) produce 355 bp fragments from the non-aroma allele. Internal Fragrant Antisense Primer (IFAP) and corresponding. External Sense Primer (ESP) produce 257 bp fragments from aroma allele (Bradbury et al. 2005).

Figure 2. Comparison of aroma nature of Iranian rice cultivars by using molecular

markers, and 1.7% KOH sensory test (+ : aromatic, - : non-Aromatic); M: Marker, 1: Fajr, 2: Shafagh, 3: Pouya, 4: Shiroudi, 5: Tabesh, 6: Nemat, 7: Neda, 8: Kadus, 9: Saleh, 10: Dorfak, 11: Sepidroud, 12: Khazar, 13: Zayandehroud, 14: Sazandegi, 15: Doroudzan, 16: Ghasredashti, 17: Jamine 85 (check), 18: Nipponbare (check); 257 bp: aromatic, 355 bp: non-aromatic.

22

Afghanistan) were aromatic, and ten cultivars (Pouya, Shiroudi, Tabesh, Nemat, Neda, Kadus, Saleh, Sepidroud, Khazar, Doroudzan, and Gulnar and Shortanby from Uzbekistan) were non-aromatic. There was a strong correlation between 1.7% KOH sensory test and PCR analysis (Table 3). Based on this result, Iranian aromatic rice cultivars could be used for breeding programs.

Plant height, the number of panicles per plant and the grain number per panicle were analyzed in the F2 generation derived from the cross between Jasmine 85 and Nipponbare. The plant height in the parents ranged from 92.1 to 115.3 cm, the F1 plants ranged from 112.4 to 120.1 cm, and the F2 plants ranged from 98.6 to 114.6 cm. The average panicles per plant and grain number per panicle were also higher in F2 plants compared with parents. The improvements of these agronomic traits and morphological characters as well as aroma nature could be useful for future breeding program in Iran and surrounding regions.

For genetic analysis of aroma nature, 1.7% KOH sensory test and PCR analysis were applied to study the segregation of aroma in the F2 population derived from the cross between Jasmine 85 and Nipponbare, and in the F3 generation to confirm the result, as well as in the F2 generation derived from the cross between LTH and Pashadi Konar (Figure 3, 4, 5). The segregation ratio in the F2 was tested by χ2 analysis for a single gene model. The results indicated a segregation ratio of 3:1 of non-aromatic and aromatic categories.

T he F2 generation obtained from the cross between Jasmine 85 and Nipponbare (df = 1, χ2 = 0.1, P = 0.50 – 0.70), F3 generation obtained from the F2 plants (df = 1, χ2 = 0.06, P = 0.90 – 0.95), F2 generation obtained from the cross between LTH and Pashadi Konar (df = 1, χ2 = 0.2, P = 0.70 – 0.80) (Table 4).

DISCUSSION

In the present study, some important agronomic characters of aromatic and non-aromatic rice cultivars from Iran and two other surrounding countries were evaluated. In rice breeding programs, breeders have attempted to reduce the plant height to increase resistance to lodging as an important character in rice varieties. In this study, tall plants did not have enough lodging resistance but the results indicated that Iranian cultivars mainly belong to the short and intermediate group.

A general comparison of Iranian cultivars with Basmati 370 and Jasmine 85 showed that most Iranian cultivars have comparable number of panicles per plant, panicle lengths and weights, and grain lengths and widths. Two landraces, Fajr and Dorfak with desirable plant type and better yield components, longer grain, and desired aroma were identified for potential use in breeding programmes.

Fragrance is a recessive trait (Lorieux et al. 1996; Garland et al. 2000; Jin et al. 2003), which suggests it is a loss, rather than gain, of gene function that is responsible for fragrance. Several mapping studies have independently identified BADH2 as the candidate gene responsible for fragrance (Bradbury et al. 2005, Vanavichit et al. 2006; Amarawathi et al. 2008; Shi et al. 2008; Sarhadi et al. 2009) and sequencing of BADH2 in each case found a deletion within the gene which would render the gene non-functional. Iranian aromatic rice cultivars and Jasmine 85, produced same marker allele, as well as Iranian non-aromatic rice cultivars and

23

y y χCross lines Aroma

response of parents

F2 plants χ2- value

Probability

F3 plants χ2- value

Probability

Total Non- aroma : Aroma

Total Non-aroma : Aroma

Jasmine85/ Nipponbare

+/- 30 25:5 (3:1)

0.1 0.50-0.70 80 61:19 (3:1)

0.06 0.90-0.95

LTH/ PashdiKonar

+/- 50 35:15 (3:1)

0.2 0.70-0.80 _ _ _ _

Table 3. Characterization of aromatic and non aromatic rice cultivars Cultivar Origin KOH Sensory test Molecular marker Fajr Iran + A Shafagh Iran + A Pouya Iran - N Shiroudi Iran - N Tabesh Iran - N Nemat Iran - N Neda Iran - N Kadus Iran - N Saleh Iran - N Dorfak Iran + A Sepidroud Iran - N Khazar Iran - N Zayandehroud Iran + A Sazandegi Iran + A Doroudzan Iran - N Ghasredashti Iran + A Lawangi Afghanistan + A Pashadi Konar Afghanistan + A Gulnar Uzbekistan - N Shortanby Uzbekistan - N Jasmine 85 (Check) Thailand + A Nipponbare (Check) Japan - N Koshihikari (Check) Japan - N

+: Aroma, -: Non-aroma, A: Aroma, N: Non-aroma Table 4. Segregation of aroma nature in F2 and F3 populations of the cross between

(J x N), and F2 population of the cross between (LTH X Pashadi Konar)

analyzed by χ2 test value

24

Figure 3. Genetic analysis of aroma using molecular marker and 1.7% KOH

sensory test in F2 individuals obtained from the cross between Jasmine 85 and Nipponbare; Lane 1-14: F2 individuals; M: Marker, J: Jasmine 85, N: Nipponbare; (+ : aromatic, -: non-aromatic); (257 bp: Aromatic, 355 bp: non-aromatic).

Figure 4. Genetic analysis of aroma using molecular marker and 1.7% KOH sensory test in F2 individuals obtained from the cross between Jasmine 85 and Nipponbare. Lane 1-14: F2 individuals, M: Marker, J: Jasmine 85, N: Nipponbare, (+: Aromatic, -: Non-aromatic), (257 bp: Aromatic, 355 bp: Non-aromatic).

Figure 5. Genetic analysis of aroma in the F2 individuals obtained from the cross

between LTH andPashadi Konar using molecular marker and 1.7% KOH sensory test; Lane 1-20: F2 individuals; M: Marker, P1: LTH: Chines cultivar, P2: Pashadi Konar, (+ : Aromatic, -: Non- aromatic), (257 bp: Aromatic, 355 bp: Non-aromatic).

25

Nipponbare also produced the same marker allele. Segregation in aroma has been detected in both F2 generations derived from the crosses between Jasmine 85 and Nipponbare, and LTH and Pashadi Konar. Genetic and molecular analyses in the present study confirmed previous studies that aroma is controlling by a single recessive gene in all type of aromatic rices, and that the DNA marker is useful for genotyping rice germplasm.

In the present study, 1.7% KOH sensory test and molecular marker analysis were used to characterize the responsible gene for aroma in Iranian, Afghan, and Uzbek rice cultivars, as well as cross combinations. We found that the 1.7% KOH sensory test is a relatively cheaper and a simpler method to distinguish between aromatic and non-aromatic rice cultivars. It is therefore practical to use in many countries including Iran.

From this study, a few Iranian cultivars (Fajr, Shafagh, Dorfak, Zayandehroud, and Sazandegi) with good agronomic traits and desirable aroma were identified. These cultivars could be used for important resources for breeding programs. Basmati from India and Jasmine 85 from Thailand are known as standard aromatic rice across the world. Iranian aromatic rice cultivars are useful for the enrichment of the genetic base of breeding programs in Iran and surrounding countries.

ACKNOWLEDGEMENTS

We would like to thank from Dr. Takeshi Motobayashi and Dr. Taiichiro Ookawa for their strong assistance and support in field preparation.

REFERENCES

Agrama, H.A., G.C. Eizaga and W. Yan. 2007. Association mapping of yield and its components in rice cultivars. Mol Breeding 19: 341-356.

Aida,Y., H. Tsunematsu, K. Doi and A. Yoshimura. 1997. Development of a series of introgression lines of Japonica in the background of Indica rice. Rice Genet. Newsl. 14: 41-43.

Bourgis, F., R. Guyot, H. Gherbi, E. Tailliez, I. Amabile, J. Salse, M. Lorieux, M. Delseny and A. Ghesquiere. 2008. Characterization of the major fragrance gene from an aromatic japonica rice and analysis of its diversity in Asian cultivated rice. Theor. Appl. Genet. 117: 353-368.

Bradbury, L.M., S.A. Gillies, D.J. Brushett, D.L. Waters and R.J. Henry 2008. Inactivation of an aminoaldehyde dehydrogenase is responsible for fragrance in rice. Plant Mol. Biol. 68: 439-449.

Clausen, K., D. Keck and W. Hiesey. 1958. ‘Experimental Studies in the Nature of Species’, vol. 3: ‘Environmental Responses of Climatic Races of Achillea’ Carnegie Institution of Washington Publication 581, 1-129.

Gravois, K.A. and R.S. Helms. 1992. Path Analysis of Rice Yield and Yield Components as Affected by Seeding Rate. Agron. J. 84: 1-4.

26

Hien, N.L.,T. Yoshihashi, W.A. Sarhadi, V.C. Thanh, Y. Oikawa and Y. Hirata, 2006. Evaluation of Aroma in Rice (Oryza sativa L.) using KOH Method, Molecular Markers and Measurement of 2-Acetyl-1-Pyrroline Concentration. Nettai Nogyo. 50: 190-198.

Jaffuel, S. and J. Dauzat. 2005. Synchronism of Leaf and Tiller Emergence Relative to Position and to Main Stem Development Stage in a Rice Cultivar. Ann. Bot. 95: 401-412.

Jain, N., S. Jain, N. Saini and R.K. Jain. 2006. SSR analysis of chromosome 8 regions associated with aroma and cooked kernel elongation in Basmati rice. Euphytica 152: 259-273.

Linghe, Z. and M.C. Shannon. 2000. Effects of Salinity on Grain Yield and Yield Components of Rice at Different Seeding Densities. Agron. J. 92: 418-423.

Madoka, Y., T. Kashiwagi, N. Hirotsu, and K. Ishimaru. 2008. Indian rice “Kasalath” contains genes that improve traits of Japanese premium rice “Koshihikari”. Theor. Appl. Genet. 116: 603-612.

Mebrouk, B., A. Abderrahmane, and Z. Jun. 2005. QTL Analysis for yield components in rice (Oriza Sativa L.) under different environments. J. Central Eur. Agric. 6: 317-322.

Nadaf, A.B., S. Krishnan, and K.V. Wakte. 2006. Histochemical and biochemical analysis of major aroma compound (2-acetyl-1-pyrroline) in basmati and other scented rice (Oryza sativa L.). Curr. Sci. 91: 1533-1536.

Nguyen, T.L. and C.B. Bui. 2002. Identification and fine mapping of SSR marker linked to fgr gene of rice. Omonrice 10: 14-20.

Sanni, K.A., I. Fawole, R.G. Guei, D.K. Ojo, E.A. Somado, D.D. Tia, O.S. Ayoni and I. Sanvhez. 2008. Geographical patterns of phenotypic diversity in Oryza sativa landraces of Cote d’Ivoire. Euphytica 160: 389-400.

Sarhadi, W.A., T. Ookawa, A.K. Madadi, W. Yousofzai, Y. Oikawa and Y. Hirata. 2009. Characterization of aroma and agronomic traits in Afghan native rice cultivars. Plant Prod. Sci. 12: 63-69.

Wang, Y., and J. Li. 2005. The plant architecture of rice (Oryza sativa). Plant Mol. Biol. 59: 75-84.

Yoon, D.B., K.H. Kang, H.J. Kim, H.G. Ju, S.J. Kwon, J.P. Suh, O.Y. Jeong and S.N. Ahn. 2006. Mapping quantitative trait loci for yield components and morphological traits in an advanced backcross population between Oryza grandiglumis and the O. sativa japonica cultivar Hwaseongbyeo. Theor. Appl. Genet. 112: 1052-1062.

Zanne. A.E., C.A. Chapman and K. Kitajima. 2005. Evolutionary and ecological correlates of early seedling morphology in East African trees and shrubs. Am. J. Bot. 92: 972-978.

27

SABRAO Journal of Breeding and Genetics 43 (1) 27–43, 2011 INTER- AND INTRA-CLUSTER HETEROSIS IN SPRING TYPE OILSEED

RAPE (Brassica napus L) HYBRIDS AND PREDICTION OF HETEROSIS USING SRAP MOLECULAR MARKERS

RIAZ AHMAD1, FARHATULLAH*2 AND CARLOS F. QUIROS1

SUMMARY

To prove that hybrid canola (B. napus L.) offers any economic yield, or oil content advantage compared to inbred lines, we examined heterosis and its genetic basis in spring canola types hybrids at the University of California, Davis, USA. Eight restorer lines and 12 maintainer inbred lines of diverse origin and selected from 5 clusters based on SRAP molecular markers were crossed to produce 96 F1 hybrids. Analysis of variance among inbred lines and their F1 hybrids revealed significant differences for seed yield, oil content, plant height and days to 100% maturity. Ninety four percent of the hybrids surpassed their respective inbred lines in respect to seed yield, except for 6 single cross combinations. Substantial mid-parent heterosis (MPH) of 127% for yield was observed, for the hybrid (Dunkeld-S6 x R121) while negative heterosis of -24% was recorded for Altex x R-117. As a whole, the average yield of all inbred lines (982 kg ha-1) was much less than the average yields of all hybrids (1522kg ha-1). We selected six hybrids i.e., Dunkeld x R121 (2302 kg ha-1), Rainbow x R121 (2204 kg ha-1), Rainbow x R-117 ( 2109 kg ha-1), Dunkeld x R110 (2017 kg ha-1), Dunkeld x R111-1 (2014 kg ha-1), Rainbow x R-111 (2002 kg ha-1) for large scale evaluation because of their maximum yield (kg ha-1) as compared to the average seed yield of 1517kg ha-1 for the 96 hybrids. These hybrids had better oil content above 43% and having early, medium and late maturity with medium height. In general, restorer line ‘R121’ was a better general combiner followed by ‘R-111’ and ‘R-111-1’. Significant advantage in average yield (1553kg ha-1) was observed from inter-clustered crosses over intra-cluster crosses (1352kg ha-1). However non-significant differences were noted for oil content, plant height (cm) and days to 100% maturity by comparing inter and intra-cluster crosses. Regression analysis revealed statistically significant relationship between genetic distances of

Key words: Oilseed rape, heterosis, genetic diversity, molecular markers, cytoplasmic male sterility (CMS), restorer of fertility.

1 Department of Plant Sciences, University of California, Davis, CA 95616 USA. 2 Department of Plant Breeding and genetic, NWFP Agriculture University, Peshawar, Pakistan * Correspondence Author: [email protected]

28

the inbred lines with seed yield in their hybrid-derived MPH or high-parent hetrosis (HPH) form. The correlation coefficient for yield (0.60), MPH (0.59) and HPH (0.67) indicated a moderate relationship between the variables, so it can be expected that some of the SRAP markers be linked to QTLs for yield. However, no significant correlation was established between the genetic distance (GD) and oil content, plant height and maturity.

INTRODUCTION

Heterosis, or hybrid vigor, refers to the phenomenon that progeny of diverse inbred varieties exhibit such as greater biomass, speed of development, and fertility than the better of the two parents. This phenomenon has been exploited extensively in crop production and has been a powerful force in the evolution of plants. The genetic basis has been discussed for nearly a century (Shull, 1908; Jones, 1917), but little consensus has emerged. Development of successful hybrids in maize in 1930s (Duvick et al., 2004) provided an important impetus for breeders of other crops to exploit the available heterosis by means of commercial hybrids in other crops. Recent investigations leading to the development of new CMS lines and identification of maintainers and fertility restoration lines have brightened the prospects of development of commercial hybrids in Brassica.

The Brassica oil crops are the world’s third most important source of edible oil. Due to a continually increasing demand for rapeseed oil for food and non-food uses, the production of hybrid cultivars with higher seed and oil yields has become increasingly important in recent years. Exploitation of yield heterosis is now being viewed as an important avenue to break the yield barrier. Heterosis in brassica has been known since 1954 (Singh and Mehta, 1954), however, its effectiveness at commercial level has been demonstrated only during the past two decades. Significant heterosis for seed yield does exist in this crop. Yield increases of 10 to 72% over mid parent value have been observed in B. napus F1 hybrids (Sernyk and Stefansson, 1983; Grant and Beversdorf, 1985; Dhillon et al., 1996; Thakur and Sagwal, 1997), and have spurred interest in the development of hybrid cultivars.

The relationship of inter and intra-cluster genotypes with hybrid performance and heterosis for other traits resulted in variable results. In oilseed rape, crosses between parents of different origins generally exhibit greater heterosis than those related by common origin (Grant and Beversdorf, 1985; Brandlia and McVetty, 1990). According to Khanna and Misra (1977) in tomato and Moll et al. (1962) in corn, comparison of hybrid yield and genetic distance revealed that in general inter-cluster crosses yielded more than intra-cluster crosses. Ali et al. (1995) observed in Canola that inter-cluster heterosis was greater than intra –cluster heterosis. It has been reported that measures of genetic similarity based on RFLPs and pedigree knowledge could be used to predict superior hybrid combinations in maize (Smith et al., 1990). Both low and high correlations between heterosis and DNA – based genetic distance have been reported in various other crops (Barbosa et al., 1996; Melchinger et al., 1990). For example, Diers et al. (1996), Knaak and Ecke (1995), Becker and Engqvist (1995) observed a significant correlation between genetic

29

distance and heterosis for seed yield in Brassica inbred lines. In contrast, Lee et al. (1989), Godshalk et al. (1990) and Melchinger et al. (1990) reported weak correlations between hybrid performance and genetic distance in corn. Similarly, weak correlations have been reported for B. napus L. (Diers et al., 1996), wheat (Martin et al., 1995) and rice (Zhang et al., 1995a). Jain et al. (1994) could not find a consistent correlation between the genetic distances of RAPDs and single cross performance in Brassica juncea L. for grain yield or heterosis. Based on these conflicting observations noted by various researchers, it will be interesting to study the relationship in genetic distance and hybrid performance in Brassica using molecular markers. If the measure of genetic similarity or genetic distance between the parents could provide a clue for predicting the progeny variance of crosses, it would increase the efficiency of breeding programs by concentrating efforts on the most promising crosses minimizing line testing in field trials.

Objectives of the present study were to (1) evaluate F1 hybrids of B. napus for seed yield, oil content, plant height and maturity (ii) determine the levels of mid-parent and high parent heterosis for these characters and (iii) establish a correlation between genetic distance revealed previously by the sequence-related amplified polymorphism (SRAP) molecular markers and heterosis for various agronomic traits in their derived F1 hybrids.


Plant material In the Brassica napus hybrid breeding program, we pollinated a cytoplasmic male sterile (CMS) line with large number of promising spring type breeding lines. In such test crosses, the parents that produce complete male sterile hybrids are classified as potential maintainers (M) and are utilized for further backcrossing to CMS lines. The lines that produce hybrids of high level of pollen fertitlity by crossing with the CMS lines hybrids of high level of pollen fertility are classified as potential restorers (R). Based on the SRAP molecular markers and UPGMA cluster analysis (Ahmad et al. in press), we select 12 maintainer lines (Dunkeld-S7, Rainbow-S6, Oscar-S6, Siren-S6, Maluka-S5, Shiralee-S6, Altex-S5, Westar-S6, Tower-S6, Salam-S6, Topas-S6, Bulbul-S5) and 8 restorer lines (R-101, R-115, R-110, R-121, R-117, R-111, R-111-1, R-103) of diverse origin with different level of genetic diversity and representing 5 different clusters (Table 1). The maintainer and restorer lines were selfed for at least 5-7 generations. By crossing maintainer lines (near isogenic lines of the cytoplamic male sterile, CMS lines) as female and restorer line as male, we developed 96 F1 hybrids. The crosses were made in the green house by hand pollination (Table 3).

Planting of hybrids The resulting F1 hybrids along with their parents were planted at the experimental fields of the Department of Plant Sciences, University of California at Davis, United States. In the field randomized complet block (RCB) design were used with three replications. The plot size was 1.5 × 5 m with a row to row distance of 45 cm. All three rows were used for seed yield and converted to kg ha-1. Bulk samples from

30

each plot were sent for oil content and were measured by nuclear magnetic resonance (NMR) analyzer (Hammond, 1991). Ten mature plants in each plot were used for plant height (cm) while days to maturity were taken when 80% of pods became dried. Standard fertilization and plant protection measures were applied.

Table 1. Plant material of spring type B. napus inbred lines used for F1 crosses.

The cluster grouping was based on the SRAP markers and UPGMA analysis. The designation of quality type given in brackets indicates high (+) or zero (0) erucic acid and high (+) or low (0) glucosinolate content, respectively. OP represents open pollinated while S showed the number of selfing.

