In this study we assayed the extent of genetic variation for five microsatellites in 94 accessions of wild (Glycine soja) and cultivated soybean (G. max). F₂ segregation analysis indicated that all five of the microsatellites were independently inherited and four loci were located in four independent linkage groups. The number of alleles per microsatellite locus ranged from five to 21. Overall, 43 more microsatellite alleles were detected in wild than in cultivated soybean. Allelic diversity for microsatellite loci was significantly higher in wild than in cultivated soybean.

In a second study, molecular markers were used to identify and characterize quantitative trait loci (QTL) controlling seed weight in soybean, and to extend reports of orthologous seed weight genes in the genus Vigna to the genus Glycine by "comparative QTL mapping". DNA samples from 150 F₂ individuals from an interspecific soybean cross were analyzed with 91 genetic markers. Three and five markers were significantly associated with seed weight variation (P<0.01) in the F₂ and F_2:3 generations, respectively. Two-way ANOVA tests for digenic interactions identified three significant epistatic interactions in both generations. In a combined analysis, the significant marker loci and epistatic interactions explained 50 and 60% of the total variation for seed weight in the F₂ and F_2:3 generations, respectively. Comparison of our results in Glycine with those reported in Vigna indicated that both genera share orthologous seed weight genes. Moreover, a significant epistatic interaction between seed weight QTLs was conserved in both genera.

The objective of the third study was to use molecular markers and interval mapping techniques to position and characterize quantitative trait loci controlling seed protein, oil, sucrose, and calcium content as well as seed weight in soybean. Two QTLs were detected for protein and calcium content, five for oil content and seed weight and six for sucrose content, respectively. Percent phenotypic variation explained by these individual QTLs ranged from 6.6 to 34.0%. The total phenotypic variation explained by all QTLs for specific traits were 42.5%, 36.7%, 49.0%, 53.1%, and 42.6% for seed weight, protein, oil, sucrose, and calcium, respectively. Of the 11 genomic intervals identified in this study, six were associated with more than one seed quality trait. These results suggest that the genetic correlations observed between seed quality traits may be due to a pleiotropic effect of a single QTL or that QTLs controlling different seed quality traits were inherited in clusters as tightly linked loci.