Evolution and association analysis of GmCYP78A10 gene with seed size/weight and pod number in soybean.

Seed-size/weight traits, controlled by multiple genes in soybean, play an important role in determining seed yield. However, the molecular mechanisms controlling the seed size and weight in soybean remain unclear. In Arabidopsis, P450/CYP78A gene family has been proved extremely relevant to seed size (such as AtCYP78A5, AtCYP78A6 and AtCYP78A9). We found that a soybean GmCYP78A10 gene underwent artificial selection during soybean breeding. The GmCYP78A10a allele mainly distributed in wild soybean (Glycine soja), but has been eliminated in the cultivars during early stage of soybean breeding, while the GmCYP78A10b allele has been accumulated and become the predominant allele in cultivated soybean (G. max). ANOVA analysis showed that the mean seed weight, seed width and seed thickness of soybean varieties with GmCYP78A10b allele was significantly heavier/bigger than those with GmCYP78A10a allele (P < 0.01). The allele could explain 7.2 % variation in seed weight. The pod number of the soybeans with GmCYP78A10b allele significantly decreased compared to those with GmCYP78A10a allele (P < 0.01, R(2) = 5.8 %), while other agronomic traits including seed weight/plant were not significantly affected by these two alleles. We speculated that during the early stage of soybean breeding, breeders selected big seed carrying GmCYP78A10b allele, but lowered pod number simultaneously. Overall, the selection did not cause the significantly change in soybean seed yield. Our results suggests that the soybean GmCYP78A10 gene may have a similar function to those genes belonging to P450/CYP78A subfamily in Arabidopsis and provides new information for the genetic control of seed size in soybean.