Code Accessions Sample type Cluster Origin Maintainer/ Restorer

Quality type

M1 Dunkeld-S7 S7-Inbred I Australia M 00 M2 Rainbow-S6 S6-Inbred I Australia M 00 M3 Oscar-S6 S6-Inbred I Australia M 00 M4 Siren-S6 S6-Inbred II Australia M 00 M5 Maluka-S5 S5-Inbred I Australia M ++ M6 Shiralee-S6 S6-Inbred I Australia M 00 M7 Altex-S5 S5-Inbred I Canada M 00 M8 Westar-S6 S7-Inbred II Canada M 00 M9 Tower-S6 S6-Inbred II Canada M 00 M10 Salam-S6 S6-Inbred V New Zealand M ++ M11 Topas-S6 S6-Inbred IV Sweden M 00 M12 Bulbul-S5 S5-Inbred I Pakistan M 00 R1 R-101 S7-Inbred III Pakistan R 00 R2 R-115 S6-Inbred I Pakistan R 00 R3 R-110 S6-Inbred II Pakistan R 00 R4 R-121 S6-Inbred II Pakistan R 00 R5 R-117 S6-Inbred I Pakistan R 00 R6 R-111 S7-Inbred III USA R 00 R7 R-111-1 S6-Inbred III USA R 00 R8 R-103 S6-Inbred II USA R 00

Statistical analysis

Data were subjected to analysis of variance, and both Mid Parent and High Parent heterosis estimates were determined by the formulae.

MPH = (F1-MP) x 100

MP

HPH = (F1-HP) x 100

HP

Where, F1 = Hybrid Mean MP = Mid Parent HP = High Parent

31

In order to determine possible associations between heterosis and genetic distance, all measured agronomic traits described above as well as their respective Mid Parent and High Parent Heterosis values were regressed to genetic distances (Nei and Li, 1979) determined by the molecular markers.

RESULTS

Performance of parents and hybrids Seed Yield (kg ha-1) The analysis of variance indicated significant differences among inbred lines and their hybrids for yield (kg ha-1). In general, inbred lines had poor yield as compared to hybrids. Restorer inbred lines had poor yield (920 kg ha-1) compared to the maintainer lines (1021 kg ha-1). On average, all inbred lines had yield of 982 kg ha-1 which was much less than the average yield of all hybrids (1522 kg ha-1) (Table 2).

Hybrids which include the maintainer line “Rainbow” as one female parent produced more average yield (1907 kg ha-1) when crossed with all restorer and as such was the best maintainer line to be used for hybrid production followed by the inbred line “Dunkeld-S7” (1906 kg ha-1). Poor general combiner in terms of yield was the maintainer inbred line “Maluka-S6” which gave poor yield of (1216 kg ha-1) combined over all restorer inbred lines followed by the Altex (1279 kg ha-1) and Shiralee (1387 kg ha-1). Similarly, restorer line “R121” was a better general combiner when crossed with all maintainer lines by giving an average yield of (1781 kg ha-1) followed by R-111 (1744 kg ha-1) and R-111-1 (1703 kg ha-1). Low average yield of 1299 kg ha-1 and 1344 kg ha-1 was recorded for restorer lines R110 and R-103 when crossed with all maintainer lines and thus were poor general male restorer (Table 3).

Six hybrids i.e., Dunkeld × R121 (2302 kg ha-1), Rainbow × R121 (2204 kg ha-1), Rainbow × R-117 (2109 kg ha-1), Dunkeld × R110 (2017 kg ha-1), Dunkeld × R111-1 (2014 kg ha-1), Rainbow × R-111 (2002 kg ha-1) gave maximum yield kg ha-1 in all 96 hybrids where the average yield was 1517 kg ha-1 (Table 3).

On average, hybrids had 56% of higher yield than the mid parent while 46% increase was recorded when compared to the better parent. Maximum yield heterosis of 127% was observed for the inbred combination (Dunkeld-S6 × R121) while negative heterosis in terms of yield kg ha-1 was recorded for Altex × R-117. Similarly, comparison of hybrids with the better parent showed that maximum gain in yield 107% was observed for Siren × R121 and Siren × R-115. Least gain in yield of -26 was recorded for the cross combination Altex x R117 when compared to the better parent (Table 3). Further, we selected a hybrid (Siren × R111-1) for further evaluation. This hybrid have medium yield performance (1895 kg ha-1), but have the highest oil content of 46.32%. This was a late maturing, medium stature hybrid.

All hybrids having yield below the average 1517 kg ha-1 were rejected. Ten hybrids (Table 3) had the worst yield performance (742-923 kg ha-1). Heterosis and heterobeltiosis were negative for these hybrids in terms of yield. Hence, they can not be considered for hybrid seed production.

32

Oil content The analysis of variance indicated significant differences among inbred lines and their hybrids for oil contents. Mean yield for the maintainer lines was 41.7% while 41.4% for the restorer lines with an average of 41.6% for all inbred lines. Restorer line R-111-1 was the best general combiner in terms of oil content by recording an average of 43.8% in hybrids where it was one of the parent followed by R121 by giving 42.8%. Poor general combiner was the early maturing, short inbred line R-103 combination with other maintainer lines by giving an average of 40.7% of oil content in hybrids where R-103 was one of the male parents (Table 2).

No significant difference in oil content was observed between the average oil content of all inbred lines (42.13%) and their respective 96 hybrids (42.19%). Maximum oil content of 46.32 was recorded for only one hybrid (Siren × R111-1) and was therefore selected for further evaluation in spite of the low yield of 1895 kg ha-1, followed by the hybrids (Dunkeld × R111), (Dunkeld × R121), (Rainbow × R121), giving 45.32, 45.09 and 45.06 % oil respectively. Low oil content of 40.02 % was observed for the hybrid (Maluka × R-103), closely followed by the (Wester × R-103) and (Tower × R103) (Table 3).

Maximum heterosis of 13.94 was observed for the hybrid (Siren × R111-1), followed by (Oscar × R111-1) with 10.725, (Siren × R111) with 9.11, (Dunkeld × R111) with 9.07 of oil content increase. Minimum heterosis of -27 was observed for hybrid (Tower × R101) (Table 3). Comparing of hybrids with the better parent showed that on average 1.26% increase in oil content was observed in hybrids. Maximum of 12.97 heterobeltiosis was observed for the cross combination (Siren × R-111-1) followed by (Oscar × R111), (Siren × R111) and (Oscar × R111-1) with increase of 9.8%, 8.3% and 9.82% respectively.

33

Table 2. Yield kg ha-1, oil content (%), plant height (cm), and days to maturity of 12 maintainer and 8 restorer inbred lines used in the development of 96 hybrids.

Inbred parents Yield kg ha-1

Oil content (%)

Plant Height (cm)

Days to maturity (%)

Dunkeld-S7 1223 44.1 120 170 Rainbow-S6 1226 43 103 165 Oscar-S6 1165 44.9 113 177 Siren-S6 921 40.3 142 187 Maluka-S5 856 40.2 115 170 Shiralee-S6 1001 41.2 140 177 Altex-S5 1006 40.9 127 177 Westar-S6 846 40.8 130 174 Tower-S6 935 40.9 120 172 Salam-S6 1006 40.3 112 190 Topas-S6 948 40.6 105 173 Bulbul-S5 1113 43.1 110 165

Maintainer Average 1020.5 41.69 119.75 174.75

R-101 1025 40.2 93 155 R-115 856 43.3 120 180 R-110 889 41.3 115 165 R-121 803 42.1 102 170 R-117 947 42.5 115 175 R-111 954 40.9 106 167 R-111-1 888 41 104 169 R-103 995 40.2 95 157 Restorer Average 919.625 41.44 106.25 167.25 All inbreds Average 982.071 41.59 114.61 171.89

34

Table 3. Genetic distance (GD), Heterosis and heterobelteosis for yield (kg ha-1), oil content%, plant height (cm) and maturity in 96 B. napus hybrids resulted from 12 maintainer and 8 restorers inbred lines.

♀ ♂ GD Yield MPH HPH Oil % MPH HPH Pt.H MPH HPH Maturity MPH HPH

Dunkeld –S7 R101 0.359 1704 52 39 43.89 6.53 4 160 50 33 165 1.53 -2.94

Rainbow-S6 0.349 1845 64 50 42.34 2.89 0.57 140 42 36 165 3.12 0

Oscar-S6 0.339 1625 48 39 41.23 2.38 2.21 162 57 43 165 -0.6 -6.77

Siren-S6 0.139 1520 56 48 42.34 5.19 5.06 165 40 16 175 2.33 -6.41

Maluka-S5 0.301 1120 19 9 40.89 1.72 1.7 160 54 39 160 -1.53 -5.88

Shiralee-S6 0.288 1114 10 8 40.9 0.49 -0.73 156 34 11 165 -0.60 -6.77

Altex-S5 0.295 1213 19 18 40.34 -0.52 -1.37 160 45 26 165 -0.60 -6.77

Westar-S6 0.129 1514 62 48 40.35 -0.37 -1.1 140 26 8 170 3.34 -2.29

Tower-S6 0.131 1304 33 27 40.44 -27 -1.12 145 36 21 170 3.97 -1.16

Salam-S6 0.368 1082 5 -9 40.39 0.34 0.22 166 62 48 175 1.44 -7.89

Topas-S6 0.359 999 2 -2.5 40.33 -0.17 -0.67 134 35 28 175 6.70 1.156

Bulbul-S5 0.353 1504 41 35 41.32 -0.79 -4.1 167 64 52 168 5 1.818

Dunkeld –S7 R115 0.043 1602 54 31 42.32 0.17 0.05 155 29 29 170 -2.85 -5.55

Rainbow-S6 0.045 1506 45 23 43.44 2.94 2.7 165 48 38 186 7.82 3.333

Oscar-S6 0.232 1622 60 39 43.32 4.84 2.41 166 43 38 172 -3.64 -4.44

Siren-S6 0.283 1902 114 107 44.32 7.3 4.78 170 30 20 187 1.90 0

Maluka-S5 0.235 1435 68 68 40.32 -2.25 -4.68 157 34 31 165 -5.71 -8.33

Shiralee-S6 0.239 1526 64 52 41.35 -0.96 -2.25 190 46 35 165 -7.56 -8.33

Altex-S5 0.238 1314 41 31 42.34 1.78 0.094 165 33 30 168 -5.88 -6.66

Westar-S6 0.286 1613 89 88 42.36 1.95 0.14 166 33 28 168 -5.08 -6.66

Tower-S6 0.285 1504 68 61 43.44 4.4 2.7 167 39 39 165 -6.25 -8.33

Salam-S6 0.609 1526 64 52 42.34 2.5 0.09 170 47 42 175 -5.40 -7.89

Topas-S6 0.474 1215 35 28 43.36 4.61 2.51 165 47 38 175 -0.84 -2.77

Bulbul-S5 0.041 1503 53 35 42.36 -0.79 -1.72 155 35 29 170 -1.44 -5.55

Dunkeld –S7 R110 0.298 2017 91 64 43.34 3.81 2.7 145 28 25 170 1.49 0

Rainbow-S6 0.293 1856 75 51 44.32 6.28 5.27 149 37 30 174 5.45 5.454

Oscar-S6 0.283 1814 77 56 43.34 6.17 4.94 150 32 30 172 0.58 -2.82

Siren-S6 0.118 1439 59 56 43.44 6.7 5.12 160 25 13 180 2.27 -3.74

Maluka-S5 0.288 839 -3 -6 40.42 -0.81 -2.13 155 35 35 165 -1.49 -2.94

Shiralee-S6 0.278 1235 31 23 40.34 -2.21 -2.09 155 22 11 165 -3.50 -6.77

Altex-S5 0.283 1019 8 2 41.34 0.5 0.09 150 24 18 163 -4.67 -7.90

Westar-S6 0.082 924 6 4 40.36 -1.68 -2.27 150 22 15 170 0.29 -2.29

Tower-S6 0.115 1011 11 8 40.39 -1.73 -2.2 152 29 27 174 3.26 1.162

Salam-S6 0.412 926 -2 -7 40.4 -0.98 -2.18 165 45 43 180 1.40 -5.26

Topas-S6 0.409 1010 9 6 41.32 0.9 0.04 150 36 30 175 3.55 1.156

Bulbul-S5 0.296 1506 50 35 41.56 -1.52 -3.57 154 37 34 172 4.24 4.242

Dunkeld –S7 R103 0.289 1704 54 39 42.6 3.39 0 162 51 35 171 4.58 0.588

Rainbow-S6 0.285 1805 62 47 41.34 2.02 1.55 145 47 41 166 3.10 0.606

Oscar-S6 0.279 1604 48 37 40.32 0.12 0 145 39 28 167 0 -5.64

35

Table 3. Genetic distance (GD), Heterosis and heterobelteosis for yield (kg ha-1), oil content%, plant height (cm) and maturity in 96 B. napus hybrids resulted from 12 maintainer and 8 restorers inbred lines (cont.)


Siren-S6 0.109 1413 47 42 40.89 1.59 0 160 35 13 177 2.90 -5.34

Maluka-S5 0.275 932 0.6 -6 40.02 -0.45 -0.45 156 49 36 160 -2.14 -5.88

Shiralee-S6 0.269 1113 12 11 41.32 1.5 0 155 32 11 167 0 -5.64

Altex-S5 0.273 1214 21 21 40.32 4.32 4.9 150 35 18 162 -2.99 -8.474

Westar-S6 0.073 796 -14 -20 40.03 -1.16 -0.42 145 28 12 172 3.92 -1.14

Tower-S6 0.105 923 -4 -7 40.03 -1.28 -0.42 150 40 25 170 3.34 -1.16

Salam-S6 0.482 1514 51 50 40.23 -0.05 0 165 59 47 173 -0.28 -8.94

Topas-S6 0.439 1635 68 64 40.32 -2.97 -2.78 160 60 52 172 4.24 -0.57

Bulbul-S5 0.287 1465 39 31 40.9 -1.8 0 155 51 41 168 4.34 1.81

Dunkeld –S7 R121 0.299 2304 127 88 45.09 5.47 4.13 150 31 21 180 5.88 5.88

Rainbow-S6 0.296 2204 117 80 45.06 5.5 4.06 149 45 45 185 10.44 10.44

Oscar-S6 0.286 1905 94 64 44.33 6 2.38 155 44 37 176 1.44 1.44

Siren-S6 0.119 1904 121 107 43.24 3.44 -0.14 167 37 18 189 5.88 5.88

Maluka-S5 0.289 1004 21 17 40.2 -3.7 -7.16 155 43 35 165 -2.94 -2.94

Shiralee-S6 0.279 1804 100 80 41.32 2.72 0.23 156 29 11 165 -4.89 -4.89

Altex-S5 0.286 1623 80 61 41.56 3.18 0.32 150 31 18 170 -2.0 -2.01

Westar-S6 0.086 1603 95 89 42.02 3.8 0.81 145 25 12 167 -2.90 -2.90

Tower-S6 0.118 1804 108 93 41.67 2.9 0.05 150 35 25 166 -2.92 -2.92

Salam-S6 0.409 1506 66 50 41.32 4.45 0.83 155 45 38 190 5.55 5.55

Topas-S6 0.398 1809 107 91 43.24 3.08 -0.14 150 45 42 180 4.95 4.95

Bulbul-S5 0.298 1902 99 71 44.44 2.87 2.63 155 46 41 175 4.47 4.47

Dunkeld –S7 R117 0.239 1909 76 56 44.32 4.65 4.28 160 36 33 172 -0.28 1.17

Rainbow-S6 0.233 2109 94 72 43.71 3.33 2.85 155 42 35 185 8.82 5.71

Oscar-S6 0.169 1804 70 55 43.67 5.43 2.75 155 36 35 172 -2.27 -2.82

Siren-S6 0.283 1604 72 69 43.34 4.68 1.97 164 28 15 189 4.41 1.06

Maluka-S5 0.158 742 -18 -22 40.35 -2.42 -5.06 150 30 30 165 -4.37 -5.71

Shiralee-S6 0.143 826 -15 -17 40.36 -3.56 -5.04 160 30 7 160 -9.09 -9.60

Altex-S5 0.093 743 -24 -26 41.34 -0.86 -2.73 155 28 22 166 -5.68 -6.21

Westar-S6 0.289 1014 13 7 41.32 -0.79 -2.78 160 31 23 167 -4.29 -4.57

Tower-S6 0.286 1213 29 28 43.22 3.64 1.69 165 40 38 166 -4.32 -5.14

Salam-S6 0.508 1636 67 63 41.32 -0.19 -2.78 170 50 48 185 1.36 -2.63

Topas-S6 0.453 1807 91 91 40.34 -2.9 -5.08 160 45 39 180 3.44 2.85

Bulbul-S5 0.237 924 -10 -17 42.34 -1.07 -1.76 165 47 43 170 0 -2.85

Dunkeld –S7 R111 0.329 1990 83 63 45.32 9.07 7.39 147 30 23 172 2.07 1.17

Rainbow-S6 0.323 2002 84 63 44.9 8.19 6.65 150 43 42 175 5.42 4.79

Oscar-S6 0.309 1887 78 62 44.89 10.51 9.76 155 42 37 172 0 -2.82

Siren-S6 0.128 1945 107 104 44.3 9.11 8.31 160 29 13 180 1.69 -3.74

Maluka-S5 0.297 1530 69 60 41.32 1.89 1.02 150 36 30 163 -3.26 -4.11

Shiralee-S6 0.289 1645 68 64 40.32 -1.78 -2.13 152 24 9 165 -4.06 -6.77

36

Table 3. Genetic distance (GD), Heterosis and heterobelteosis for yield (kg ha-1), oil content (%), plant height (cm) and maturity in 96 B. napus hybrids resulted from 12 maintainer and 8 restorers inbred lines (cont.)


Altex-S5 0.295 1620 65 61 41.32 1.03 1.02 150 29 18 164 -4.65 -7.34

Westar-S6 0.115 1530 70 60 41.87 2.49 2.37 154 31 18 170 -0.29 -2.29

Tower-S6 0.126 1630 73 71 41.32 1.03 1.03 155 37 29 176 3.83 2.32

Salam-S6 0.385 1625 66 62 40.32 -0.69 -1.42 160 47 43 190 6.44 0

Topas-S6 0.376 1730 82 81 41.32 1.39 1.03 156 48 47 170 0 -1.73

Bulbul-S5 0.325 1805 75 62 44.02 4.81 2.13 152 41 38 172 3.61 2.99

Dunkeld –S7 R111-1 0.331 2014 91 65 44.78 7.64 6.11 150 34 25 173 2.06 1.76

Rainbow-S6 0.326 1930 83 57 44.39 6.83 5.43 152 47 46 176 5.38 4.14

Oscar-S6 0.312 1776 73 52 45.03 10.72 9.82 153 41 35 172 -0.57 -2.82

Siren-S6 0.131 1895 110 106 46.32 13.94 12.97 158 28 11 182 2.24 -2.67

Maluka-S5 0.299 1430 64 61 42.43 4.51 3.48 152 39 32 164 -3.24 -3.52

Shiralee-S6 0.291 1560 65 56 42.44 3.2 3.01 155 27 11 167 -3.46 -5.64

Altex-S5 0.296 1425 50 41 41.3 0.85 0.73 149 30 17 164 -5.20 -7.34

Westar-S6 0.129 1625 87 83 41.89 2.42 2.17 152 37 17 169 -1.45 -2.87

Tower-S6 0.131 1730 89 85 43.44 6.08 5.95 153 49 28 176 3.22 2.32

Salam-S6 0.379 1726 82 72 44.46 9.37 8.43 161 50 44 189 5.29 -0.52

Topas-S6 0.371 1625 77 71 44.43 8.89 8.37 157 42 50 172 0.58 -0.57

Bulbul-S5 0.329 1703 70 53 44.39 5.56 2.99 152 38.6 38 174 4.19 2.95

Average 1517 56 46 42.19 2.23 1.25 156 38.6 29 172 0.56 -2

Max 2304 127 107 46.32 13.94 12.97 190 64 52 190 10 10

Mini 742 -24 -26 40.02 -27 -7.16 134 22 7 160 -9 -9

Plant height The analysis of variance indicated significant differences among inbred lines and their hybrids for plant height (cm). Average plant height for maintainer lines was 119.8 cm while for restorer inbred lines was 106 cm with an average of 114.6cm. Two inbred lines, Rainbow (103 cm) and Topas (105 cm) were the shortest while Siren and Shiralee were the tallest with plant height of 142 cm and 140 cm, respectively. Restorer inbred lines R101 and R103 were the shortest with plant height of 93cm and 95cm, respectively while R-115 and R-110 were the tallest with plant height of 120 cm and 115 cm (Table 2).

The short statured inbred lines like R-101 and R103 mostly resulted in shorter hybrid when combined with shorter maintainer lines like Rainbow and Topas. In general, average plant height for hybrids (155.9 cm) was significantly more than the average inbred line height of 114.6cm. Tallest hybrids of 190 cm was result of cross combination “Shiralee” and “R-115”, followed by Salam × R117 (170 cm), Salam × R-115 (170 cm) and Siren × R115 (170 cm). Shorter hybrid of 134 was Topas × R101 followed by Rainbow × R101 and Westar × R101 of 140cm plant height (Table 3).

37

Positive heterosis and heterobeltiosis of 64% and 52% for plant height was observed in the hybrid (Bulbul × R101) while negative heterosis of 22% towards shortness was observed for the hybrids (Westar × R117), (Shiralee × R110), (Westar × R110) while negative heterobeltiosis of 7% reduction in plant height was observed for the hybrid (Shiralee × R117) (Table 3).

Days to maturity The analysis of variance indicated significant differences among inbred lines and their hybrids for days to maturity when 90% of the pods were turned yellow. Inbred lines Rainbow and Bulbul were the early maturing maintainer lines while Salam and Siren were the late. In restorer lines, R101 and R103 mature early within 155 and 157 days respectively while restorer R115 was late maturing in 180 days (Table 2).

