Genetic Diversity and Association Analysis of Protein and Oil Content in Food-type Soybean

1. Genetic Diversity and Association Analysis of Protein and Oil Content in Food-type Soybean Ainong Shi, Pengyin Chen, Bo Zhang, and Anfu Hou University of Arkansas

2. Background • Food-type soybean has generated tremendous public interests in various soyfoods • - Large seeds for tofu, soymilk, and edamame • - Small seeds for natto and bean sprouts

3. Few have reported the genetic structure and diversity of food-type soybean. • Seed protein and oil contents are two main quality traits in soybean breeding effort. • Association analysis based on linkage disequilibrium (LD) has recently become an alternative approach to map QTL in plants. • SSR markers have been widely used to analyze genetic diversity and detect QTLs through linkage mapping and association analysis in many crops including soybean.

4. Objective • To analyze genetic diversity in 105 food-type soybeans using 65 SSRs • To conduct association analysis of protein and oil content in soybean

5. Materials and Methods • 105 food-type soybeans collected from Japan, South Korea, and 6 states of USA • 65 SSRs located on 20 soybean MLGs used for genotying in the 105 soybeans • Allele diversity including allele number, allele frequency, gene diversity, and PIC value calculated by PowerMarker 3.25 • AMOVA implemented in Arlequin 3.11 • Genetic structure analyzed by Structure 2.2 • Association tests run with the mixed linear model method in TASSEL 2.0.1

6. Results • Phenotypic and genotypic data • Protein: 42.9% (36.96-50.0%) • Oil: 19.0% (13.8-22.5%) • r = -0.67 (protein-oil) • Allele number: 10 (5-16) • Gene diversity: 0.82 (0.57-0.91) • PIC: 0.79 (0.42-0.89)

7. Genetic Diversity • 105 food-type soybeans divided into three clusters (I, II and III), and further into six groups • Cluster I: S. Korea and Japan • Cluster II: USA lines • Cluster III: S. Korea The genetic background in USA food-type lines were different from those lines from Japan and South Korea.

8. Cluster I: I-S and I-SJ • Cluster II: II-N, II-MK and II-IO • Group I-S: 7 Southern Korea lines • Group I-SJ: 10 South Korea, 3 Japan, and 1 Missouri lines • Group II-N: 3 North Dakota, 2 South Korea, and 1 Virginia lines. • Group II-MK: 8 Missouri, 3 Kansas, and 1 Virginia lines • Group II-IO: 18 Iowa, 12 Ohio, 1 Missouri and 1 Virginia lines • Cluster III: 26 South Korea, 2 Japan, and 2 Virginia. • Two lines, K32 (KS4302sp) and O105 (HF-AG5381) not clustered into any group

9. 18 Iowa + 12 (out of 14) lines clustered into the group II-IO • 3 (4) Kansas + 10 Missouri lines into II-MK • 3 North Dakota lines into II-N • 5 Virginia lines into III, II-MK, II-N, and II-IO. • 5 Japanese lines into I-SJ and III • 46 Korean lines into four groups: • 8 to I-S, 10 to I-SJ, 2 to II-N, and 26 to III

10. AMOVA for geographic origin in the 105 soybeans • The largest variance due to within soybean lines (88.41%) • Little variance due to among countries (less than 1%) • The variance among the six states of USA (10.60%) indicated genetic diversity existed among the states

11. LD parameter R2 was significant for most of the pairwise comparisons among 65 SSR markers. Linkage disequlibrium (LD) value (R2, above diagonal line) and probability value (P, below diagonal line) for 65 markers in 105 soybean lines

12. 19 putative QTLs for protein content (15 previously mapped using linkage analysis and 4 new identified)

13. 13 putative QTLs for oil content (10 previously mapped using linkage analysis and 3 new identified)

14. Summary • 105 soybean lines were divided into three clusters and further clustered into six groups. • A negative correlation was obtained between protein and oil contents (r = -0.67). • 13 SSR markers distributed on 11 MLGs were identified to be significantly associated with oil content (p < 0.001 and R2% = 14.4 – 43.5). • 19 SSR markers distributed on 14 MLGs were identified with protein content (P < 0.001, R2% = 14.3 – 45.6).

15. Thank you!