Selective Genotyping to Identify Soybean Seed Protein/Oil QTLs in 48 High Protein Germplasm Accessions

Selective Genotyping (SG) describes a QTL mapping method in which all population members are phenotyped, but only the most informative individuals - those in the lowest and highest tails of a phenotyped trait distribution - are genotyped.  In this research, 48 F2 soybean populations of about 250 F2 plants were derived from mating a high protein germplasm accession with an agronomic cultivar of ordinary protein.  The F2.3 seed progenies were phenotyped for seed protein/oil content using a NIR method, and the 22 highest and 22 lowest F2.3 progenies for seed protein were selected in each population for genotyping with 1536 SNPs.  Using R/qtl software for QTL mapping, allelic segregation for a high/low protein phenotype was assessed with a SG-stratified 1900 permutation-test-derived genome-wide (alpha=0.05) LOD score significance criterion.  The well-known QTL located on Chr 20 (LG-I) was detected as the most highly significant QTL in 25 of the 48 populations (and it was detected at a LOD score >3.0 in 10 more of the 48).  QTLs on Chr 15 (LG-E), Chr 10 (LG-O), Chr 6 (LG-C2) were detected with moderate frequency (8-12 of the 48), with QTLs on Chr 18 and Chr 19 detected with a lower frequency (4-7), and other QTLs detected in only a few cases (1-3). SG-based surveys of accessions for a given trait to determine the QTL allele status in multiple accessions appears feasible, so long as the trait is highly heritable (oligenic) and is inexpensive to phenotype.  A major finding in this research was the discovery that 17 of the 48 high protein germplasm accessions were 99.9% identical in terms of their SNP genotypes, despite differences in geographical origin and land race names.  This finding has now been confirmed, based on an unpublished 50K SNP chip genotyping of those 17 parents.