No significant effect for average maturity was observed between the inbred lines (172 days) and in all 96 hybrids (171.88 days). Effect of early inbred lines R101 and R103 was evident in all those hybrids where these two were one of the male parents. Late hybrids with days to maturity of 190 days were observed in hybrids (Salam × R121) and (Salam × R111) followed closely by the hybrids (Siren × R121) and (Siren × R117) with 189 days of maturity. We also observed early maturing hybrids (Shiralee × R117), (Maluka × R103), (Maluka × R101) with 160 days to maturity (Table 3). Maximum heterosis of (10.45) for maturity was recorded in the hybrid (Rainbow × R121). Minimum heterosis and heterobeltiosis was recorded for the hybrid (Shiralee × R117) (Table 3).

Genetic distance and hybrid performance We observed that hybrids resulting from inter-cluster parents yielded significantly more seed (1553 kg ha-1) as compared to the intra-cluster crosses (1352 kg ha-1). For example, inter-cluster crosses, Rainbow × R121 (2204 kg ha-1), Dunkled x R121 (2304 kg ha-1), Oscar × R121 (1905 kg ha-1) had higher yields than intra-cluster crosses Westar × R121 (1603 kg ha-1), Siren x R121 (1904 kg ha-1) and Tower × R121 (1804 kg ha-1). Similar results were also found for other hybrids like the cross combinations of restorer R117 within the cluster with Altex, Maluka, resulted in low yield of 743 kg ha-1, 742 kg ha-1, as compared to the crosses from the inter-cluster inbred lines Siren (1604 kg ha-1), Tower (1213 kg ha-1), Salam (1636 kg ha-1) and Topas (1807 kg ha-1) (Table 4). Similarly combinations of intra-cluster crosses of restorer line R103 resulted in much lower yield of 796 kg ha-1, 1413 kg ha-1 with the maintainer lines Westar and Siren, respectively as compared to the inter-cluster crosses Rainbow × R103 (1805 kg ha-1), Dunkeld × R103 (1805 kg ha-1), Topas × R103 (1635 kg ha-1) and Salam × R103 (1514 kg ha-1). The analysis indicated that there was a statistically significant relationship at the 95% confidence level between genetic distance and seed yield, its respective MPH or its HPH (Table 4, Figure 1). The R2 statistic indicates that the model as fitted explains 35.40% of the variability for seed yield, 35.12% in mid parent and 37.04% in high parent heterosis. The correlation coefficient for yield (0.60), MPH (0.59) and HPH (0.67) indicates a moderate relationship between the variables.

Non significant increase in oil content was observed by comparing the intra- cluster hybrids (41.71%) to the inter-cluster hybrids (42.28%). Similarly no

38

significant relationship was established for plant height and maturity between inter- and intra-cluster crosses. Correlation coefficients for genetic distance and oil content (0.20), plant height (0.19), maturity (0.25), and their respective MPH and HPH show relatively weak relationships (Table 4).

DISCUSSION

For a reliable and economical hybrid seed production, three biological requirements must be fulfilled: (i) presence of hybrid vigor, (ii) prevention of self-pollination of the female parent, (iii) adequate pollination by the male parent (Wright, 1980). To produce hybrids in canola, we developed Cytoplasmic Male Sterile (CMS) lines to prevent self pollination. These CMS lines were back crossed with the maintainer lines for 5-7 generations and were nearly isogenic to the CMS lines. We achieved adequate pollination of the CMS lines by developing the restorer lines. Majority of the inbred lines used in the present study were derived from double-low seed quality (zero erucic acid, low glucosinolate content) spring type oilseed rape. We examined 96 hybrids along with their parents for yield performance, oil content, plant height and maturity and mid-parent and high parent heterosis. The analysis of variance indicated significant differences among inbred lines and their hybrids for yield (kg ha-1), oil contents (%), plant height (cm) and days to maturity. Ninty four percent of the hybrids surpassed their respective inbred lines in respect to seed yield, except for 6 single cross combinations. In general, inbred lines were less vigorous due to selfing for 5-7 generations and accumulation Table 4. Correlation coefficient between genetic distance (GD) resulting from

SRAP markers, and yield (kg ha-1), oil content (%), plant height (cm) and maturity in 96 B. napus hybrids and their respective mid-parent (MPH) and high-parent (HPH) heterosis estimates.

Yield MPH HPH Oil % MPH HPH Pt.H MPH HPH Maturity MPH HPH

Correlation

Coefficient

0.60 0.59 0.67 0.20 0.35 0.25 0.19 0.30 0.19 0.25 0.19 0.18

R2 35.40 35.12 37.04 8.02 14.34 0.30 4.3 12.45 3.02 10.58 6.75 5.28

39

Figure 1. Simple linear regression of yield (kg ha-1) on genetic distance (GD) resulting from SRAP molecular markers.

of deleterious alleles which resulted in low mean seed yield of 982 kg ha-1 from inbred lines as compared to the average yield of 1517 kg ha-1 for the hybrids. In spite of the adverse effect of inbreeding on the inbred lines, average yield for these lines were still sufficient to maintain these lines for production of hybrid seeds. Substantial mid-parent heterosis (MPH) of 127% for yield was observed, for the hybrid (Dunkeld-S6 x R121) while negative heterosis of -24% was recorded for Altex x R-117. Normally, seed yield and oil content are negatively correlated, but we were able to select six hybrids (Dunkeld x R121, Rainbow x R121, Rainbow x R-117, Dunkeld x R110, Dunkeld x R111-1, Rainbow x R-111) for high yield advantage and better oil content. These hybrids were recommended for large scale evaluation in different environment since these include early, medium and late maturity hybrids with medium stature which were resistant to lodging. We further noted that three restorer line “R121” “R-111’, R-111-1 were excellent general combiner. The one late hybrid (Siren x R111-1) having the highest oil content of 46.32%, was also selected to be used in a breeding program in spite of its low yield (1895 kg ha-1). Heterosis in canola seed yield has also been reported by Grant and Beversdorf (1985) and Lefort- buson et al. (1987) but to a little lesser extent than the values reported in our study. Starmer et al. (1998) observed that canola hybrids displayed 17% higher yield than mid parent performance. The heterotic advantage in our hybrids over inbred lines was sufficiently large to justify increased cost of hybrid seed production.

No specific pattern with respect to oil content was observed for hybrids and their respective inbred lines. Oil content of several hybrids was higher than that observed for the inbred lines but the difference was not significant. However, some possible heterosis over the mid and high parent was observed for the cross combination (Siren x R111-1, Oscar x R111-1, Siren x R111, Dunkeld x R111). According to Gupta and Khanna (1982) and Fick and Miller (1997), heterosis for

40

seed yield is usually significant in sunflower, while heterosis for seed oil concentration is seldom significant. Our present finding about the heterosis in oil content are according to the result of Starmer (1998) that oil content showed heterotic advantage, although to a lesser degree, perhaps due to its oligogenic character.

Inbred lines were significantly shorter than the hybrids because of inbreeding for five to seven generations. In general, shorter hybrids were produced by crossing of shorter restorer lines like R-101 and R-103 with the shorter maintainer lines like Rainbow and Topas. Reverse observations were noted for crossing of tall into tall inbred lines. In canola, shorter hybrids are preferred due to lodging problem, but shorter hybrids normally produced less yield. So breeding objectives are directed towards selection of medium stature hybrids with more yield or selection of tall hybrids with resistance to lodging.

For different environments, selections for early, medium and late hybrids are of utmost importance. There is also a negative correlation between yield and maturity. Breeding objectives are always geared towards the selection of early or medium hybrids with less penalty for yield and oil content. One of our hybrids, (Dunkeld-S7 x R110) was in the desired direction towards earliness without the penalty of significant reduction in yield (2017 kg ha-1) and oil content (43.34 %).

Significant advantage in average yield (1553 kg ha-1) was observed from inter-clustered crosses over intra-cluster crosses (1352 kg ha-1). However non-significant differences were noted for oil content (%), plant height (cm) and days to 100% maturity by comparing inter and intra-cluster crosses. The relationship of inter and intra-cluster genotypes with hybrid performance and heterosis for other traits resulted in variable results. According to Khanna and Misra (1977) in tomato and Moll et al. (1962) in corn, comparison of hybrid yield and genetic distance revealed that in general inter-cluster crosses yielded more than intra-cluster crosses. Ali et al. (1995) observed in Canola that inter-cluster heterosis was greater than intra –cluster heterosis. Crosses between parents of different origins generally exhibit greater heterosis than those related by common origin (Grant and Beversdorf, 1985; Brandle and McVetty, 1990).

In order to pursue the idea of any relationship between inter and intra-cluster genotypes and heterosis, we perform a simple regression analysis between genetic distance of inbreds and their hybrid performance. The reason for the non-significant correlation of GD with oil concentration compared to the significant correlation with seed yield is that GD can only account for variation in performance due to dominance effects (Bernardo, 1992) or linkage effects. We assume that there was greater heterosis and likely dominance effects for seed yield than for oil concentration. Many authors for various traits in Brassica have reported the relationship of heterosis to genetic distance. Bernardo (1992) indicates that molecular marker heterozygosity would be most useful for predicting hybrid performance in crop species where: (1) dominance effects are strong; (2) heterotic groups are complementary and allele frequencies at individual loci in the parental inbreds are negatively correlated; (3) trait heritability is high; (4) average parental allele frequencies vary only within a narrow range; (5) at least 30-50% of the quantitative trait loci (QTL) are linked to molecular markers; and (6) not more than 20-30 % of the molecular markers are randomly dispersed or unlinked to QTL. In

41

Brassica crops some of these conditions may not be met and current knowledge may not yet be sufficient to establish the usefulness of molecular markers as predictors for the heterosis expression of seed yield and oil content.

CONCLUSION

The present study showed a potential hybrid advantage for seed yield in B. napus hybrids that should provide farmers with an opportunity to improve productivity, particularly in potential high-yield areas and where conventional breeding has apparently reached a yield plateau. The CMS system has been found to be the most effective and practical one for developing brassica hybrids. We also found that only hybrid seed yield was moderately correlated with genetic distances.

ACKNOWLEDGEMENTS

The authors are thankful to Victor Wutor, Lethbridge Research Center, Agri. Food Canada and Janssens Marc, University of Bonn, Germany, for internal technical review of the manuscript. We are also indebted to Doug Walker, Vince D Antonio and Gary, Farm Manager, University of California, at Davis for maintaining plant material in the field and green house. We appreciate Mr. Yasir Khan for his valuable help in formatting this manuscript. We are grateful to two anonymous reviewers for their inputs.

42

REFERENCES

Ali, M, L.O. Copeland and S.G. Elias. 1995. Relationship between genetic distance and heterosis for yield and morphological traits in winter canola (Brassica napus L). Theor. Appl.. Genet. 91:118-121.

Brandle, J.E and P.B.E. Mcvetty.1990. Geographical diversity, parental selection and heterosis in oilseed rape. Canad. J. Plant Sci. 70: 935-940.

Barbosa. N.J.f., M.E. Sorrells and G. Cisar.1996. Prediction of heterosis in wheat using coefficient of parentage and RFLP based estimates of genetic relationships. Genome. 39:1142-1149.

Becker , H., G. M. Engqvist. 1995. The potential of resynthesized rapeseed for hybrid breeding. Proc 9th int’l. rapeseed congr. GCIRC, Cambridge, England. pp.113-115.

Bernardo, R. 1992. Relationship between single-cross performance and molecular marker heterozygosity. Theor. Appl. Genet. 83, 628-634.

Dhillon, B.S., S.S. Banga, B.K. Mangat, Allah-Rang, L.S. Randhawa, T.S. Bharaj and A. Rang. 1996. Hybrid breeding in crop plants. J. Res., Punjab Agric. Univ. 33: 1-4, 1-21.

Diers, B.W., P.B.E. Macvetty and T.C. Osborn.1996. Relationship between heterosis and genetic distance based on restriction fragment length polymorphism markers in oilseed rape (Brassica napus L.). Crop Sci. 36: 79-83.

Duvick D.N., J.S.C. Smith, M. Cooper, 2004. Changes in performance, parentage, and genetic diversity of successful corn hybrids, 1930-2000. pp. 65-97. In: C.W. Smith, J. Betrán, E.C.A. Runge (Eds.), Corn. Origin, History, and Production. John Wiley, Hoboken, N.J.

Fick, G.N. and I.F. Miller. 1997. Sunflower breeding in: sunflower technology and production. Crop Sci. Soc. America, Madison, Wisconsin. pp. 395-439.

Godshalk, E.B., M. Lee and K.R. Lamkey. 1990. Relationship of restriction fragment length polymorphisms to single –cross hybrid performance in maize. Theor. Appl. Genet. 80: 273-280.

Grant, I. and W.D. Beversdorf. 1985. Heterosis and combining ability estimates in spring oilseed rape (Brassica napus L.). Canad. J. Genet. Cytol. 27: 472-478.

Gupta, K.K and K. R. Khanna.1982. Gene action and heterosis for oil yield and component characters in sunflower . Indian J. Genet. Plant Breed. 42: 265-271.

Hammond, E.G. 1991: Organization of rapid analysis of lipids in many individual plants. In: H. F. Linskens, and J. F. Jackson (eds), Modern Methods of Plant Analysis, Vol. 12. Essential Oils and Waxes, 321-330. Springer-Verlag, Berlin.

Jain, A., S. Bathia, S.S. Banga, S. Prakash and M. Laksmikumaran. 1994. Potential use of random amplified polymorphic DNA (RAPD) technique to study the genetic diversity in Indian mustard (Brassica juncea) and its relationship to heterosis. Theor. Appl. Genet. 88: 116-122.

Jones, D.F. 1917. Dominance of linked factors as a means of accounting for heterosis. Genetics 2: 466–479.

Khanna, K.R and C.H. Misra. 1977. Divergence and heterosis in tomato. SABRAO J. Breed. Genet. 9: 43-50.

43

Knaak , C and W. Ecke.1995. Genetic diversity and hybrid performance in European winter oilseed rape (Brassica napus L.). Proc 9th Int’l Rapeseed Congr. GCIRC, Cambridge, England. pp. 110-112.

Lee, M., E.B. Godshalk., K.R. Lamkey and W.W. Wodmar.1989. Association of restriction fragment length polymorphisms among maize inbreds with agronomic performance of their crosses . Crop Sci. 29: 1067-1071.

Lefort-buson, M. and Y. Dattee.1982. Genetic study of some agronomic characters in winter oilseed rape ( Brassica napus L. ). Heterosis. Agronomie. 2: 315-322.

Martin, J.M., L.E. Talbort., S.P. Lanning and N.K. Blake. 1995. Hybrid performance in wheat as related to parental diversity. Crop Sci. 35: 104-108.

Melchinger, A.E., M. Lee., K.R. Lamkey and W.L. Woodman. 1990. Genetic diversity for restriction fragment length polymorphisms: relation to estimated genetic effects in maize inbreds. Crop Sci. 30: 1033-1040.

Moll, R.H., W.S. Salhuana and H.F. Robinson. 1962. Heterosis and genetic diversity in variety crosses of maize. Crop Sci. 2: 197-198.

Nei, M., and W. Li, 1979: Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76, 5256-5273.

Sernyk, J. L. and B.R. Stefansson. 1983. Heterosis in summer rape (Brassica napus L.). Canad. J. Plant Sci. 63: 407-413.

Shull, G.H. 1908. The composition of a field of maize. American Breeders Assoc. Report. 4: 296–301.

Singh, D. and T.R. Mehta. 1954. Studies in breeding of brown sarson I. comparison of F1’s and theirs parents. Indian J Genet. Plant Breed. 14: 74-77.

Smith, O.S., J.S.C. Smith, S.L. Bowen, R.A. Tenborg and S.J. Wall. 1990. Similarities among a group of elite maize inbreds as measured by pedigree F1 grain yield, heterosis and RFLPs. Theor. Appl. Genet. 80, 833-840.

Starmer, K.P., J. Brown and J.B. Davis. 1998. Heterosis in spring Canola hybrids grown in Northern Idaho. Crop Sci. 38, 376-380.

Thakur, H.L. and J.C. Sagwal. 1997. Heterosis and combining ability in rapeseed B. napus. Indian J. Genet. Plant Breed. 57:163-167.

Wright. 1980. In Hybridization Of Crop Plants. Ch. 8: 161-176. Fehr et al., Eds. Zhang. Q., V.J. Qao, M. A.S. Maroof, S.H. Yang and J.X. Li. 1995. Molecular

divergence and hybrid performance in rice. Mol. Breed. 1: 133-142.

44

SABRAO Journal of Breeding and Genetics 43 (1) 44–58, 2011

USING INFORMATION FROM RELATIVES AND PATH ANALYSIS

TO SELECT FOR YIELD AND SEED SIZE IN SOYBEAN (Glycine max L. Merrill)

EDIZON JAMBORMIAS1,2,3*, SURJONO H. SUTJAHJO2,

M. JUSUF3, SUHARSONO3

SUMMARY

Selection of yield and seed size in soybean was carried out by using information from relatives of a S3 selected generation. The research objectives were to determine the selection criteria for yield and seed size, to establish a selection index, and to estimate the selection response based on the information from relatives from an experiment based on pedigree selection on the S3 selected generation of a Slamet × Nakhonsawan cross. The results of this study were: (1) number of filled seeds was a useful selection criterion for yield and seed size; (2) the coefficient of determination of selection index model was low, which was caused by the difference between the phenotypic values of some individuals with poor performance in the families included with the best performance, that was selected through the use of a selection index model based on information from relatives; and (3) selection response of yield was high, but was low for seed size, so that the selection based on yield still needs to be carried out in the next generation, whereas selection based on seed size should not be used.

Key words: hierarchy, variance, covariance, heritability, coheritability, selection.

INTRODUCTION

Soybean (Glycine max L. Merrill) consumed as tofu, tempe, soy sauce and milk is a valuable source of plant protein and an important food source for Indonesians (Nugraha et al., 2000). National productivity is not sufficient to supply the national demand, marked by the increase of the total imports of soybean. Since 2008, Indonesia still imports about 60% of soybean (Prabowo, 2008). This has created the slowing down of soybean farming activity because of the low productivity and quality that is less competitive with imported soybeans. Therefore, some efforts are needed to increase the productivity of soybean.

1 Department of Agronomy, Agriculture Faculty, Pattimura University Ambon, Maluku, Indonesia. 2 Department of Agronomy and Horticulture, Agriculture Faculty, Bogor Agricultural University, Bogor,

Indonesia. 3 Research Center for Bioresources and Biotechnology, Bogor Agricultural University, Bogor, Indonesia. * Corresponding author: [email protected]

45

Genetic improvement through plant breeding is an approach to increase soybean productivity. An important objective is to increase the size of seeds, indicated by the characteristic of new released varieties of soybean that have high productivity and bigger seed size (Suhartina, 2003). To achieve this objective, varieties of Slamet and Nakhonsawan have been crossed since 2003. Slamet is a high yielding Indonesian variety with resistance to aluminum but produces a small seed size (Sunarto, 1995), while Nakhonsawan is a variety from Thailand with the characteristics low yield but large-sized seeds. F2 generation, the filial of F1 generation from crossing two parental lines, is a base population for selection (selected generation S0). A field experiment in this generation showed that high heritability of yield and seed size, as well as phenotypic selection that resulted in significant genetic gains for both traits (Suharsono et al., 2006). Selected generations of S1 and S2 were the selected set representing F3 and F4 generations. Field experiments for the two selected generations also tended to show high heritability (Suharsono et al., 2007). Further phenotypic selection on selected generations S3 and S4, (i.e. selected set of F5 and F6 generations) showed that the yield heritability remained high, but decreased and equalled zero in the two selected generations respectively (Jambormias et al., 2004; 2007). Genetic gain for yield was also high on the S3 selected generation, but the performance of the F6 generation was not significantly correlated with the F5 generation. This indicated the failure of phenotypic selection to select for fixed transgressive segregants at early selected generations (Jambormias et al., 2004; 2007).

Based on the above data, the low percentage of selected lines probably corresponds to the selection of dominant or over-dominant genes in the set of transgressive segregants. The dominance was caused by heterozygosity. High frequency of heterozygosity in each selected generation will be segregating and producing variance for each generation. This causes the heritability and genetic gain to be high, although the performance is almost the same as in the previous generation (Jambormias et al., 2004; Jambormias and Riry, 2009). Therefore, the selection of phenotypes with low selection intensity tends to give low selection efficiency.

Increasing the efficiency of selection can be carried out through a selection approach which involves information from relatives among individuals and interrelationships between quantitative traits (Jambormias and Riry, 2009). A method of analysis that provides better understanding of interrelationship between quantitative traits is path analysis (Bizeti et al., 2004). Wright (1921) proposed this analysis as a method that partitions the estimated correlation into direct and indirect effects of some traits toward an affected trait. It is a result of various growth components and yield production influencing traits.

Path analysis was first carried out in plants by Dewey and Lu (1959) in wheat grass, and by Pandey and Torrie (1973) in soybean. In its development, the results were used to identify selection criteria for yield (Sedghi and Amanpour-Balaneji, 2010). The selection criteria was used subsequently for carrying out the indirect selection on yield (Oz et al., 2009), generating the selection index in direct selection (Wirnas et al., 2006) and developing new varieties based on the concept of "ideal plant types” (Sumarno and Zuraida, 2006).

46

Analysis of relatives among individuals can be made by recording family information from some selection methods. One example is the pedigree selection method (Poehlman and Sleper, 1996). The application of the selection method will result in the population of plants in every selection generation. The population can be partitioned by each selected generation based on its ancestor. Setting the individuals according to the ancestor hierarchical structure results in sets of close relatives of individuals (i.e. families of the investigated generation) (Jambormias and Riry, 2009).

Every phase of the family hierarchy above has genetic parameters, such as variance and covariance or heritability and coheritability. Coheritability is the covariance ratio of genotypic and phenotypic of two traits. If the heritability measures the amount of variance contributions of a trait caused by genetic factors, then coheritability is associated with the same two traits simultaneously. Generally, coheritability is used to estimate the broad sense heritability, correlation and path analysis (Malik et al., 2007).

