Influence of population size, genetic maps, and environments on QTL detection and mapping in sorghum

Most economically important traits in field crops such as sorghum [ Sorghum bicolor (L.) Moench] have complex inheritance with significant influence of environments. Selection for such traits can be enhanced through use of molecular markers strongly associated with traits of interest. Quantitative trait loci (QTL) detection and mapping to pinpoint genomic regions associated with the traits are essential requisites to design an effective breeding strategy. However, several factors can affect efficient QTL mapping. The objectives of this study were to evaluate major factors affecting QTL mapping including population size, genetic linkage maps and environments. Joint-trait analysis was explored for its efficiency in QTL mapping using empirical data from multi-environment testing of a large population. Recombinant inbred lines (RI lines) were developed from an intra-specific cross between SRN-39, an African caudatum, and Shan Qui Red (SQR), a Chinese kaoliang line. A total 528 RI lines and the two parents were evaluated for several agronomic traits: Striga resistance, seedling vigor, flowering, physiological maturity, plant height, grain yield and yield components (kernel number, kernel weight, and plant population). The experiment was conducted in 2005 and 2006 at two planting dates (early and late) at the Purdue University Agronomy Center for Research and Education (ACRE), West Lafayette, Indiana. A Striga lab experiment was also undertaken in 2007 at Purdue University parasitic weed containment facility. Population size had significant impact on the number of QTL mapped, additive genetic effects expressed, and the genetic variations explained by QTL (R 2). Overall, increase in population size resulted in increased number and power of QTL detection, though exceptions were noted. Genetic linkage maps had no significant effect on the number of QTL detected. However, maps showed influence on the additive genetic effects expressed and the variation explained by QTL (R2). The power of QTL detection tended to be reduced when genetic maps developed from the largest sample population were used indicating that genetic maps estimated from small sample size can be used for the initial QTL mapping study. Results also indicated that environments had significant effect on the number of QTL detected, estimates of additive genetic effects, and variation explained by QTL (R2). Discrepancy varied from 0 to 33% in the number of QTL detected across traits and environments, and this variation tended to depend on the heritability of traits. Joint-trait analysis substantially improved the total number of QTL discovered. A number of QTL with high likelihood ratio (LR) of QTL-by-environment interaction (QEI) were detected showing high instability of a number of QTL segregating in the population studied across diverse environments. Several genomic regions having likely pleiotropic effects on sorghum grain yield were mapped for future studies targeting improvement of sorghum grain yield