Identification of quantitative trait loci (QTL) in maize on the basis of F3 and testcross progeny performance

Abstract

Detection of quantitative trait loci (QTL) is expected to vary depending on the type of progeny evaluated phenotypically. The objective of this paper was to compare the effectiveness of QTL detection using P3 and testcross (TC) progeny evaluation on the basis of estimated QTL positions and magnitude of effects. One hundred eighty-two Fg individuals derived from a B52 X B73 cross were genotyped at 87 segregating RFLP loci. The performance of F3 lines per se and TC progenies, generated by topcrossing F3 lines with the inbred testers, Mol7 and Oh43, was evaluated at two to six environments for grain yield and several other agronomic traits. Interval mapping was used to determine the positions of putative QTL on a linkage map and to estimate their effects. The positions of QTL, as indicated by log-likelihood peaks exceeding a 2.2 threshold, frequently mapped to similar regions (< 20 cM apart) in F3 line and both TC progeny evaluations. Regions with significant peaks detected by Mol7 TC and Oh43 TC evaluation but not F3 line evaluation were also common. One notable difference in QTL detection among the progeny types was for grain yield, where analysis using a single-QTL model detected 5 significant peaks in F3 lines, 8 in Oh43 TCs, and 19 in Mol7 TCs. Subsequent analysis using multiple-QTL models reduced the number of likely QTL sites to four in F3 lines and Oh43 TC, and five for Mol7 TC, however variance accounted for by a multiple-QTL model remained high in Mol7 TC evaluation (65.4) relative to that of Oh43 TC (41.9) and F3 line (37.2) evaluation. These results suggest that both F3 and TC progeny evaluation are effective in detecting QTL for most agronomic traits, although TC evaluation may be superior for grain yield. The effect of tester choice may also be greatest for grain yield and other traits where specific combining ability is important.