Interrelationship analysis among traits and relatives information can be used by plant breeders in the future to accelerate the generation of transgressive segregants for multiple traits selection at the beginning of a selected generation with a low level of selection. The inclusion of these two analyses in carrying out selection will optimally take advantage of total genetic information in the population, both covariance of generations and traits (Falconer and Mackay, 1996). In this study, the analysis can be used for carrying out selection on soybean yield and seed size by involving only two traits alone or several other quantitative traits simultaneously. A multiple trait selection approach which involves all the genetic information in a selection index model will increase the efficiency of selection. Simultaneous selection based on index selection will be more efficient compared to the selection based on single character (Soh et al., 1994; Moeljopawiro, 2002).

The simultaneous use of economic value of quantitative characters, information of relationships between traits (covariance) and variance of information from relatives will enable a selection index to be generated that could increase economically valuable quantitative traits. The main soybean breeding objectives in Indonesia – yield and large seed size (Suhartina, 2003) – were selected for investigation in this study. Therefore, yield and seed size were defined as economically valuable traits. Because yield is a priority, then its economic value is set two times larger than the size of seeds, and four times greater than other traits that are not economically valuable.

These aims of this research were to: (1) determine the selection criteria for yield and seed size of soybean in the S3 selected generation derived from a Slamet × Nakhonsawan population; (2) arrange the selection index based on information from relatives in the selection of S3 selected generation of the Slamet × Nakhonsawan population; (3) estimate the ‘direct selection responses’ and ‘correlated selection responses’ of yield and seed size, and their partition in each family hierarchy of S3 selected generation derived from the Slamet × Nakhonsawan population.

47


Experiment Material Field data experiments were obtained from selected generations S3 or F5 generation of Slamet × Nakhonsawan (Jambormias et al., 2004). The genetic materials included in this experiment were 250 families of S3 selected generation and both parents. Measured variables were time of harvest (days, X1), height of plant (cm, X2), number of branch (X3), number of nodes (X4), fertile nodes (X5), number of pods (X6), number of filled pods (X7), number of seeds (X8) and number of filled seeds (X9). Selection variables were seed size (g/100 seeds, Y2) and the yield (g per plant, Y1).

Experiment Procedures These experiments used the pedigree selection method with randomization of rows of the families and selected generation S3 as the experimental unit, which followed a nested design in a completely randomized design. Each row of a family was considered a single experimental unit without replication, except for rows of each parent, which were replicated. The experiment was set up such that the next generation families were nested in the previous generation families. Set of families of selected generation S3 were arranged in 3-stage fully nested design pattern (Montgomery, 2001). In this case, F5 families were nested within the set of F4 families, and families were nested in set of F3 families. The linear additive model for the experiment of the result of the cross and each parent was: selected generation S3:

( )( ) ( ) 53 4 5l ijkijkl i j i k ij FY F F F Wμ= + + + +

for i = 1, 2, …, 70; j =1, 2, …, 74; k = 1, 2, …, 250; and l = 1, 2, …, rk

where: ijklY = measured value of individual F5 to l nested respectively based on a hierarchical family according to F5 to k, F4 to j and F3 to i; μ = mean; F3i = effect

of set of F3 families to i; )(4 ijF = effect of set of F4 families to j nested in set of F3

families to i; )(5 ijkF = effect family F5 to k nested in set of F4 families to j and F3

to k; and ( )5l ijkFW = effect of individuals in family F5 to l nested in family F5 to k, F4

to j and F3 to i. Parents: ijiij TY εμ ++= for i = 1, 2; and j = 1, 2, 3, ..., ri where μ = mean, ijY = measured value of parent to i at replication to j, Ti = effect of

parent to i, and εij = error.

The model for the selected generation was assumed to be random. Estimation of genetic variance-covariance was based on information from relatives, and was performed according to a nested design model, whereas the genetic information of population S3 selected generation without partition based on information from relatives was used to estimate variance-covariance based on individuals.

Generating the matrix of phenotypic and genotypic variance and covariance was done as follows:

48

The partition of the expected mean square [E(MS)] nested design model gives the variance component between families 2

3Fσ , 24Fσ , 2

5Fσ and variance within F5

families 25WFσ , and the E(MS) of completely randomized design including parents

resulted in the variance of within parents ( 2eσ ) (Table 1).

Table 1. Analysis of variance structure and the expected mean squares according

to the general linear model of 3-stages nested in selected generation S3 Source DF MS E(MS) Experiment using the population of selected generation S3 F3 (a-1) MSF3 ls 2

εσ + kd 2Cσ + jc 2

Bσ + ib 2Aσ

F4 (F3) )1(∑ −ib MSF4 ls 2εσ + kd 2

Cσ + jc 2Bσ

F5 (F4,F3) )1(∑ −jc MSF5 ls 2εσ + kd 2

Cσ

Within family F5 ∑ − )1( ls MSWF5 ls 2εσ

Total ∑ −1ls Experiment using parents: Parents (t-1) MSTetua 2

εσ + ir2τσ

Intraparents )1( −Σ ir MSE 2εσ

Total 1−Σ ir

Estimation of between family variance in each family hierarchy (i.e. F3, F4, and F5) and within family F5 and environment was based on the following equations:

2 3 43ˆ F F

Fi

MS MSb

σ−

=

2 4 54ˆ F F

Fj

MS MSc

σ −=

2 5 55ˆ F WF

Fk

MS MSd

σ−

=

2 55ˆ WF E

WFl

MS MSs

σ−

=

2ˆEσ = MSE

Estimation of additive genetic variance was obtained through the

information from relatives derived from Falconer and Mackay (1996), where the

49

correlation of the breeding value of a self pollinated crop was assumed to be equal with twins breeding value correlation (i.e. equal to unit). This derivation gave the variance component between families as additive genetic variance for families

2ˆ( )iAFσ and the variance within families minus the variance within parents gave the

additive variance within families 2ˆ( )iAWFσ . The phenotypic variance ( 2ˆPσ ) was the

sum of between families, within F5 family and within-parental variances, respectively. The phenotypic and genotypic covariances were calculated in the same line as the calculation of the above variances, i.e. by partitioning the expected mean cross product [E(MCP)] model with the structure of analysis of covariance in the same line with the structure of analysis of variance in Table 1.

Choosing selection criteria Generating the selection index for yield and seed size was determined by the best regression model using stepwise regression, followed by path analysis (Singh and Chaudhary, 1979). The quantitative traits selected as selection criteria were the k of the p quantitative traits showing the similar direction of phenotypic and genotypic path coefficients based on individual and information from relatives with the phenotypic regression coefficient based on the individual information (having the same sign, either positive or negative).

Generation of the selection index The selection index was generated by using the procedure by Becker (1975) and Falconer and Mackay (1996) with a minor modification. Generally, the selection index is written in a matrix equation: I = b’Y, with the selection index variance ( 2

Iσ ) and genotypic variance ( 2Hσ ) from the matrix equation: 2

Iσ = b’Pb and 2Hσ

= a’Ga (Falconer and Mackay 1996; Moeljopawiro 2002), where: P = matrix of phenotype variance and covariance, G = matrix of genotype variance and covariance, b = index value vector, derived from the manipulation of algebra matrix equation b = P-1Ga, and a = economic value vector, its elements e are stated 4 : 2 for the yield and seed size, and 1 for the selection criteria if the phenotypic and genotypic path coefficient value is positive, and zero if the value is negative.

The rank of selection index as a model is measured by the coefficient of

determination, by equation 2HIR = 2

2

H

I

σσ

100 (Moeljopawiro, 2002).The higher the

coefficient of determination, the more accurate the estimation of selection index model.

The elements of vector and matrix needed for generating of the selection

index were based on the information from relatives on yield and seed size as follows: b = (bY2F3 bY1F3 bXiF3 ... bXkF3 bY2F4 bY1F4 bXiF4 ... bkF4 bY2F5 bY1F5 bXiF5 ... bXkF5 bY2WF5 bY1WF5 bXiWF5 ... bXkWF5 )’

50

( )2 4 2 4 2 4 2 4a 'i k i k i k i k

x x x x x x x xe e e e e e e e= L L L L

where e = economically weighted selection criteria.

4 4

Pk k×

=P Σ =

0 0 0

0 0 0

0 0 0

0 0 0

Pk k

Pk k

Pk k

Pk k

×

×

×

×

⎛ ⎞⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎝ ⎠

Σ

Σ

Σ

Σ

4 4

Gk k×

=G Σ =

3

4

5

5

0 0 0

0 0 0

0 0 0

0 0 0

Fk k

Fk k

Fk k

WFk k

×

×

×

×

⎛ ⎞⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎜ ⎟⎝ ⎠

Σ

Σ

Σ

Σ

where P and G are the phenotypic and genotypic covariance matrices based on the information from relatives. Elements of matrix P are the null matrix with (k × k) dimensions and phenotypic covariance matrix ( )P

k k×Σ , while elements of matrix G

were the null matrix with (k × k) dimensions, genotypic covariance matrix between families ( )Fi

k k×Σ and within families 5( )WF

k k×Σ . Elements of matrices P

k k×Σ , Fi

k k×Σ and

5WFk k×Σ are the variance in the main diagonal and covariance in the non main

diagonal of matrix of variables Y2, Y1, Xi … Xk of traits, i.e. yield, seed size, and the traits of the selection criteria.

Estimation of selection responses Estimation of correlated selection responses of yield (Y1) and seed size (Y2) traits for

the selection index model is found in equation CRYi = 1G II

i σσ

(Falconer and

Mackay, 1996), where i = selection intensity, in this experiment it was 2.1 (percentage of 5%); 2

Iσ = b’Pb, for b = weighted selection index coefficient vector,

and P = matrix of phenotype variance and covariance, 1G Iσ is genetic variance and

covariance of trait i, for 1G Iσ = b’G*, and G* = matrix of phenotype variance and

51

covariance with the column element i’ not relevant with trait i valued zero. Selection responses were components of direct selection response of trait Yi and the correlated responses with other quantitative traits through trait Yi.

RESULTS AND DISCUSSION

Choosing the selection criteria of yield and seed size based on information from relatives The stepwise regression analysis based on individual traits of yield and seed size resulted in the equation: yield: Y1 = -3.925 + 0.109 X9 + 0.285 Y2 - 0.0412 X5 + 0.0233X7 + 0.00968 X1 – 0.00711 X6 (R2 = 96.9%) and seed size: Y2 = 7.033 + 1.85 Y1 - 0.223 X9 + 0.0728 X4 + 0.0543 X6 + 0.03 X1 - 0.05 X7 + 0.065 X5 (R2 = 57.5%). Harvest time, number of fertile nodes, number of pods, number of filled pods, number of filled seeds and seed size are quantitative traits that are related to the yield. Meanwhile, seed size has a relationship with quantitative traits or variables including harvest time, number of nodes, number of fertile nodes, number of pods, number of filled pods, number of filled seeds, and yield.

The quantitative traits showed a functional relationship with yield and seed size according to stepwise regression analysis based on information from selected individuals as candidate selection criteria. It was then interpreted by using phenotypic and genotypic path coefficients based on information from populations and relatives. The candidate selection criteria which showed similarities of direction between the regression coefficient with each phenotypic and genotypic path coefficient were chosen as selection criteria (Figures 1 and 2). The results showed that only number of filled seeds (X9) showed co-directional consistency of path coefficients with yield and seed size traits, at the level of individuals and information from relatives. Therefore, only this trait was chosen for the selection criteria.

52

0.01Phenotypic

-0.33Genotypic

0.06Phenotypic

0.33Genotypic

0.08Phenotypic-0.06Genotypic

0.90Phenotypic

1.03Genotypic

Figure 1. Direct effect of phenotypic and genotypic information based on individual information of some quantitative traits of yield and seed size. The arrows indicated direction of the direct effect of each variable, while the subscript values indicated the phenotypic and genotypic path coefficient of variables.

53

Y1

Y2

X1

X4

X5

X6

X7

X9

0.06F3

-0.01F4

1.54WF5

0.003F5

-0.45F3

0.03F4

0.08F5

0.69WF5

0.43F3

0.05F4

0.30F5

0.16WF5

-0.02F3 0.05F4 -0.04F5 2.21WF5

-0.49F3-0.15F4

-0.196F50.10WF5

0.85F3

0.34F4

0.146F5

-3.73WF5

0.60F3

0.71F41.018F5

3.17WF5

0.17F30.

29F4

0.205

F5

-1.71W

F5

-0.42F3

-0.24F4-0.21F5

1.12 WF5

3.05F 3 1.38F4 0.78F5 0.43WF 5-4.17F3 -2.25F4

-0.28F5-1.20WF5

-3.96F3

-1.85F4

-3.89F5

-1.21WF5

5.59F 33.

03F4

3.19F5

2.08

WF5

Figure 2. The path coefficient of genotypic information based on information from relatives of some quantitative traits for yield and seed size. The arrows indicated the direction of the direct effect of each variable, while the values of the subscript Fi indicated the genotypic path coefficient of variables of each family or set of family hierarchy and within families.

The interrelationship between seed size and yield in this research is different from what has been found by Iqbal et al. (2003), Bizeti et al. (2004) and Wirnas et al. (2006). They showed that there was no correlation between the seed size and yield. Similarly, there was a negative correlation between plant height and yield as reported by Shrivastava et al. (2001), however the result of this research showed that there was no correlation between the two traits, although it showed a similar positive correlation between days of harvest and yield. The same result was also reported by Bizeti et al. (2004) who reported a positive correlation between the number of total nodes and yield.

54

Generation of the selection index model and simultaneous selection based on the information from relatives

The selection index model based on the information from relatives used the value of economic weighting for the number of filled seeds was null. This was caused by the negative similarity values between regression and path coefficient of seed size. The equation of the selection index involving such economic weight vectors is as follows:

I = 0.04Y2.F3 + 0.23 Y2.F4 + 0.26 Y2.F5 + 0.08 Y2.WF5 + 1.29 Y1.F3 + 0.69 Y1.F4 + 0.87 Y1.F5 + 0.62 Y1.WF5 – 0.05 X9.F3 + 0.08 X9.F4 + 0.10 X9.F5 + 0.08 X9.WF5

(R2 = 69.38).

The coefficient of determination of the selection index model of 69.38% showed the low precision of this selection index model. This tendency occurred as a result of the contribution of using hierarchical information from relatives. The estimated values came from the contribution of family means; so that it was far from the phenotypic values. Furthermore, according to the matrix equation b = P-1Ga, it contains the contribution of the heritability, coheritability and economic weight values. The traits in the selection are the traits with low heritability and coheritability which yielded the low coefficient of determination. Nevertheless, the inclusion of hierarchy information in the selection index model used additive effects in the between family effect compared to the within family which is affected by the non-additive genes and the environment (Falconer and Mackay, 1996).

The selection index model showed the high contribution of yield for family hierarchy when compared with seed size and number of filled seeds. The largest contribution was from the F3 family hierarchy, followed by a hierarchy of family F5, F4 and within-family F5. This large contribution can be understood as a consequence of economical weighting of yield. It may also have been due to the effect of gene action in the selection index weighting matrix equation (b = P-1Ga), where the heritability of the yield of 0.93 was higher than the seed size of only 0.49 (Jambormias et al., 2004). Theoretically, based on the matrix equation above, the multiplication between elements of the matrix P-1G can produce joint heritability. The joint heritability represents the accumulation of heritability on the main diagonal and co-heritability on multiplication of rows and columns that correspond to the main diagonal elements.

Seed size was ranked second in contribution of gene action effect and economic value on the selection index model, especially the between-family hierarchy F4 and F5, meanwhile the number of seeds per plant gave a small contribution. The lower economic weight of these two traits compared to the economic weight value of yield resulted in the small contribution for seed size. Even the trait number of seeds per plant contributes only through its coheritability effect to the yield and the seed size according to the matrix b = P-1Ga because its economic weight value was zero. This condition is ideal to maintain the high yield and seed size that assure high productivity.

55

Yield and seed size in the selected generation, according to the selection index model with selection percentage of 5%, were 14.87 and 13.32 g, respectively (Table 2). The standard deviation of yield was higher than seed size, which ranged between 4.21 to 36.06 g and 2.61 to 30.07 g for seed size. The low minimum values show the existence of individuals with undesirable phenotypes. This implies that an individual with a low phenotypic value which comes from the F3, F4 and F5 family with high mean value tend to have high breeding values. The low value of individual phenotypes may come from environmental effects. For seed size for example, Mursito (2003) stated that the maximum seed size is determined genetically, however the actual seed size formed depends on the environment during seed filling. Therefore, undesirable phenotypes are not entirely derived from genetic effects, but also due to environmental effects.

Table 2. Statistics of selected generation yield and seed size of the selection index model based on the information from relatives selected generation F5

Traits Mean Standard deviation

Minimum value

Maximum value

Yield (g) 14.87 4.21 4.87 36.08 Seed size (g) 13.32 2.61 7.34 30.07

Analysis of selection responses show that there is still a tendency of high

selection response of yield, whereas seed size decreases (Table 3). The high response of yield came from the correlated response number of filled seeds trait considering family hierarchy. The highest response came from relatives of F3 and F5 families.

Direct and correlated selection responses of the yield and seed size traits based on information from relatives

Selection response is the change of expected population mean in the next generation as a result of selection applied in the previous generation (Wei et al., 1996). Response to selection incorporating two correlated traits consisted of two types: direct response and correlated response (Falconer and Mackay, 1996). Direct selection response is determined by the selection intensity, the economic and heritability value, meanwhile the correlated selection response is determined by the selection intensity, heritability, coheritability, and economic value on the phenotype value.

56

Table 3. Total direct and correlated responses of yield and seed size from the selection index model based on the information from relatives

Yield Selection response (gram)

Seed size Selection response (gram)

Total Direct F3 Direct F4 Direct F5 Direct Within family F5 Correlated Y2→Y1.F3 Correlated Y2→Y1.F4 Correlated Y2→Y1.F5 Correlated Y2→Y1.WF5 Correlated X9→Y1.F3 Correlated X9→Y1.F4 Correlated X9→Y1.F5 Correlated X9→Y1.WF5

17.861 1.141 0.066 0.678 0.070 0.062 0.016 0.133 0.012 9.038 0.547 5.521 0.577

Total Direct F3 Direct F4 Direct F5 Direct within family F5 Correlated Y1→Y2.F3 Correlated Y1→Y2.F4 Correlated Y1→Y2.F5 Correlated Y1→Y2.WF5 Correlated X9→Y2.F3 Correlated X9→Y2.F4 Correlated X9→Y2.F5 Correlated X9→Y2.WF5

0.682 -0.054 0.091 0.121 0.005 -0.010 0.039 0.074 0.005 0.057 0.050 0.276 0.028

Note: Y1 = yield, Y2 = seed size, and X9 = number of filled seeds; WF5 = within family F5.

The high correlated selection response of number of filled seeds to yield due to the high covariance or coheritability between yield and number of filled seeds for family hierarchy showed a high response. This condition is probably due to additive gene action and linkage between the two traits. Therefore, further selection is needed for the next generation to increase the yield, especially by focusing on information from relatives of family F3 or F5. Carrying out of family selection with respect to information from relatives is a form of selection to quickly get transgressive segregants at earlier selected generations (Jambormias and Riry, 2009).

On the other hand, the low seed size selection response in almost each level of hierarchy indicates the small additive gene action in increasing the seed size. This suggests that since there is no additive gene action, further selection to increase the seed size would not be successful.

CONCLUSIONS AND SUGGESTIONS

The conclusions drawn from the results of this research are as follows: 1. The number of filled seeds was the only quantitative traits chosen as selection

criteria among the quantitative traits, because it had consistency of similar direction in the interrelationship based on individual stepwise regression analysis, which was based on individual information from path coefficients of individual phenotype and genotype (based on the individual and information from relatives at all family hierarchy levels).

57

2. Generating the selection index model incorporating information from relatives resulted in a low coefficient of determination (R2 = 69.38%). This was due to the large differences between selection index based on phenotypic values of certain individuals with poor phenotypic performance, which are included in the families with good performance.

3. Selection response of yield is categorized as high, mainly derived from the contribution of correlated responses of number of filled seeds through hierarchy family F3 and F5. The additive gene action probably existed in the F3 and F5 families, which indicated family selection through the two hierarchy families. On the other hand, response of total seed size was low, which indicates the lower contribution of additive genes to increase seed size, therefore the seed size selection can be stopped at the F6 stage.

In order to obtain earlier transgressive segregants, it is necessary to check for genotypes which are pure lines, and augmented experimental designs. Performance of check genotypes are used to indicate the value of selection of individual phenotypes with the best performance. The variance of check genotypes is also used to determine the environmental variance, in selecting the best performing individuals.

REFERENCES

Becker, W.A. 1975. Manual of Quantitative Genetics. 3th Ed. Washington State University Press, Washington (pp. 144-145).

Bizeti, H.S., C.G.P. de Carvalho, J. Souza and D. Destro. 2004. Path analysis under multicollinearity in soybean. Braz. Arch. Biol. Tech. 47: 669-676.

Dewey, D.R. and K.H. Lu. 1959. A correlation and path coefficient analysis of components crested wheatgrass seed production. Agron. J. 51: 515-518.

Falconer, D.S. and T.C.F. Mackay. 1996. Introduction to Quantitative Genetics. 4th Ed. Adison-Wesley Longman, Harlow, United Kingdom (pp. 228-240 and 312-325).

Iqbal, S., M. Ariq, M. Tahira, M. Ali, M. Anwar and M. Sarwar. 2003. Path coefficient analysis in different genotypes of soybean (Glycine max (L.)) Merr). Pak. J. Biol. Sci. 6: 1085-1087.

