An evaluation of inbreeding methodologies and molecular markers associated with developing food-grade soybeans.

摘要

Soybean [Glycine max (L.) Merr.] production in the U.S. is primarily aimed at animal feed and industrial processes, although there is a growing market for soy-based food products directed at human consumption. Desirable soybean cultivars used to make soy foods such as tofu, soymilk, and edamame generally have the following seed compositional characteristics; high protein content, large seed weight, yellow seed coat and hilum color, high sugar content, and proper seed texture. Soybean breeders utilize molecular markers, associated with quantitative trait loci (QTL), for marker assisted selection in their breeding programs to enhance efficient selection of quantitatively inherited traits. The objectives of this research was to determine the most effective and efficient inbreeding methodology for developing large-seeded, high-protein, and high-sucrose soybean cultivars, to validate previously reported genomic regions with QTL associated with seed protein concentrations in a F4-derived recombinant inbred line (RIL) population in Arkansas, and to identify QTL associated with soybean seed sucrose in a F4-derived RIL population. The population R95-1705 (high-protein) x MFL552 (moderately high-sucrose) was used for developing five sub-populations based on inbreeding and selection method including mass selection (Mass), bulk harvest (Bulk), pedigree breeding (PED), single-seed descent (SSD), and single-pod descent (SPD). Soybean plant selections from each sub-population were grown in a replicated strip-plot design with a complete randomized block arrangement during the summer of 2008 at Stuttgart, AR and Keiser, AR. Overall, the success of the SSD and SPD inbreeding methods was unmatched for all traits when compared to the Mass, Bulk, and PED inbreeding methods. The SSD and SPD sub-populations consistently had higher average means, both before and after selection, and higher calculated genetic gain per cycle and per year. These results are also consistent with theoretically predicted values of genetic gain and genetic variance for these methods. A high-protein, MG V Arkansas germplasm, R95-1705, was crossed with a MG V, Round-up ReadyRTM germplasm, Hutcheson-RR, to form a segregating RIL population for protein QTL identification and mapping. This RIL population was grown in a randomized complete block design (RCBD) with two replications, at Fayetteville, AR and Stuttgart, AR, during the summers of 2007 and 2008. Composite interval mapping for the SSR markers on LG I revealed that the putative QTL associated with seed protein in this population lies on an approximate 3 cM region between flanking SSR markers Satt614 and Sat_219. A moderately high-yielding, MG V germplasm, S97-1688 was crossed with a moderately high-sucrose MG III plant introduction, P1243545, to form a segregating RIL population consisting of 121 individuals for sucrose QTL identification and mapping. The 121 F4:6 RIL and the parents were grown in single-row 3.1 m plots, with a seeding rate of 33 seed per meter, in a randomized complete block design (RCBD), with two replications at two locations during the summers of 2007 and 2008. Composite interval mapping showed that a putative QTL associated with seed sucrose on MLG Dlb in this population lies on an approximate 4 cM region between flanking SSR markers Sat_069 and Sat_183 and explains approximately 11% of the variation in seed sucrose. These results will help improve the genetic understanding of the inheritance of protein, sucrose, and seed weight, as well as increase the efficiency of current inbreeding methodology and selection when marker assisted selection is desired.