Jambormias, E., S.H. Sutjahjo, M. Jusuf and Suharsono. 2004. Keragaan, keragaman genetik dan heritabilitas sebelas sifat kuantitatif kedelai (Glycine max L. Merrill) pada Generasi Seleksi F5 zuariat persilangan varietas Slamet × Nakhonsawan. J. Pertanian Kepulauan. 3: 115-124.

Jambormias, E., S.H. Sutjahjo, M. Jusuf and Suharsono. 2007. Keragaan dan keragaman genetik sifat-sifat kuantitatif kedelai (Glycine max L. Merrill) pada Generasi Seleksi F6 persilangan varietas Slamet × Nakhonsawan. Bul. Agron. pp. 168-175.

Jambormias, E. and J. Riry. 2009. Penyesuaian data dan penggunaan informasi kekerabatan untuk mendeteksi segregan transgresif sifat kuantitatif pada tanaman menyerbuk sendiri (suatu pendekatan dalam seleksi). J. Budidaya Pertanian. 5: 11-18.

58

Malik, F.A.M., M. Ashraf, A.S. Qureshi and A. Ghafoor. 2007. Assessment of genetic variability, correlation and path analyses for yield and its components in soybean. Pak. J. Bot. 39: 405-413.

Moeljopawiro, S. 2002. Optimizing selection for yield using selection index. Zuriat. 13: 35-42.

Montgomery, D.C. 2001. Design and Analysis of Experiments. 5th Ed. John Wiley & Sons Inc., New York, USA (pp. 566-567).

Mursito, D. 2003. Heritabilitas dan sidik lintas karakter fenotipik beberapa galur kedelai (Glycine max (L.) Merrill)). Agrosains. 6: 58-63.

Nugraha, S.U., S.D. Djoko and S. Widiarti. 2000. Pengembangan mutu kedelai untuk agroindustri. Prosiding Lokakarya Penelitian dan Pengembangan Mutu Kedelai di Indonesia. BPPT. Jakarta 6-7 Agustus 1996.

Oz, M., A. Karasu, A.T. Goksoy and Z.M. Turan. 2009. Interrelationships of agronomical characteristics in soybean (Glycine max) grown in different environments. Int. J. Agric. Biol. 11: 85-88.

Pandey, J.P. and J.H. Torrie. 1973. Path coefficient analysis of seed yield components in soybeans (Glycine max (L) Merrill). Crop Sci. 13: 505-507.

Poehlman, J.M., and D.A. Sleper. 1996. Breeding Field Crops. 4th Ed. Iowa State University Press, Iowa, USA. 165p.

Prabowo, E.H. 2008. Kedelai: Komoditas Yang Salah Urus. http://kompas.com/kompas-cetak/0801/16/ekonomi/4168629.htm. Accessed January 25, 2008.

Sedghi, M. and B. Amanpour-Balaneji. 2010. Sequential path model for grain yield in soybean. Not. Sci. Biol. 2: 104-109.

Shrivastava, M.K., R.S. Sukla and P.K. Jain. 2001. Path coefficient analysis in diverse genotype of soybean (Glycine max L.). Plant Sci. 4: 47-51.

Singh R.K., and B.D. Chaudhary. 1979. Biometrical Methods in Quantitative Genetic Analysis. Kalyani Publishers, Ludhiana, India. 73p.

Soh, A.C., C.S. Chow, S. Iyama, and Y. Yamada. 1994. Candidate traits for index selection in choice of oil palm ortets for clonal propagation. Euphytica. 79: 23-32.

Suhartina. 2003. Perkembangan dan Deskripsi Varietas Unggul Kedelai 1918 – 2002. Badan Penelitian dan Pengembangan Pertanian. Pusat Penelitian dan Pengembangan Tanaman Pangan, Balai Penelitian Tanaman Kacang-kacangan dan Umbi-umbian, Indonesia. 2p.

Suharsono, M. Jusuf and A.P. Paserang, 2006. Analisis ragam, heritabilitas dan pendugaan kemajuan seleksi populasi F2 dari persilangan kedelai kultivar Slamet × Nokonsawon. J. Tanaman Tropika. 9: 86-94.

Suharsono, M. Jusuf and Dasumiati. 2007. Analisis ragam dan seleksi populasi F3 dari persilangan kedelai kultivar Slamet × Nokonsawon. J. Tanaman Tropika. 10: 21-28.

Sumarno, and N. Zuraida. 2006. Hubungan korelatif dan kausatif antara komponen hasil dengan hasil kedelai. J. Penelitian Pertanian Tanaman Pangan. 25: 38-44.

Wei, M., A. Caballero, and W.G. Hill. 1996. Selection response in finite populations. Genetics 144: 1961-1974.

Wirnas, D., I. Widodo, Sobir, Trikoesoemaningtyas and D. Sopandie. 2006. Pemilihan karakter agronomi untuk menyusun indeks seleksi pada 11 populasi Generasi F6. Bul. Agron. 34: 19-24.

Wright, S. 1921. Correlation and causation. J. Agric. Res. 20: 557-585.

59


COMBINING ABILITY FOR OLEIC ACID IN PEANUT

(Arachis hypogaea L.)

N. SINGKHAM1, S. JOGLOY1*, T. KESMALA1, P. SWATSITANG2, P. JAISIL1, N. PUPPALA3 AND A. PATANOTHAI1

SUMMARY

Elevated oleic acid concentration increases the quality and shelf-life of peanut and their derived products. Incorporation of this character is an important objective of most peanut breeding programs. The objectives of this study were to examine the general combining ability (GCA) and specific combining ability (SCA) effects for oleic acid concentration and identify promising crosses for developing high oleic peanut varieties. Twenty crosses in the F2 and F3 generations from a full diallel matting of five parental genotypes and their parents were evaluated under field conditions in a randomized complete block design with four replications for two seasons along with the original parental lines. Seed samples were analyzed for fatty acid compositions by gas liquid chromatography. The GCA effects were significant for oleic acid and O/L ratio in both F2 and F3 generations. The SCA and reciprocal effects were also significant, but their relative contributions to variation among crosses were much smaller than those of the GCA effects. The results suggested that additive gene action was important in the inheritance of oleic acid concentration, and selection for high-oleic acid should be effective. The crosses of the line [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 with SunOleic 97R and Georgia-02C are most promising to exploit genetic variation for oleic acid, which in this population appears to be beyond the genetic control of the ol genes.

Key Words: Peanut breeding, fatty acids, general combining ability, inheritance, seed quality, specific combining ability

1 Department of Plant Science and Agricultural Resources, Faculty of Agriculture, Khon Kaen University, Khon Kaen, 40002, Thailand

2 Department of Biochemistry, Faculty of Science, Khon Kaen University, Khon Kaen, 40002, Thailand 3 Agricultural Science Center at Clovis, New Mexico State University, Clovis, New Mexico, 88101, USA * Corresponding author: Phone: +66 43 364 367. Fax: +66 43 364 367. Email: [email protected]

60

INTRODUCTION

Peanut (Arachis hypogaea L.) is an important oil-bearing crop and the kernels contain 50% oil. Oleic and linoleic acids are the primary fatty acids in peanut oil accounting for 80% of fatty acid composition (Andersen and Gorbet, 2002). Oleic acid was positively correlated with the ratio of oleic to linoleic acids (O/L ratio), and it was negatively associated with linoleic acid (Dwivedi et al., 1993).

Maintaining quality is a concern for peanut products worldwide. Low quality and short storability are generally associated with rancidity of the kernels. High-oleic acid peanut has much longer shelf life and flavor stability than normal-oleic peanut (O’Keefe et al., 1993; Mugendi et al., 1998). In addition, iodine value (IV) provides a good measurement of the degree of oil unsaturation and stability (Andersen and Gorbet, 2002). High oleic oil may provide health benefits since it has been associated with lowered blood serum cholesterol especially with low density lipoproteins (LDL) in humans (O’Byrne et al., 1997).

Most studies supported a two-gene model of qualitative inheritance of oleic acid controlled by two recessive genes (ol1 and ol2) (Moore and Knauft, 1989; Isleib et al., 1996; Lόpez et al., 2001). In the case of qualitative inheritance, breeding for high oleic acid would be relatively easy.

However, quantitative inheritance of this character has also been reported in normal oleic acid peanut germplasm (Upadhyaya and Nigam, 1999). Upadhyaya and Nigam (1999) reported that additive × additive epistasis was detected for oleic acid concentration in peanut. Mercer et al. (1990) found that general combining ability (GCA) was detected for oleic acid in the F1 and F2 generations and maternal effects were significant in the F1 generation but dissipated in the F2 generation. In the case of quantitative inheritance, breeding for this character would be complex and difficult, particularly, if the character has low heritability.

The previous results were dependent on materials used (with or without high oleic acid ol genes) and methods of studies (qualitative or quantitative), and genetic control of oleic acid may be beyond the two major genes with large effect (Isleib et al., 2006a). The O/L ratio in a segregating population with high oleic acid showed continuous variation ranging from 2.2 to 25.4 (Lόpez et al., 2001). In addition to ol1 and ol2, oleic acid concentration may be influenced by environmental and additional genetic factors might have modified the expression of this character (Isleib et al., 2006a), causing the progenies described by Lόpez et al. (2001) to segregated continuously.

High oleic acid can improve peanut oil quality similar to that of olive oil, and is an important objective of most breeding programs. Since the discovery of high oleic acid in peanut, Georgia-02C (Branch, 2003) and SunOleic 97R (Gorbet and Knauft, 2000) were developed and used as parents in many peanut breeding programs to transfer the high oleic character. However, a breeding line ([(NC17090 × B1)-9-1 × KK 60-3]F6-8-3) developed at Khon Kaen University,has intermediate oleic acid concentration, and it may carry either ol1 or ol2 or another oleic modifying gene.

61

The objectives of this study were to determine GCA and SCA effects for oleic acid concentration in peanut crosses, and to determine the effect of an intermediate oleic genotype when crossed with high oleic genotypes and normal oleic genotypes.


Plant materials Five peanut genotypes differing in oleic acid concentration were used in the crossing program to produce F1 hybrids. They included SunOleic 97R, Georgia-02C, [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3, KKU 1 and KK 5. SunOleic 97R and Georgia-02C are high oleic varieties which are known to carry two recessive genes, ol1 and ol2 (Chu et al., 2009). [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 is a breeding line developed at Khon Kaen University and genetic control of oleic acid in this peanut genotype has not been studied. KKU 1 and KK 5 are local varieties with normal oleic acid concentration. These five genotypes were crossed in a full diallel mating design to produce 20 F1 hybrids. Field experiment for the F2 generation Twelve seeds for each of 20 F1 hybrids and five parental lines were planted in the rainy season 2008 at the Field Crop Research Station of Khon Kaen University (KKU) in Northeast Thailand (16°26′N, 102°50′E,190 masl) for producing F2 seeds for fatty acid evaluation and generation advance. The experiment was arranged in a randomized complete block design with four replications. Plots consisted of two-rows, 1.2 m in length with spacing of 50 cm between rows and 20 cm between plants within a row. The variety Kalasin 2 was planted to provide a border plant at the end of each row to demarcate plots. Routine cultural practices were followed to grow the plots (Singkham et al., 2010). Supplementary irrigation was given during the dry periods in the rainy season with an overhead sprinkler system.

The F1 plants for each cross and the parental lines were harvested at maturity. Border plants in each plot were discarded, and ten F1 plants form each plot were harvested manually. The seeds of parental lines were bulked for each plot, while the F2 seeds from each F1 plant in the plot were bulked and divided into two sets. The first set was analyzed for fatty acids and the other set was used for planting plots to produce the F3 generation. Field experiment for the F3 generation The remnant F2 seeds of 20 crosses and seeds of parental lines were planted in the dry season (2008/09) at the Field Crop Research Station of KKU. A randomized complete block design with four replications was used. Seeds were planted in two-row plots 1.2 m long with spacing of 50 cm between rows and 20 cm between plants within a row. The variety Kalasin 2 was used to plant border plants at the ends of the rows. Field crop management was similar to that in the F2 experiment. Ten plants in each plot were harvested at maturity. The seeds from each plot were bulked for fatty acid analysis.

62

Fatty acid analysis Fifty mature kernels for each sample in the F2 and F3 generations were bulked for the determination of oil concentration and fatty acid compositions. The samples were ground and oven-dried at 70 °C for about 15 to 20 h. Moisture concentration was then determined by weight difference between sample prior to oven-drying and after oven-drying. Oil was extracted from seeds by the Soxtec extractor (50 mL of petroleum ether was used as a solvent).

Percentage of oil = oil weight (g) × 100 ground seed weight (g)

The extracted oil was analyzed for fatty acid concentration by gas liquid chromatography (GLC). The protocol of fatty acid analysis was modified from Bannon et al. (1982). Fatty acid methyl esters (FAME) were prepared by adding 1 mL of 2.5% H2SO4/MeOH in 10 mg of oil sample and 100 μL of 0.01 g/mL C17:0 an internal standard. The mixture was incubated at 80 °C for 2 h. After incubation, 200 μL of 0.9% (w/v) NaCl and 200 μL heptane were added to the mixture and mixed well. The FAME was extracted into heptane. The concentration of oil sample was 33 μg, which was dissolved in a 1 μL of FAME. The FAME sample (2 μL) was injected to GLC (with Flame Ionization Detector: FID) for fatty acids analysis. Fatty acid analysis was conducted on Shimadzu Gas Chromatograph GC – 14B – CR7A and SGE fort GC capillary column (30 m × 0.25 mm ID BPX70 0.25 μm) was used. Helium was the carrier gas at a flow rate of 30 mL/min. Hydrogen and air were used at the rate of 30 and 300 ml/min, respectively for the ignition of the FID. Oven temperature was maintained at 130 °C for 2 min. Then it was programmed at 5

°C/min to 220 °C and held at this temperature for 8 min. The injector temperature and detector temperature were 250 °C and 300 °C, respectively. The standard fatty acids that were used to identify the fatty acid concentration in peanut varieties consisted of myristic, palmitic, stearic, oleic, linoleic, linolenic, arachidic, eicosenoic, behenic, erucic and lignoceric acids. Fatty acids are reported here in concentration (percentage as related to total fatty acids).

O/L ratio, IV and the ratio of unsaturated to saturated fatty acids (U/S ratio) (Singkham et al., 2010) were computed as below:

O/L ratio = % oleic acid / % linoleic acid, IV = (% oleic acid × 0.8601) + (% linoleic acid × 1.7321) + (% eicosenoic

acid × 0.7854), U/S ratio = (% oleic acid + % linoleic acid + % eicosenoic acid) / (% palmitic

acid + % stearic acid + % arachidic acid + % behenic acid + % lignoceric acid). Statistical analysis Analysis of variance for each character was performed according to the procedures by Hoshmand (2006). When the differences were statistically significant, Duncan’s multiple range test (DMRT) was used to compare mean differences. Estimates of combining ability in the F2 and F3 generations were computed by using Method 1 Model 2 of Griffing (1956). The relative importance of GCA and SCA effects was calculated as GCA/SCA mean squares. Tests for significance of GCA and SCA effects were done using the t-test. Simple correlation was used to determine the relationship between parental line performance and their progenies.

63

RESULTS

Parental means for fatty acid composition and oil characters Parents were significantly different for most characters in both seasons (rainy 2008 and dry 2008/09) except for % oil in the rainy season (Table 1). The parental means for oleic acid concentration varied from 45.3 to 80.5% of total fatty acid. Georgia-02C and SunOleic 97R had the highest oleic, eicosenoic, lignoceric acid, O/L ratio and U/S ratio in both seasons. Georgia-02C and SunOleic 97R also had low palmitic, stearic, linoleic, arachidic acids and low IV in both seasons. KK 5 and KKU 1 had relatively low oleic acid and O/L ratio, and they had higher linoleic acid in both seasons. [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 had intermediate-oleic acid in both the rainy and dry seasons (71.6 and 67.3%, respectively). Combining ability for fatty acid composition and other oil characters GCA was significant for all characters in both the F2 and F3 generations (Table 2). SCA was also significant for palmitic, stearic, oleic, and linoleic acids, as well as % oil, O/L ratio, IV and U/S ratio in both F2 and F3 generations. The reciprocal effects were significant for palmitic, oleic, linoleic acids, % oil, O/L ratio and IV in both the F2 and F3 generations. The ratios of GCA/SCA mean squares for oleic acid and linoleic acid were high in both the F2 and F3 generations, and, therefore, GCA contributed greater to the variation in these characters than did SCA.

SunOleic 97R, Georgia-02C and [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 had high GCA effects for oleic acid (6.41, 7.20 and 3.16, respectively) in the F2 generation and also in the F3 generation (8.02, 7.72 and 2.16, respectively) (Table 3). In contrast to oleic acid, these genotypes had low GCA effect for linoleic acid (Table 3). For O/L ratio, GCA effects were high for SunOleic 97R and Georgia-02C in both the F2 and F3 generations. The correlation coefficients between fatty acid composition of the parents and their progenies were positive and significant, ranging from 0.39 to 0.59 except % oil in both the F2 and F3 generations.

For SCA effects in the F2 generation, SunOleic 97R × Georgia-02C had the highest SCA effects for oleic acid (2.16) and O/L ratio (5.30), whereas KKU 1 × Georgia-02C had the lowest SCA effect for oleic acid (–4.13) but it had the highest SCA effects for linoleic acid (3.75) and IV (2.80) (Table 4). For SCA effects in the F3 generation, SunOleic 97R × Georgia-02C had the highest for oleic acid (3.47) and for O/L ratio (8.40), but it had the lowest SCA for linoleic acid (–3.68) and IV (–3.39) (Table 5). KK 5 × SunOleic 97R had the highest SCA effect for linoleic acid (8.54) but it had the lowest SCA effect for oleic acid (–3.68).

The GCA effect of intermediate oleic genotype,( (NC 17090 x B1)-9-1 x KK 60-3)F6-8-3 was positive and significant ranging from 2.16 to 3.16 for olelic acid in F2 and F3(Table 3). The SCA effects of intermediate oleic genotype ([(NC17090 × B1)-9-1 × KK 60-3]F6-8-3) with high oleic genotypes (SunOleic 97R and Georgia-02C) were both negative for oleic acid concentration, and the crosses with normal oleic genotypes (KK 5 and KKU 1) also exhibited negative for oleic acid concentration in both F2 and F3 generations (Table 4 and 5).

64

Tab

le 1

. M

ean

of fa

tty a

cid

conc

entra

tion

(% o

f tot

al fa

tty a

cid)

, % o

il, th

e ra

tio o

f ole

ic to

lino

leic

aci

ds (O

/L ra

tio),

iodi

ne v

alue

(IV

), un

satu

rate

d to

satu

rate

d fa

tty a

cids

ratio

(U/S

ratio

) of p

aren

tal l

ines

in th

e ra

iny

seas

on 2

008

and

the

dry

seas

on 2

008/

09 in

Tha

iland

.

Pare

ntal

line

Pa

lmiti

c ac

id

Stea

ric

acid

O

leic

ac

id

Lino

leic

ac

id

Ara

chid

ic

acid

Ei

cose

noic

ac

id

Beh

enic

aci

d Li

gnoc

eric

ac

id

% o

il O

/L ra

tio

IV

U/S

ratio

R

ainy

seas

on 2

008

SunO

leic

97R

6.

3 d

3.1

d 78

.7 a

4.

5 d

1.4

d 1.

8 a

2.8

b 1.

7 a

47.2

18

.0 b

76

.8 d

5.

5 a

Geo

rgia

-02C

6.

5 d

2.8

d 79

.8 a

3.

1 e

1.3

d 1.

8 a

2.7

b 1.

7 a

46.5

27

.1 a

75

.4 d

5.

7 a

F6-8

-3a

7.3

c 4.

5 b

71

.6 b

11

.0 c

1.

9 b

1.0

b 2.

7 b

1.1

b 45

.0

6.6

c 81

.3 c

4.

8 b

KK

5

11.5

a

4.0

c 46

.9 c

31

.0 a

1.

7 c

0.7

c 3.

0 ab

1.

3 b

47.6

1.

5 c

94.6

a

3.7

c K

KU

1

10.6

b

5.4

a 47

.5 c

28

.4 b

2.

2 a

0.6

d 3.

2 a

1.2

b 50

.3

1.7

c 90

.5 b

3.

4 d

Mea

n 8.

6 4.

0 63

.8

16.4

1.

7 1.

2 2.

9 1.

4 10

.5

47.3

84

.1

4.5

F-te

st **

**

**

**

**

**

**

**

**

**

**

Dry

seas

on 2

008/

09

SunO

leic

97R

6.

1 e

2.3

c 80

.5 a

3.

8 d

1.3

c 1.

8 a

2.5

b 1.

7 a

47.8

ab

21.8

a

77.1

c

6.2

a G

eorg

ia-0

2C

6.9

d 2.

6 c

78.3

a

4.8

d 1.

3 c

1.8

a 2.

7 b

1.6

a 51

.4 a

17

.1 b

77

.2 c

5.

6 b

F6-8

-3a

8.6

c 4.

1 b

67.3

b

14.4

c

1.8

b 1.

0 b

2.6

b 1.

2 b

45.8

b

4.7

c 83

.5 b

4.

5 c

KK

5

11.9

a

4.1

b 45

.3 c

30

.2 a

1.

7 b

0.8

bc

3.2

a 1.

3 b

43.6

b

1.5

c 91

.8 a

3.

4 d

KK

U 1

10

.7 b

5.

2 a

49.2

c

27.0

b

2.1

a 0.

7 c

3.3

a 1.

3 b

45.4

b

1.9

c 89

.7 a

3.

4 d

Mea

n 8.

8 3.

6 64

.1

16.0

1.

6 1.

2 2.

9 1.

4 9.

4 46

.8

83.8

4.

7 F-

test

**

**

**

**

**

**

**

**

**

* **

**

a T

he v

arie

ty [(

NC

1709

0 ×

B1)-

9-1

× K

K 6

0-3]

F6-8

-3

*

and

** si

gnifi

cant

at t

he 5

% a

nd 1

% p

roba

bilit

y le

vels

, res

pect

ivel

y M

eans

in th

e sa

me

colu

mn

follo

wed

by

the

sam

e le

tter (

s) a

re n

ot si

gnifi

cant

ly d

iffer

ent (

at P

< 0

.05)

by

DM

RT

65

Tabl

e 2.

M

ean

squa

res f

or fa

tty a

cid

conc

entra

tion,

% o

il, th

e rat

io o

f ole

ic to

lino

leic

aci

ds (O

/L ra

tio),

iodi

ne v

alue

(IV

), un

satu

rate

d to

satu

rate

d fa

tty a

cids

ratio

(U/S

ratio

) in

the F

2 and

F3 g

ener

atio

ns.

Sour

ce

d.f.

Palm

itic

acid

St

earic

ac

id

Ole

ic a

cid

Lino

leic

ac

id

Ara

chid

ic

acid

Ei

cose

noic

ac

id

Behe

nic

acid

Li

gnoc

eric

ac

id

% o

il O

/L ra

tio

IV

U/S

ratio

F 2

gen

erat

ion

GCA

a 4

13.6

0**

1.96

**

609.

50**

39

3.69

**

0.21

**

0.67

**

0.16

**

0.14

**

1.15

**

219.

78**

15

8.31

**

2.51

**

SCA

b 10

0.

26**

0.

14**

5.

75**

4.

23**

0.

01

0.02

0.

03

0.01

7.

22**

37

.86*

* 2.

46**

0.

10*

Reci

proc

als

10

0.09

* 0.

06

8.58

**

5.60

**

0.01

0.

01

0.01

0.

01

4.93

**

0.62

**

2.94

**

0.02

GCA

/SC

Ac

52

.26

14.0

7 10

5.98

93

.1

21.4

7 34

.5

5.72

22

.62

0.16

5.

8 64

.38

24.9

7 Er

ror

48

0.04

0.

02

0.89

0.

58

0.00

3 0.

005

0.01

0.

004

1.17

1.

05

0.61

0.

01

F 3

gen

erat

ion

GCA

4

15.9

1**

3.34

**

723.

20**

43

4.55

**

0.31

**

0.64

**

0.28

**

0.15

* 11

.48*

* 21

7.39

**

155.

70**

3.

64**

SC

A

10

0.80

**

0.10

* 17

.10*

* 19

.44*

* 0.

01

0.03

0.

05

0.01

6.

14**

41

.63*

* 17

.93*

* 0.

16**

Re

cipr

ocal

s 10

0.

50**

0.

09

22.5

3**

17.7

1**

0.02

0.

01

0.05

0.

02

3.32

**

7.33

**

12.4

2**

0.05

G

CA/S

CA

19.8

6 31

.93

42.2

8 22

.35

44.2

6 21

.88

5.96

15

.25

1.87

5.

22

8.68

23

.43

Erro

r 48

0.

08

0.04

6.

19

3.59

0.

01

0.00

4 0.

01

0.01

1.

54

1.54

4.

11

0.02

*

and

** si

gnifi

cant

at t

he 5

% an

d 1

% p

roba

bilit

y le

vels,

resp

ectiv

ely

a Gen

eral

com

bini

ng a

bilit

y b S

peci

fic c

ombi

ning

abi

lity

c The

ratio

of g

ener

al c

ombi

ning

abi

lity

mea

n sq

uare

s and

spec

ific

com

bini

ng a

bilit

y m

ean

squa

res

66

Tab

le 3

. G

ener

al c

ombi

ning

abi

lity

effe

cts f

or fa

tty a

cid

conc

entra

tion,

% o

il, th

e ra

tio o

f ole

ic to

lino

leic

aci

ds (O

/L ra

tio),

iodi

ne v

alue

(IV

), un

satu

rate

d to

satu

rate

d fa

tty a

cids

ratio

(U/S

ratio

) and

cor

rela

tion

betw

een

fatty

aci

d co

mpo

sitio

ns o

f par

enta

l lin

es a

nd th

eir p

roge

nies

in

the

F 2 a

nd F

3 gen

erat

ions

.

Pare

ntal

line

Pa

lmiti

c ac

id

Stea

ric

acid

O

leic

ac

id

Lino

leic

a

cid

Ara

chid

ic

aci

d Ei

cose

noic

ac

id

Beh

enic

ac

id

Lign

ocer

ic

acid

%

oil

O/L

ratio

IV

U

/S ra

tio

F 2

gen

erat

ion

SunO

leic

97R

–0

.98*

* –0

.41*

* 6.

41**

–5

.01*

* –0

.15*

* 0.

24**

–0

.09*

* 0.

11**

0.

14

3.85

**

–2.9

7**

0.47

**

Geo

rgia

-02C

–0

.98*

* –0

.52*

* 7.

20**

–5

.97*

* –0

.15*

* 0.

30**

0.

01

0.14

**

–0.2

8 5.

94**

–3

.91*

* 0.

46**

F6-8

-3a

–0.5

6**

0.22

* 3.

16**

–2

.46*

* 0.

06**

–0

.07*

* –0

.13*

* –0

.15*

* 0.

01

–1.3

1**

–1.6

0**

0.13

**

KK

5

1.41

**

0.22

* –8

.57*

* 7.

14**

0.

06**

–0

.24*

* 0.

01

–0.0

6**

–0.3

6 –4

.30*

* 4.

82**

–0

.49*

*

KK

U 1

1.

11**

0.

50**

–8

.21*

* 6.

30**

0.

18**

–0

.24*

* 0.

20**

–0

.05*

* 0.

48

–4.1

9**

3.66

**

–0.5

7**

rb 0.

59**

0.

47**

0.

59**

0.

58**

0.

50**

0.

56**

0.

39*

0.47

**

–0.0

9 0.

47**

0.

57**

0.

56**

F 3

gen

erat

ion

SunO

leic

97R

–1

.12*

* –0

.64*

* 8.

02**

–5

.75*

* –0

.21*

* 0.

26**

–0

.16*

* 0.

11**

0.

41

5.58

**

–2.8

7**

0.69

**

Geo

rgia

-02C

–1

.11*

* –0

.54*

* 7.

72**

–6

.33*

* –0

.14*

* 0.

28**

–0

.02

0.14

**

1.30

**

4.41

**

–4.1

0**

0.50

**

F6-8

-3a

–0.4

3**

0.32

**

2.16

**

–1.8

5**

0.10

**

–0.1

0**

–0.1

5**

–0.1

1**

0.52

–1

.93*

* –1

.41*

0.

02

KK

5

1.59

**

0.15

**

-9.4

8**

7.28

**

0.04

–0

.19*

* 0.

13**

–0

.11*

* –1

.11*

* –3

.86*

* 4.

31**

–0

.59*

*

KK

U 1

1.

07**

0.

71**

-8

.42*

* 6.

65**

0.

21**

–0

.25*

* 0.

21**

–0

.04

–1.1

2**

–4.2

1**

4.07

**

–0.6

2**

rb 0.

53**

0.

54**

0.

56**

0.

54**

0.

52**

0.

55**

0.

36*

0.44

**

0.16

0.

45**

0.

47**

0.

56**

*

and

** si

gnifi

cant

diff

eren

t fro

m z

ero

at th

e 5%

and

1%

pro

babi

lity

leve

ls, r

espe

ctiv

ely

a The

var

iety

[(N

C17

090

× B1

)-9-

1 ×

KK

60-

3]F6

-8-3

b C

orre

latio

n co

effic

ient

bet

wee

n fa

tty a

cid

cont

ents

of p

aren

ts a

nd th

eir p

roge

nies

67

Tabl

e 4.

Sp

ecifi

c co

mbi

ning

abi

lity

effe

cts f

or fa

tty a

cid

conc

entra

tion,

% o

il, th

e rat

io o

f ole

ic to

lino

leic

aci

ds (O

/L ra

tio),

iodi

ne v

alue

(IV

), un

satu

rate

d to

satu

rate

d fa

tty a

cids

ratio

(U/S

ratio

) in

the F

2 gen

erat

ion.

* an

d **

sign

ifica

nt d

iffer

ent f

rom

zer

o at

the 5

% a

nd 1

% p

roba

bilit

y le

vels,

resp

ectiv

ely

a The

var

iety

[(N

C170

90 ×

B1)

-9-1

× K

K 6

0-3]

F6-8

-3

Cros

s Pa

lmiti

c ac

id

Stea

ric

acid

O

leic

ac

id

Lino

leic

ac

id

Ara

chid

ic

acid

Ei

cose

noic

ac

id

Behe

nic

acid

Li

gnoc

eric

ac

id

% o

il O

/L ra

tio

IV

U/S

ratio

Su

nOle

ic 9

7R ×

Geo

rgia

-02C

–0

.27

–0.1

4 2.

16**

–1

.87*

* –0

.05

0.11

* –0

.09

–0.0

3 –1

.68

5.30

**

–1.2

9*

–0.0

6 Su

nOle

ic 9

7R ×

F6-

8-3a

0.05

0.

01

–1.1

8*

0.96

0.

01

–0.0

1 –0

.05

–0.0

7 1.

87*

–2.1

2*

0.64

–0

.20*

* Su

nOle

ic 9

7R ×

KK

5

0.47

**

0.49

**

–2.4

8**

1.98

**

0.12

**

–0.1

8**

0.07

–0

.03

1.54

–3

.43*

* 1.

15

–0.1

3*

SunO

leic

97R

× K

KU

1

0.33

–0

.35*

* –0

.75

0.97

–0

.07

–0.0

7 0.

03

0.02

–0

.72

–3.2

2**

0.99

–0

.02

Geo

rgia

-02C

× S

unO

leic

97R

0.

07

0.10

–0

.09

–0.2

1 0.

05

–0.0

1 0.

002

–0.0

5 1.

26

1.09

–0

.45

0.24

**

Geo

rgia

-02C

× F

6-8-

3a 0.

18

–0.0

1 –1

.55*

1.

44*

0.03

–0

.04

0.23

**

–0.0

1 2.

84**

–3

.63*

* 1.

11

–0.0

9 G

eorg

ia-0

2C ×

KK

5

0.11

0.

04

–0.9

9 0.

73

0.01

–0

.08

0.00

2 –0

.03

–2.0

5*

–5.0

6**

0.36

–0

.15*

G

eorg

ia-0

2C ×

KK

U 1

0.

39*

0.17

–1

.40

1.18

0.

08*

–0.0

7 0.

12

0.07

1.

79*

–4.9

4**

0.78

–0

.10

F6-8

-3a ×

Sun

Ole

ic 9

7R

0.08

0.

26*

–0.7

9 0.

34

0.10

* 0.

08

0.11

0.

09

1.08

–0

.30

–0.0

4 0.

01

F6-8

-3a ×

Geo

rgia

-02C

0.

02

0.20

–0

.16

0.19

0.

08*

–0.0

3 –0

.02

0.12

* –1

.70

–0.3

8 0.

15

–0.1

4*

F6-8

-3a ×

KK

5

0.06

0.

003

1.13

–1

.62*

–0

.02

0.06

–0

.08

0.03

0.

46

1.84

* –1

.80*

* 0.

05

F6-8

-3a ×

KK

U 1

0.

17

–0.1

3 –0

.08

–0.1

6 –0

.08*

0.

03

–0.0

5 0.

04

–2.1

7**

1.57

–0

.31

–0.0

5 K

K 5

× S

unO

leic

97R

0.

26

–0.0

8 –3

.48*

* 2.

70**

0.

07

–0.0

1 0.

17*

0.05

–0

.88

–0.5

6 1.

68**

–0

.35*

* K

K 5

× G

eorg

ia-0

2C

0.39

* 0.

14

–2.9

2**

2.31

**

0.04

–0

.06

0.07

–0

.002

0.

40

–0.5

8 1.

43*

–0.1

4*

KK

5 ×

F6-

8-3a

–0.0

5 –0

.05

–0.2

8 0.

75

0.00

4 –0

.003

–0

.05

0.01

–0

.35

–0.1

3 1.

08

–0.0

1 K

K 5

× K

KU

1

–0.4

3*

–0.2

4*

1.91

* –1

.28*

–0

.02

0.11

* 0.

05

0.02

–0

.27

3.38

**

–0.4

9 –0

.13*

K

KU

1 ×

Sun

Ole

ic 9

7R

–0.1

6 0.

16

–0.4

4 0.

41

0.04

–0

.02

0.02

–0

.01

1.58

–0

.09

0.32

–0

.10

KK

U 1

× G

eorg

ia-0

2C

0.39

* 0.

06

–4.1

3**

3.75

**

–0.0

2 –0

.18*

* 0.

03

–0.0

2 –2

.25*

–0

.99

2.80

**

–0.2

9**

KK

U 1

× F

6-8-

3a 0.

12

–0.1

7 –1

.70*

0.

57

–0.0

2 0.

02

0.04

0.

04

2.48

**

–0.1

8 –0

.48

–0.0

1 K

KU

1 ×

KK

5

0.16

0.

34**

–1

.21

0.31

0.

12**

–0

.07

0.02

–0

.04

–2.0

6*

–0.0

6 –0

.54

0.26

**

68

Tab

le 5

. Sp

ecifi

c co

mbi

ning

abi

lity

effe

cts f

or fa

tty a

cid

conc

entra

tion,

% o

il, th

e ra

tio o

f ole

ic to

lino

leic

aci

ds (O

/L ra

tio),

iodi

ne v

alue

(IV

), un

satu

rate

d to

satu

rate

d fa

tty a

cids

ratio

(U/S

ratio

) in

the

F 3 g

ener

atio

n.

* an

d **

sign

ifica

nt d

iffer

ent f

rom

zer

o at

the

5% a

nd 1

% p

roba

bilit

y le

vels

, res

pect

ivel

y a T

he v

arie

ty [(

NC

1709

0 ×

B1)-

9-1

× K

K 6

0-3]

F6-8

-3

Cro

ss

Palm

itic

acid

St

earic

aci

d O

leic

aci

d Li

nole

ic

acid

A

rach

idic

ac

id

Eico

seno

ic

acid

B

ehen

ic

acid

Li

gnoc

eric

ac

id

% o

il O

/L ra

tio

IV

U/S

ratio

Su

nOle

ic 9

7R ×

Geo

rgia

-02C

–0

.53*

*–0

.14

3.47

–3.6

8*–0

.05

0.15

**–0

.07

–0.0

1–2

.57*

8.40

**–3

.29*

0.32

**Su

nOle

ic 9

7R ×

F6-

8-3a

0.03

–0.0

8–1

.58

0.92

–0.0

2–0

.10*

0.09

–0.0

9–0

.12

–4.0

8**

0.18

–0.0

7Su

nOle

ic 9

7R ×

KK

5

0.47

**0.

08–1

.92

1.46

–0.0

3–0

.07

0.18

*0.

032.

94**

–3.1

6**

0.81

–0.3

3**

SunO

leic

97R

× K

KU

1

1.16

0.23

–3.3

85.

24**

0.05

–0.1

2*–0

.16

–0.0

4–0

.55

–5.4

1**

6.10

**–0

.34*

*G

eorg

ia-0

2C ×

Sun

Ole

ic 9

7R

0.00

–0.1

0–0

.54

0.69

–0.0

60.

040.

01–0

.05

–0.4

1–4

.85*

*0.

760.

07G

eorg

ia-0

2C ×

F6-

8-3a

–0.0

10.

081.

14–0

.38

0.04

–0.0

30.

040.

001.

10–1

.94

0.30

–0.0

3G

eorg

ia-0

2C ×

KK

5

0.47

**0.

31**

–2.8

93.

31*

0.11

–0.1

8**

0.08

–0.0

20.

44–4

.50*

*3.

12–0

.27*

*G

eorg

ia-0

2C ×

KK

U 1

0.

44**

–0.2

8–3

.50

2.46

–0.0

4–0

.12*

0.08

0.09

–1.1

1–3

.87*

*1.

13–0

.24*

F6-8

-3a ×

Sun

Ole

ic 9

7R

0.46

**0.

02–2

.59

2.10

0.01

–0.0

40.

06–0

.04

–2.1

3*–1

.20

1.38

–0.1

8F6

-8-3

a × G

eorg

ia-0

2C

–0.1

10.

21–0

.17

–1.1

30.

08–0

.05

0.00

–0.0

1–0

.36

0.65

–2.1

4–0

.13

F6-8

-3a ×

KK

5

0.28

–0.1

2–0

.88

0.82

–0.0

20.

04–0

.32*

*–0

.01

–1.1

81.

620.

720.

05F6

-8-3

a × K

KU

1

–0.2

30.

31*

–0.5

7–0

.23

0.05

0.04

0.14

0.05

2.08

*2.

18*

–0.8

9–0

.05

KK

5 ×

Sun

Ole

ic 9

7R

1.14

–0.0

4–9

.38*

*8.

54**

–0.1

6*–0

.20*

*0.

01–0

.21*

*–0

.89

–3.3

3**

6.58

**–0

.21*

KK

5 ×

Geo

rgia

-02C

0.

26–0

.13

–2.2

41.

16–0

.03

0.01

0.03

0.08

0.07

–0.2

10.

05–0

.10

KK

5 ×

F6-

8-3a

0.85

**–0

.43*

*–1

.74

1.35

–0.0

90.

10*

0.11

0.02

0.25

–0.2

90.

95–0

.10

KK

5 ×

KK

U 1

–0

.47*

–0.3

4*2.

47–1

.98

–0.0

70.

11*

–0.0

5–0

.11

–1.3

83.

16**

–1.2

20.

35**

KK

U 1

× S

unO

leic

97R

0.

23–0

.05

–0.8

31.

840.

06–0

.01

0.23

*0.

061.

24–0

.19

2.45

*-0

.06

KK

U 1

× G

eorg

ia-0

2C

–0.0

90.

111.

230.

690.

03–0

.10*

0.08

0.05

–0.9

10.

062.

150.

05K

KU

1 ×

F6-

8-3a

0.38

0.41

–2.7

31.

500.

19**

–0.0

10.

34**

0.08

–2.9

3**

–0.2

50.

20–0

.31*

KK

U 1

× K

K 5

–0

.09

0.08

–0.3

20.

48–0

.05

–0.0

1–0

.20*

–0.1

8*0.

04–0

.03

0.51

0.08

69

DISCUSSION

The parental lines could be classified into three groups based on their impact on oleic acid concentration (Table 1). SunOleic 97R and Georgia-02C were classified as genotypes with high oleic acid. [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 was an intermediate genotype, whereas KKU 1 and KK 5 were classified as the normal oleic group.

The difference in the parental lines for oleic acid was not surprising as SunOleic 97R and Georgia-02C are known to carry ol1 and ol2 genes inhibiting linoleic biosynthesis in peanut (Yu et al., 2008), and, thus, these two recessive genes increase oleic acid in these two genotypes. These genotypes have been used in peanut breeding programs to improve oleic acid concentration. SunOleic 97R was developed from the naturally occurring mutants for ol1 and ol2 (Gorbet and Knauft, 2000), whereas Georgia-02C was developed from artificial mutants induced by gamma irradiation (Branch, 2003).

The surprising genotype was the line [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 which was intermediate between high group and normal group. It showed consistently higher oleic acid concentration than did KKU 1 and KK 5. [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 was derived from KK 60-3 which was selected from a blend population of NC 7. NC 7 also showed high combining ability for oleic acid (Mercer et al., 1990).

It is unlikely that NC 7 transmission of any ol gene to [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 because oleic acid of NC 7 was in a range of the normal oleic group (Isleib et al., 2006b). Therefore, it would carry neither ol1 nor ol2. If one of the ol genes exist in [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3, it should be from the other parents, NC 17090 or B1, but their pedigree is not known. The probability of finding a natural mutant of the ol1 allele is quite high with 31.6% in mini-core accessions in the United States (Chu et al., 2007). The cross NC 17090 × B1, was a segregating population from the NCSU peanut breeding program (Jogloy, personal communication). Therefore, one of its progenitors may carry a mutant gene. If so the line [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 should be normal instead of intermediate because the dominance expression of the mutant ol gene. The expression of ahFAD2A transcript in transgenic high oleic peanut genotype was sufficient to alter high oleic to normal oleic (Jung et al., 2000). The most plausible explanation is that it carries other genetic factors that modify the oleic acid concentration. The presence of an intermediate phenotype indicated that the inheritance of oleic acid concentration might modified beyond the genetic control of just the ol genes.

Normal-oleic peanut genotypes have 36 to 70% oleic acid and 15 to 43% linoleic acid, whereas high oleate peanut genotypes have approximately 80% oleic acid and approximately 2% linoleic acid. The first report of high oleic peanut genotypes found only two genotypes with elevated oleic acid concentration out of 494 accessions (Norden et al., 1987) that were high-oleic. Segregation patterns in the F2 population indicated that high oleic acid concentration was digenically inherited in Spanish-type peanut, but there seems to be more allelic variation both within and among them. In addition, variation within the high and normal oleate classes indicated that other factors may be involved in determining the precise O/L ratio (Lopez et al., 2001).

70

In the absence of ol mutant genes in normal oleate peanut genotypes, there is genetic variation for oleic characters. In the presence of two ol mutant genes, there is also genetic variation in oleic acid concentration. This type of variation should not exist if only two ol genes control this character because all peanut genotypes are homologous for these genes. The results might indicate that oleic acid concentration in peanut is modified beyond the control of these two genes although with relatively small effect (Lopez et al., 2001; Moore and Knauft, 1989; Norden et al., 1987).

GCA and SCA effects were significant for oleic, linoleic acids, % oil, O/L ratio, IV and U/S ratio in both the F2 and F3 generations. The results suggested that additive and nonadditive gene action contributed to the variation in these characters. However, GCA effects contributed greater than the SCA effects for these characters. The comparative magnitude of GCA and SCA mean squares is an indicator of the relative importance of the additive and nonadditive gene action in the inheritance of character (Kornegay and Temple, 1986). The results of this study showed high ratios of GCA and SCA mean squares for oleic and linoleic acids suggesting that additive gene action was more important than nonadditive gene action for these characters. The magnitude of GCA variance for % oil was smaller than SCA in the F2 generation indicating a preponderance of nonadditive gene action for this character.

Reciprocal effects were also detected for oleic, linoleic acids, % oil, O/L ratio and IV in both the F2 and F3 generations, suggesting that cytoplasmic factors were also important for these characters. Similar results have been previously reported by Mercer et al. (1990) that GCA effects and reciprocal effects were significant for oleic acid, linoleic acid, O/L ratio and IV in the F1 and F2 generations, but SCA effects were not significant for oleic and linoleic acid. In addition, they found that maternal effects were significant only in the F1 generation for these characters.

SunOleic 97R and Georgia-02C were the best parents for high oleic acid concentration and O/L ratio because they had high GCA effects for these characters. This was not unexpected because they are high oleic genotypes. The crosses with SunOleic 97R and Georgia-02C are expected to give high oleic progenies to exploit the merit of ol genes from their parents.

The crosses with [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 as a parent are also interesting especially those with SunOleic 97R and Georgia-02C because [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 also had significant and positive combining ability, whereas KK 5 and KKU 1 had significant and negative combining ability for oleic acid concentration and O/L ratio.

Moreover, the correlation between the performance per se of parental lines and their progenies was positive and significant for oleic, linoleic acids and O/L ratio, indicating that performance per se can be used as a predictor to select good combining parents for these characters.

In conclusion, additive gene action was more important than nonadditive gene action in determining fatty acid concentration. Genotypes with high oleic concentration can be selected in early generations especially in crosses with SunOleic 97R or Georgia-02C as parents because they carry the two recessive mutant genes (ol1 and ol2) for high oleic acid. The crosses SunOleic 97R × [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 and Georgia-02C × [(NC17090 × B1)-9-1 × KK 60-3]F6-8-3 might hold promise in identifying genetic control of oleic acid concentration beyond two major ol genes.

71

ACKNOWLEDGEMENTS

This study was funded by the Thailand Research Funded through the Royal Golden Jubilee Ph.D. Program (Grant No. PHD/0035/2548) and greatful acknowledgement is made to the Thailand Research Fund (TRF), the Commission for Higher Education(CHE) and Khon Kaen University for providing financial support to this research through the Distingish Research Professor Grant of Professor Dr Aran Patanothai. Thanks are extended to Department of Biochemistry, Faculty of Science, Khon Kaen University,Khon Kaen, Thailand for providing laboratory facilities. Dr. Corley C. Holbrook is also acknowledged for his critical reading and valuable suggestions.

REFERENCES

Andersen, P.C., and D.W. Gorbet. 2002. Influence of year and planting date on fatty acid chemistry of high oleic acid and normal peanut genotypes. J. Agric. Food Chem. 50: 1298-1305.

Bannon, C.D., G.J. Breen, J.D. Craske, N. Hai, N.L. Harper, and K.L. O’Rourke. 1982. Analysis of fatty acid methyl esters with high accuracy and reliability. III. Literature review of an investigation into the development of rapid procedures for the methoxide-catalysed methanolysis of fats and oils. J. Chromatogr. 247: 71-89.

Branch, W.D. 2003. Registration of ‘Georgia-02C’ peanut. Crop Sci. 43:1883-1884. Chu, Y., C.C. Holbrook, and P. Ozias-Akins. 2009. Two alleles of ahFAD2B control

the high oleic acid trait in cultivated peanut. Crop Sci. 49: 2029-2036. Chu, Y., L. Ramos, C.C. Holbrook, and P. Ozias-Akins. 2007. Frequency of a loss-

of-function mutation in oleoyl-PC desaturase (ahFAD2A) in the mini-core of the U.S. peanut germplasm collection. Crop Sci. 47: 2372-2378.

Dwivedi, S.L., S.N. Nigam, R. Jambunathan, K.L. Sahrawat, G.V.S. Nagabhushanam and K. Raghunath. 1993. Effect of genotypes and environments on oil content and oil quality parameters and their correlation in peanut (Arachis hypogaea. L.). Peanut Sci. 20: 84-89.

Gorbet, D.W., and D.A. Knauft. 2000. Registration of ‘SunOleic 97R’ peanut. Crop Sci. 40: 1190-1191.

Griffing, B. 1956. Concept of general and specific combining ability in relation to diallel crossing systems. Aust. J. Biol. Sci. 9: 463-493.

Hoshmand, A.R. 2006. Design of experiments for agriculture and the natural sciences. 2nd edn. Chapman & Hall/CRC, Boca Raton, FL.

Isleib, T.G., C.T. Young, and D.A. Knauft. 1996. Fatty acid genotypes of five Virginia-type peanut cultivars. Crop Sci. 36: 556-558.

Isleib, T.G., R.F. Wilson, and W.P. Novitzky. 2006a. Partial dominant, pleiotropism, and epistasis in the inheritance of the high-oleate trait in peanut. Crop Sci. 46: 1331-1335.

Isleib, T.G., H.E. Pattee, T.H. Sanders, K.W. Hendrix, and L.O. Dean. 2006b. Compositional and sensory comparisons between normal- and high-oleic peanuts. J. Agric. Food Chem. 54: 1759-1763.

72

Jung, S., G. Powell, K. Moore, and A. Abbott. 2000. The high oleate trait in the cultivated peanut (Arachis hypogaea L.). II. Molecular basis and genetics of the trait. Mol. Gen. Genet. 263: 806-811.

Kornegay, J.L., and S.R. Temple. 1986. Inheritance and combing ability of leaf hopper defense mechanism in common bean. Crop Sci. 26: 1153-1158.

Lόpez, Y., O.D. Smith, S.A. Senseman, and W.L. Rooney. 2001. Genetic factor influencing high oleic acid content in Spanish market-type peanut cultivars. Crop Sci. 41: 51-56.

Mercer, L.C., J.C. Wynne, and C.T. Young. 1990. Inheritance of fatty acid content in peanut oil. Peanut Sci. 17: 17-21.

Moore, K.M., and D.A. Knauft. 1989. The inheritance of high oleic acid in peanut. J. Hered. 80: 252-253.

Mugendi, J.B., C.A. Sims, D.W. Gorbet, and S.F. O’Keefe. 1998. Flavor stability of high-oleic peanut stored at low humidity. J. Am. Oil Chem. Soc. 75: 21-25.

Norden, A.J., D.W. Gorbet, D.A. Knauft, and C.T. Young. 1987. Variability in oil quality among peanut genotypes in Florida breeding program. Peanut Sci. 14: 7-11.

O’Byrne, D.J., D.A. Knauft, and R.B. Shireman. 1997. Low fat-monounsaturated rich diets containing high-oleic peanuts improve serum lipoprotein profiles. Lipids. 32: 687-695.

O’Keefe, S.F., V.A. Wiley, and D.A. Knauft. 1993. Comparison of oxidative stability of high- and normal-oleic peanut oils. J. Am. Oil Chem. Soc. 70: 489-492.

Singkham, N., S. Jogloy, T. Kesmala, P. Swatsitang, P. Jaisil, and N. Puppala. 2010. Genotypic variability and genotype by environment interactions in oil and fatty acids in high, intermediate and low oleic acid peanut genotypes. J. Agric. Food Chem. 58: 6257-6263.

Upadhyaya, H.D., and S.N. Nigam. 1999. Detection of epistasis for protein and oil contents and oil quality parameters in peanut. Crop Sci. 39: 115-118.

Yu, S., L. Pan, Q. Yang, P. Min, Z. Ren, and H.Zhang. 2008. Comparison of the Δ12 fatty acid desaturase gene between high-oleic and normal-oleic peanut genotypes. J. Genet. Genomics. 35: 679-685.

73


GENETIC DIVERSITY IN MUTAGINIZED FAMILIES OF

FENUGREEK (Trigonella foenum-graecom L.)

S.K. KAUSHIK1* AND E.V.D. SASTRY2

SUMMARY

The present experiment was conducted to estimate the mutagenic effects of gamma rays on yield and its attributes in M7-generation of fenugreek cultivar RMt-1. The experiment comprised of 29 M7 mutant lines generated through irradiation with γ-rays along with three checks namely; RMt-1, RMt-143 and local check (desi Methi) was conducted to estimate the magnitude and pattern of diversity present in the crop using metroglyph analysis and index score values. The experiment was carried-out in completely randomized block design replicated thrice during rabi 1997-98. The observations were recorded on 5 randomly selected competitive plants for plant height, number of branches, pods per plant, pod length, seeds per pod, test weight, harvest index and seed yield. Whereas four traits namely; days to 50% flowering, protein content (%), powdery mildew incidence and days to maturity were recorded on plot basis. The mean values of each trait were used for statistical analysis. ANOVA revealed significant variability for all the traits except for branches per plant, pod length, days to maturity and harvest index. Index score values for different families varied from 14 - 24 indicating wide range of diversity among the lines. Scattering diagram revealed yield superiority of all families except two (13-2 and UM-305) over local check and 12 families (37.50%) over RMt-1. Families of cluster-1 (168-2,551-1) and cluster-6 (50-1,764-2) exhibited maximum genetic divergence indicated that hybridization between the families of these two clusters would give desired segregants. All the families grouped into three groups based on the frequency distribution of index score values. 11 Families of 3rd group showed good coordination between index score values and scatter diagram. The family No. 20 (551-1) found to be best among all families and further evaluation of this family in yield trial may prove beneficial.

Key words: Fenugreek, M7-generation, Meteroglyph, Index score, Scatter diagram, seed yield. 1 Krishi Vigyan Kendra-Ujjain, Madhya Pradesh, India 2 Department of Plant Breeding & Genetics, SKN College of Agriculture (SKRAU)-Jobner, Rajasthan,

India * Corresponding author: [email protected]

74

INTRODUCTION

The land of spices India, is the largest producer, consumer and exporter of spices in the globe. As per Spice Board of India, 52 spices are grown in India. Fenugreek, popularly known as “Methi”, is an important autogamous spice crop largely grown in the northern India during rabi season. Among the various spices grown, fenugreek occupies a prime position with respect to area as well as production is concerned. Earlier studies indicated that the amount of variability is limited particularly for yield and yield-contributing attributes which often act as a limiting factor in the improvement of this crop. Although genetic variability can be generated through hybridization, success has been limited owing to poor seed set and the small size of flowers of the crop. Furthermore, germplasm studies indicated that indigenous collections may not be useful in creating adequate variability. Hence, use of mutagens is the only option to induce the variability. Induced mutations have been proved to be a valuable approach in generating genetic variability for breeding improved varieties.

Mutation breeding has been very useful in inducing new variability which is an essential requirement of any plant breeding programme in changing agriculture pattern of the day (Micke et al. 1990). Ionising radiation is a potent tool to induce genetic variability. Therefore, the present investigation is based on advanced lines generated through radiation treatment (29 M7-families) along with parental variety RMt-1, RMt-143 and local check for comparison. Genetic divergence among large numbers of germplasm collections can be analysed by two methods: (i) D2 –analysis or (ii) the metroglyph technique. Since it is difficult to compute D2 values among large numbers of accessions, another relatively simple and easy method, the metroglyph technique, was tried upon the set of 32 genotypes. It has been suggested that the metroglyph technique would be suitable for preliminary grouping prior to undertaking D2 analysis (Chandra, 1977).

Metroglyph analysis was developed by Anderson (1957) for displaying a set of multivariate responses. In the graph each genotype vector is represented by a circle of fixed radius (called as glyph) with rays emanating from its periphery. Each variable is assigned a sign (ray position) and the length of the ray represents the index score (long ray, short ray, no ray for high, medium, low scores, respectively) of the variate based on the range of variability. An index can be prepared for each glyph by assigning values to long ray(say3), short ray(say2) and no ray(say 1). The glyphs are positioned in the graph by selecting two most variable characters and plotting means of one character against means of the other character. Each character was illustrated by a ray at a fixed position on each glyph. The range of the character was represented by a different length of the ray. The index score were obtained by allotting numerical values (1, 2 or 3) to three grades of expression (i.e. low, medium and high) recognized in respect of each character and finally summing up the scores obtained by each genotype for all characters under study.

The meteroglyph analysis is useful for germplasm classification and divergence studies. Coupled with index score, this method is applicable for the screening and classification of genotypes when large number of accessions are available. Therefore, the meteroglyph analysis was used to study the mutagenized

75

morphological diversity and to identify high yielding superior families combining desirable combination of various attributes. Earlier Vijay Kumar and Ramankutty (1987), Gill and Tripathi (1988), Kailash Chandra (1992) and Kumhar (1996) used this method to identify high yielding superior varieties combining desirable combinations of various attributes.


The 29 mutant lines of M7-generation, selected for present study, were developed by using different doses of γ-radiation to fenugreek variety RMt-1. The material utilized in the present study was derived from 10, 20, 30, 40 and 50 Kr. doses of γ-radiation to parental variety RMt-1 (Table 1). These 29 mutant lines along with RMt-1, RMt-143 and local check were grown for evaluation in a complete randomized block design with three replications at Agriculture Research Farm. S.K.N. College of Agriculture, Jobner, during rabi 1997-98. The plot size of 4 x 1.2 m, with a spacing of 30 cm between the rows and intra- row spacing of 10 cm maintained by thinning. Recommended agronomical practices were followed to raise a good crop.

Observations were recorded on days to 50 % flowering, days to maturity, powdery mildew incidence, test weight (g), seed yield (kg) and harvest index on plot basis whereas plant height (cm), number of branches per plant, number of pods per plant, pod length and seeds per pod were recorded on ten competitive randomly selected plants from each experimental plot. For estimating the protein content, protein content (%) the nitrogen content in the seeds of each plot was estimated by Calorimetric method (Snell and Snell, 1939) using blue filter. It was multiplied by a factor of 6.25 to get percent protein in each sample (A.O.A.C., 1960). Powdery mildew incidence was recorded on a zero to nine scale (Mayee and Datar, 1986) under field conditions and then finally per cent of disease intensity was calculated for each family.

Genetic divergence was carried out using meteroglyph analysis as per the method given by Anderson (1957). In the present investigation two most variable characters based on coefficient of variation (i.e. number of branches per plant and harvest index) were selected for the purpose of analysis. Mean values of number of branches per plant on x axis for each genotype were plotted on the graph against the mean values of harvest index on y axis. Thus, each genotype occupied a definite position called glyph on the graph. Variations for all other characters of each genotype were represented by rays on the respective glyph. Variation for each character is depicted by the length of ray. For this purpose, the range for each character was divided into three equidistant scores as depicted in Table 2. Thus, the length of ray for a particular character on the glyph may be short or medium or long depending on the index score value of a genotype (Table 2). The usefulness of a genotype was calculated by sum total of index score values with regard to all the traits of respective family (Table 3).

76

Table 1. Mutant lines evaluated in the present study. S. No. Line No. S. No. Line No. 1. 1-2 17. 543-3 2. 13-1 18. 549-1 3. 13-2 19 549-2 4. 50-1 20. 551-1 5. 50-3 21. 551-3 6. 168-1 22. 552-1 7. 168-2 23. 552-2 8. 244-1 24. 554-1 9. 248-2 25. 454-2 10. 249-1 26. 682-2 11. 249-4 27. 743-1 12. 293-4 28. 764-2 13. 504-1 29. 765-2 14. 204-4 30. Local Check 15. RMt-1 31. UM-305 16. 540-3 32. RMt-143

Table 3. Different groups of families based on index score values of a family. Index Score value Families Frequency Group 14 30 1 I 17 16, 19, 23 3 II 18 3, 6, 9, 10 4 20 2, 7, 18, 27, 29, 31 6 III 21 11, 12, 15, 21, 24 5 22 13, 14, 22, 8 4 23 1, 5, 17 3 24 4, 20, 25, 26, 28, 32 6

77

Tab

le 2

. The

rang

e of

var

iabi

lity

and

clas

s-in

terv

als o

f diff

eren

t cha

ract

ers i

n M

7-ge

nera

tion

of fe

nugr

eek

Cha

ract

ers

Var

iabi

lity

(R

ange

) Sc

ore

1 Sc

ore

2 Sc

ore

3 C

.V.

(%)

Rang

e Si

gn

Rang

e Si

gn

Rang

e Si

gn

Num

ber o

f bra

nche

s

3.87

-5.7

0 3.

87-4

.48

4.

48-5

.09

5.

09-5

.70

14

.91

Har

vest

Inde

x

22.5

3-32

.87

22.5

3-25

.98

25

.98-

29.4

3

29.4

3-32

.87

14

.58

Day

s to

50%

flow

erin

g

45.6

7-68

.33

68.3

3-60

.78

60

.78-

53.2

3

53.2

3-45

.67

4.

81

Plan

t hei

ght(c

m)

75

.37-

102.

30

75.3

7-84

.31

84

.31-

93.2

5

93.2

5-10

2.30

9.28

Num

ber o

f pod

s per

pla

nt

20

.13-

46.6

8 20

.13-

28.9

8

28.9

8-37

.83

37

.83-

46.6

8

10.2

7

Pod

leng

th (c

m)

8.

87-1

2.74

8.

87-1

0.16

10.1

6-11

.45

11

.45-

12.7

4

12.3

3

Seed

s per

pod

13.6

3-17

.53

13.6

3-14

.93

14

.93-

16.2

3

16.2

3-17

.53

7.

04

Day

s to

mat

urity

124.

67-1

36.6

7 13

6.67

-132

.67

13

2.67

-128

.67

12

8.67

-124

-67

4.

09

Test

wei

ght (

g)

10

.10-

14.9

3 10

.10-

11.7

1

11.7

1-13

.32

13

.32-

14.9

3

2.97

Seed

yie

ld p

er p

lot (

kg)

0.

57-1

.15

0.57

-0.7

6

0.76

-0.9

5

0.95

-1.1

5

12.5

1

Prot

ein

cont

ent (

%)

10

.94-

44.2

7 10

.94-

22.0

5

22.0

5-33

.16

33

.16-

44.2

7

11.5

2

Pow

dery

mild

ew

20

.50-

40.2

5 40

.25-

33.6

7

33.6

7-27

.09

27

.09-

20.5

0

17.0

4

78


The analysis of variance revealed significant differences among genotypes for all traits studied except for branches per plant, pod length, days to maturity and harvest index suggesting ample scope to identify desirable genotypes for these variable traits. The frequency of mutagenized families with their respective index score values are depicted in Figure 1. Index score values of different families ranged from a low of 14 to a high of 24 indicating wide range of diversity among the families. Lines with an index score of 20 were highest in frequency. None of the lines exhibited index score values of 15, 16 and 19. Local check had the lowest index score, indicating the effectiveness in creating superior lines. Further, it is also clear that the mutant lines varied widely in comparison to parental line RMt-1. Perusal of Figure 2 indicated that four families gave higher harvest index than RMt-1, whereas all families except one (i.e. 764-2), revealed harvest index superiority over local cultivar. The superiority of the mutant lines over local check was attributable to this character. Four families (50-1, 244-1, 552-1 and 682-2) showed superior seed yield, number of branches per plant and index score values than RMt-1. All families except one (UM-305) revealed superior seed yield over the local cultivar.

All families including checks based on frequency distribution of index score values grouped into three groups (Table 2). The first group represented by local cultivar only had a minimum score (i.e.14) and showed least useful genotype. The second group displayed by 20 per cent of total families with a mean score value of 17.5, indicating these genotypes were more useful than the local cultivar. The third group included parental variety RMt-1and RMt-143 comprised of nearly 75 per cent of total families with an average score value of 22. Since 13 families of the third category were better than RMt-1, this suggested that induced mutation may be effectively used for procuring superior families in this crop. The genotypes used as parents in the hybridization programme should be chosen from different groups representing wide genetic variability.

In general, the scattering of glyphs did not reflect any clear clustering patterns. Although, an attempt was made to cluster them in six different groups based on common characteristics (Figure 2). The families included in cluster 1 are highly desirable because they had high harvest index than both parental varieties (i.e. RMt-1 and local check). Therefore, the analysis showed that superior families could be identified if selection is made for families having high seed yield, number of branches per plant, harvest index and high index score values.

In the present investigation family number 20 (551-1) appears to be the best among all as it disclosed significant superiority over both parental varieties and local check, and it had high index score value for most of the economic characters. Further evaluation of this family in yield trials may prove to be promising with regard to yielding ability. Gill and Tripathi (1988), Kailash Chandra (1992) and Kumhar (1996) used this technique to identify high yielding superior variety combining desirable combinations of various attributes.

79

Figure 1. Frequency distribution of index score values

22

23

24

25

26

27

28

29

30

31

32

33

HAR

VES

T IN

DEX

(%)

3 4 5 6NUMBER OF BRANCHES

1

23

4

5

6

Figure 2. Meteroglyph diagram of 32 M7 lines of Fenugreek

80

REFERENCES

Anderson, E. 1957. A semi-graphical method for the analysis of complex problems. Proc. Nat. Acad. Sci. USA 43: 923-927.

A.O.A.C. 1970. Official methods of analysis. AOAC, P.O. 13540. Benjamin Franklin Station Washington DC; 20044, U.S.A.

Chandra, S 1977. Comparison of Mahalanobis's method and metroglyph technique in the study of genetic divergence in Linum usitatissimum L. germ plasm collection. Euphytica 26: 141-148.

Gill, SS and Tripathi, BK 1988. Meteroglyph and index score characters analysis of some exotic sugarcane hybrids. J. Agr. Sci; 58: 806.

Kailash Chandra 1992. Genetic variation and association among yield and yield related characters in fenugreek. M.Sc. (Ag.); Thesis, R.A.U. Bikaner.

Kumhar, BL 1996. Variation for seed yield and yield related morphological characters in M1- generation of fenugreek (Trigonella foenum-graecum L.). M.Sc. (Ag.); Thesis, R.A.U. Bikaner.

Mayee, CD and Datar, VV 1986. Phytopathometry. Technical Bulletin-I (Special Bulletin 3), pp: 218 – 223; Marathwara Agric. Univ. Parbhani.

Snell, FB and Snell, CT 1939. Calorimetric method of analysis II A.D. Vannostrand Co. Inc. New York.

Vijay Kumar, NK and Ramankutty, NN 1987. Meteroglyph analysis of hybrid attributes in certain tall indica rice (Oriza sativa) varieties in North Kerla. J. Agr. Sci. 85: 452.

81

SABRAO Journal

of Breeding and Genetics

43 (1) 81-90, 2011

ASSESSMENT OF GENETIC VARIABILITY OF AGRONOMIC

CHARACTERS IN INDIGENOUS AND EXOTIC COLLECTION

OF BLACK SOYBEAN (Glycine max (L.) Merrill.)

ANURADHA BHARTIYA1*

, J. P. ADITYA1, G. SINGH

1, A. GUPTA

1,

P. K. AGARWAL1 and J.C. BHAT

1

SUMMARY

Genetic divergence was studied for yield and different yield

contributing traits in 282 black soybean accessions collected from

different eco geographic regions of the world. Based on non-

hierarchical euclidean cluster analysis all accessions were grouped into

9 clusters. The first four principle component axes (PCA) accounted for

70.28% of total variance. Minimum mean value for days to 50%

flowering (46), plant height (82.75 cm) and days to maturity (121) were

obtained in cluster I. So this cluster can be very useful to develop early

maturing genotypes. Cluster IX contained eight accessions showed

maximum mean value for pod length (4.26 cm), 100 seed weight (14.87

g) and seed yield per plant (8.09 g). So, from yield point of view, this

cluster can be used to develop high yielding as well as high grain

weight genotypes. Inter cluster distance was found maximum between

cluster IV and IX (6.39). Hence genotypes from these clusters can be

used in hybridization to get desirable recombinants. Accessions VBS

25, VBS 48 from cluster VII and VBS161, VBS152 from cluster VIII

found as exceptionally superior donor which can be used in multiple

crossing programmes to get transgressive segregants for desirable traits.

Key words: Black soybean (Glycine max (L.) Merrill.), principal

component axes, cluster analysis, genetic diversity.

INTRODUCTION

Black soybean (Glycine max (L.) Merrill.) is an important food crop of northern

India especially Uttarakhand. Black seed-coat soybean, locally known as Bhat /

Bhatmash, is grown in Kumaon region and in its bordering states and countries in

the Himalayas (Shah, 2006). It is believed that soybean was brought in via Burma

by traders from Indonesia. As a result, it has been traditionally grown on a small

scale in H.P., Kumaon hills of Uttarakhand, Eastern Bengal, Khasi hills and parts

of central India (Singh, 2006).

1 Vivekanda Parvatiya Krishi Anusandhan Sansthan (ICAR), Almora 263 601

*Corresponding Author: E-mail: [email protected]

(Revised online version June, 2012)

82

The selection practiced in black soybean over the years by hill farmers

had led to the evolution of a genotype with high nutritional quality and low anti-

nutritional factors. The protein deficiency disease kwashiorkor, which is

otherwise prevalent in India, was rarely seen among children in Kumaon Hills.

This may be due to regular consumption of black soybean in their daily meals in

one form or the other (Hymowitz, 1969). Black soybean produced in hills has

better taste and nutritive value, fetches higher market prices and also have huge

market potential. Traditionally cultivated black soybean (bhat) is much lower

than normal soybean in yield which can be improved by involving diverse

germplasm of exotic as well as indigenous black and normal soybean in breeding

programme.

Genetic improvement can be achieved only if the knowledge about

magnitude of genetic divergence is available for agronomically important traits.

Role of genetic diversity for crop improvement programme has been emphasised

by Joshi and Dhawan (1966). Therefore to ensure the successful genetic

improvement, it is logical as well as essential to estimate the genetic variability

for planning the hybridization programme. The present study aimed at to

determine the extent of variability in indigenous as well as exotic black soybean

accessions collected from different eco geographic regions of the world and to

identify agronomically important promising accessions for utilization in future

breeding programmes.


Experimental material consists of 282 germplasm lines of black soybean

collected from India (153), Taiwan (26), Philippines (48) and USA (55). These

germplasm lines were evaluated during Kharif season with three checks VL Soya

1, VL Soya 47 and VL Bhat 65 in augmented block design at VPKAS

experimental farm, Hawalbagh, Almora (29035’N and 79039’E at an elevation of

1250 m.s.l.) under mid hill condition. Experimental design with three checks

repeated systematically after every 10 lines of test entries. The checks were

distributed evenly in each block. Each germplasm and check was grown in rows

of 3 m length and spaced 45 cm apart. Distance of plants within a row was

maintained at 10 cm. All the recommended cultivation measures were followed

to raise a good crop.

Observations were recorded for 11 quantitative and qualitative traits

each. Data on number of primary branches per plant, pods per plant and grain

yield per plant were taken on five randomly selected plants. Pod length and

number of seeds per pod were recorded on ten pods selected randomly from the

harvest of five plants. 100 seeds were selected randomly and weighed. Days to

50% flowering and days to maturity were recorded on plot basis. The mean,

range, coefficient of variation for each quantitative character were calculated for

282 accessions using standard statistical procedures. Principle component

analysis (Rao, 1984) and correlation coefficient between principle component

axes scores and adjusted mean values of the traits were computed. Genetic

divergence was computed using non hierarchical euclidean cluster analysis

(Spark, 1973).

83


In the present study, moderate to high variability was observed for qualitative

traits in 282 accessions of black soybean. Frequency distribution of traits (Table

1) showed that accessions were predominantly tall (0.90), violet flowered (0.91),

green leaf colour (0.76), light tawny pubescence (0.99), spherical seed shape

(0.64), small seed size (0.52), resistance to pod blight (0.73) and tolerance to frog

eye leaf spot (0.52). Transformation of quantitative traits into principle

component yielded eleven Eigen vectors and Eigen roots. First Eigen vector

having highest Eigen value 3.265 and only first four of the eleven principle

component axes (PCA) had given Eigen value more than one. Only principal

component with Eigen value more than one were considered in determining the

agro morphological variability in accessions (Kaiser, 1960). The first four

principle component axes (PCA) together accounted for 70.28% of total variance

and PCI, PCII, PCIII and PCIV accounted for 27.21%, 22.52% 11.63% and

8.92% of total variation respectively. PCA I mainly correlated to traits namely

days to maturity, number of pods per plant, and 100 seed weight. PCA II mainly

defined by seed yield per plant, number of primary branches per plant and oil

content. PCA III was determined by days to 50% flowering, number of pods per

plant, plant height and days to maturity. PCA IV was correlated with 100 seed

weight, oil content and days to maturity. PCA I, III and IV were positively

correlated with days to maturity and PCA I, II and IV were positively correlated

with 100 seed weight. (Table 2).

Maximum range was observed for quantitative traits (Table3) viz., plant

height (58-195 cm) followed by number of pods per plant (11-81) and days to

50% flowering (34-77 days) while trait pod length (3.10-4.70 cm), number of

primary branches (3-5) and number of seeds per pod (2-3) showed lowest range

value. A wide range in oil content (8.55-21.32 %) was also observed. Protein

content (31.42-46.66 %) was observed which is within the range (29.2–51.2 %)

reported by Hafez (1983) and Kakade et al. (1972) of different soybean cultivars.

Coefficient of variability was highest for seed yield per plant (58.31 %) followed

by number of pods per plant (26.19 %) and lowest for days to maturity (4.55 %).

The variation within groups expressed as percent coefficient of variation, gives

rough estimate of diversity and variability within groups (Fruman et al., 1997).

For traits viz., days to maturity, pod length, and protein content percent

coefficient of variation was observed less than 10 percent hence it is suggested

that these traits having less variability can be improved through introgression of

variability from other sources. The rest of the traits ranged in CV value from

12.55 % for days to 50 % flowering to 58.31 % for seed yield per plant. Genetic

variability is of great interest to the plant breeder as it plays a vital role in framing

a successful breeding programme (Mehetre et al., 1994). Presence of sufficient

amount of variability for traits like number of primary branches per plant, number

of pods per plant, plant height and seed yield per plant suggest improvement of

these traits through selection.

84

Table 1. Class partition and frequencies of 11 non metric traits for 282

accessions of black soybean.

Qualitative trait Classes Frequencies

Plant growth habit Tall

Medium

Dwarf

0.90

0.10

0.00

Early plant vigour Poor

Good

Very good

0.28

0.54

0.18

Flower colour White

Violet

Dark purple

0.09

0.91

0.00

Leaf colour Green

Light green

Dark green

0.76

0.08

0.16

Leaf size Broad

Medium

small

0.27

0.62

0.11

Pubescence colour Grey

Tawny

Light tawny

0.01

0.00

0.99

Seed shape Spherical

Spherical-flattened

Elongated

Elongated-flattened

0.64

0.32

0.00

0.04

Seed size Small (<10g)

Medium(10-13g)

Large(> 13g)

0.52

0.42

0.06

Lodging

susceptibility

None

Low

Medium

high

0.10

0.28

0.36

0.24

Resistance to pod

blight

Highly resistant

Resistant

Moderately resistant

Tolerant

Susceptible

Highly susceptible

0.00

0.73

0.26

0.01

0.00

0.00

Resistant to frog eye

leaf spot

Highly resistant

Resistant

Moderately resistant

Tolerant

Susceptible

Highly susceptible

0.00

0.21

0.21

0.52

0.06

0.00

85

Table 2. Correlation coefficients between the first four principal components

(PCI, PCII, PCIII & PC IV) and 11 quantitative characters of black

soybean accessions.

*, **: Significant at 5% and 1% probability levels, respectively.

Characters PC I PC II PC III PC IV Eigen

root

Days to 50% flowering 0.12* 0.096 0.797** 0.178** 3.265

Plant height (cm) -0.01 -0.205 0.692** -0.386** 2.702

No. of primary branches

per plant 0.153** 0.769** 0.066 -0.166**

1.395

Days to maturity 0.251** 0.129 0.374** 0.322** 1.07

Pod length(cm) 0.018 0.14 -0.233** 0.007 0.902

Oil content (%) 0.114 0.301** 0.175** 0.615** 0.734

Protein content (%) 0.159** -0.031 -0.278** -0.205** 0.641

No. of seeds per pod -0.025 -0.019 -0.03 0.037 0.571

100 seed weight (g) 0.271** 0.389** -0.19** 0.561** 0.455

Seed yield per plant (g) 0.164** 0.876** -0.09 -0.269** 0.19

No. of pods per plant 0.384** 0.291 0.292** 0.02 0.057

86

Non hierarchical euclidean cluster analysis was found useful in

estimating the genetic diversity on the basis of similarity in agro-morphological

traits in 282 black soybean accessions. Estimation of genetic divergence through

cluster analysis was emphasized by Murty and Arunachalam (1966). Sapra et al.

(2006) emphasised the use of non hierarchical algorithms for clustering of

germplasm and sampling within discrete groups for maximising diversity in

soybean. On the basis of F test 9 clusters were found to be more suited for this

study (Table 4). Cluster II consisted of maximum (67) accessions followed by

Cluster IV, VI, VII which consisted 48, 38, 36 accessions respectively.

Distribution of genotypes into different clusters, suggested the presence of

substantial genetic divergence among the germplasm and indicated that this

material may serve as a good source for selecting the diverse parents for

hybridization programme, aimed at isolating desirable recombinants for seed

yield as well as other characters (Raje and Rao, 2001). Average distances of

clusters from cluster centroids ranged from 2.32 to 2.97. It was found minimum

in cluster IV and maximum in cluster I. It suggested that genotypes in cluster I

were relatively more diverse among themselves, however, in most of the cases, the inter cluster distances were greater than intra cluster distances implying

greater degree of genetic diversity between the genotypes of the two clusters than

the genotypes present within a single cluster. Similar result has been reported by

Tyagi et al. (2008). So far as inter cluster distance is concerned, cluster IV and IX

centroid were the farthest (6.39) from each other i.e. these two clusters are

genetically more diverse followed by cluster I-VIII (5.928) and cluster VII-IX

(5.81). Greater the diversity between two clusters wider the genetic diversity

between genotypes (Mian and Bahl, 1989). Hence selection of genotypes should

be done from the two clusters with wider inter cluster distance to get more

variability and heterotic effect (Pradhan and Roy, 1990). The genotypes belong to

distant clusters with desirable agro-morphological traits can be used to make

multiple crosses and genes for desirable traits can be transferred to a common

genetic background. Minimum inter cluster distance was observed in cluster VII

and II (1.86) (Table 4). It may be due to unidirectional selection pressure

practiced by hill farmers to meet certain quality parameters in past might have

resulted in limited divergence with similar features among the genotypes.

Cluster mean for various characteristics exhibited significant differences.

Minimum mean value for traits (Table 5) viz., Days to 50% flowering (45.5),

plant height (82.75 cm) and days to maturity (121.33) were obtained in cluster I

which included twelve genotypes. So accessions from this cluster can be very

useful to develop early maturing genotypes. Cluster IX contained eight

accessions showed maximum mean value for pod length (4.26 cm), 100 seed

weight (14.87 g) and seed yield per plant (8.09 g). Considering yield, this cluster

can be used to develop high yielding as well as high grain weight genotypes.

Cluster I and IX can be used in hybridization to get high yielding and early

maturing recombinants. Similarly, cluster VII for plant height (144.19 cm),

cluster VIII for number of primary branches per plant (4.09), cluster VI for

number of pods per plant (59.95), Cluster V for oil content (17.79%) and protein

content (42.84%) had maximum mean values. The rest of the clusters have

moderate mean values for these traits.

87

Table 3. Estimates of range, mean, and CV (%) for 11 quantitative characters in

black soybean accessions

Table 4. Intra (diagonal) and inter-cluster distances D values for 11 quantitative

traits in black soybean accessions.

Characters Mean Minimum Maximum SD CV (%)

Days to 50% flowering 68.04 34.00 77.00 8.54 12.55

Plant height (cm) 131.72 58.00 195.00 26.59 20.19

No. of primary

branches per plant 3.38 3.00 5.00 0.52 15.43

No. of pods per plant 51.14 11.00 81.00 13.40 26.19

Days to maturity 129.28 116.00 139.00 5.88 4.55

Pod length(cm) 3.47 3.10 4.70 0.30 8.77

Oil content (%) 15.34 8.55 21.32 2.17 14.13

Protein content (%) 39.32 31.42 46.66 2.50 6.36

No. of seeds per pod 2.65 2.00 3.00 0.48 18.12

100 seed weight (g) 10.18 5.44 18.21 2.08 20.38

Seed yield per plant (g) 4.32 0.12 13.75 2.52 58.31

Cluster I II III IV V VI VII VIII IX

I 2.968

II 4.468 2.404

III 3.424 2.827 2.884

IV 4.776 2.493 3.085 2.316

V 5.237 3.517 2.922 3.686 2.793

VI 5.677 2.862 3.817 4.177 3.011 2.635

VII 4.169 1.856 2.559 2.75 3.513 3.87 2.382

VIII 5.928 3.594 4.004 3.258 2.934 2.65 4.34 2.738

IX 5.462 5.323 5.042 6.385 4.94 4.177 5.813 5.361 2.773

88

Table 5. Cluster means for 11 quantitative traits in 282 black soybean

accessions

Cluster Days to

50 %

flower-

ing

Plant

height

(cm)

No. of

primary

branches

per plant

No. of

pods

per

plant

Days to

maturity

Pod

length

(cm)

Oil

content

(%)

Protein

content

(%)

No. of

seeds

per

pod

100

seed

weight

(g)

Seed

yield

per

plant

(g)

I 45.5 82.75 3.08 38.75 121.33 3.52 13.96 38.65 2.92 10.43 1.53

II 71.54 140.22 3.19 57.7 130.48 3.53 15.13 38.89 3.00 10.20 3.11

III 56.48 135.12 3.12 43.16 125.8 3.54 13.85 40.98 2.60 8.45 4.67

IV 72.15 138.23 3.02 48.54 130.52 3.2 14.85 39.29 2.00 10.34 2.68

V 65.79 129.64 3.64 43.5 126.43 3.35 17.79 42.84 2.57 9.62 7.06

VI 72.08 119.34 4.00 59.95 132.61 3.67 15.39 39.5 3.00 10.82 6.77

VII 69.81 144.19 3.03 38.83 127.61 3.56 16.24 38.3 2.94 8.50 2.87

VIII 69.29 131.18 4.09 58.53 131.47 3.21 15.97 38.67 2.03 11.31 6.53

IX 55.38 92.88 3.50 53.25 121.88 4.26 15.53 39.43 3.00 14.87 8.09

89

Table 6. List of promising accessions for different quantitative characters.

Characters Promising

accessions

Cluster

number

Country

Days to 50% flowering

(<40 days)

VBS 25

VBS 48

VBS 102

VBS 152

VBS 161

VII

VII

VII

VIII

VIII

USA

USA

USA

USA

India

Plant height (<62cm) VBS152

VBS161

VIII

VIII

USA

India

Number of pods per plant

(>76)

VBS74

VBS 9

VBS15

VBS132

II

I

I

VI

India

India

India

Philippines

Days to maturity (<118 days) VBS25

VBS29

VBS48

VBS66

VII

VII

VII

VIII

USA

USA

USA

USA

Pod length (>4.5 cm) VBS 237 III India

Oil content (>21 %) VBS 269

VBS 185

VBS 184

VI

VIII

II

India

USA

India

Protein content (>45 %) VBS 198

VBS 121

VBS 314

VBS 302

VBS148

II

II

IV

IV

II

USA

India

India

India

India

100 Seed weight (>16.5 g) VBS320

VLS1

VLS 47

IV

IV

V

India

India

India

Seed yield per plant (>12 g) VBS 334

VBS285

VBS 342

VBS 175

IV

IV

IV

VIII

India

India

India

USA

90

Majority of promising accessions (Table 6) for different traits were in

cluster VIII and cluster IV and originated in India and USA. Accession VBS 25

and VBS 48 belongs to cluster VII had least values for days to 50% flowering

and days to maturity i.e. these are early maturing accessions. These accessions

can be used in a crossing programme to transfer early maturity to high yielding

background. Similarly, accession VBS161 and VBS152 belong to cluster VIII

had least days to 50% flowering and plant height. Promising genotypes belonging

to different clusters identified in this study can be used as donor for different

agro- morphological traits to develop suitable genotypes for hills of Uttarakhand.

REFRENCES Fruman B.J., C.O. Qualset, B. Skoumand, J.H. Heaton, H. Corke, D.M. Wesenberg.1997.

Characterization and analysis of North American Triticale genetic resources.

Crop Sci. 37:1951-1957.

Hafez, Y.S. 1983. Nutrient composition of different varieties and strains soybean. Nutr.

Rep. Int. 28: 1197–1206.

Hymowitz, T. 1969. The soybean of the Kumaon Hills of India. Economic Botany. 23(1):

50-54.

Joshi A. B. and S.L. Dhawan. 1966. Genetic improvement in yield with spaceal reference

to self fertilizing crops. Indian J. Genet. 26A: 101-131.

Kaiser H.F. 1960. The application of electronic computers to factor analysis. Educ.

Psychol. Meas. 20:141–151.

Kakade, M.L., N.R. Simons, I.E. Liener and J.W Lambert. 1972. Biochemical and

nutritional assessment of different varieties of soybean. J. Agric. Food Chem.

20, 87–90.

Mehetre S.S., C.R. Mahajan, P.A. Patil and D.N. Hajare.1994. Genetic divergence in

soybean (Glycine max). Indian J. Genet.54 (1):83-8.

Mian M. A. K. and P.N. Bhal. 1989. Genetic divergence and hybrid performance in chick

pea. Indian J. Genet.49 (1):119-124.

Murty B. R. and Y. Arunachalam. 1966. The nature of genetic divergence in relation to

breeding system in crop plants. Indian J. Genet.26A: 188-98.

Pradhan, A.K. and Roy, A. 1990. Genetic divergence in rice. Oryza, 27 (4): 415-418.

Raje R. S. and S.K. Rao. 2001. Genetic diversity in germplasm collection of mungbean

(Vigna radiata L. Wilczek). Indian J. Genet. 61(1): 50-52.

Rao C.R. 1984. The use and interpretation of principal component analysis in applied

research. Sankhya. 26:329-358.

Sapra R.L., S.K. Lal, Akshay Talukdar and K.P. Singh. 2006. Selecting accessions in

Soybean collection with high diversity. Indian J. Plant Genet. Resources. 19(2):

171-174.

Shah, N.C. 2006. Black soybean: An ignored nutritious and medicinal food crop from the

kumaon region of India. Asian Agri-Hist. 10, 33–42.

Singh B.B. 2006. Success of soybean in India: The early challenges and pioneer

promoters. Asian Agri-Hist. 10, 43–53.

Spark D.N. 1973. Euclidean cluster analysis algorithm. Applied Statistics. 22: 126-130.

Tyagi J.P., Tejbir Singh, Sanjay Singh and V.P. Singh. 2008. Genetic diversity among

indigenous land races of rice for certain quality traits. J. Plant Genet. Resources

21(1):28-30.

Volume 43 No. 1 June 2011

Documents

Transcript of Volume 43 No. 1 June 2011