Recurrent selection (RS) is often used in crop improvement to increase the frequency of favourable alleles. Molecular analysis of allele shifts under RS can reveal quantitative trait loci (QTLs) and allow favourable alleles to be identified from a rich pool. Full-sib S1 RS was conducted by the International Maize and Wheat Improvement Center (CIMMYT) in eight diverse tropical maize populations to improve the level of quantitative resistance to northern corn leaf blight (NCLB). Through four cycles of RS, significant increases in resistance were achieved in each cycle; the mean area under the disease progress curve decreased by an average of 17% per cycle across all cycles and populations.

For one population, we performed a genome-wide scan with simple sequence repeat (SSR) markers to identify loci exhibiting significant differences in allele frequency shifts between the initial (C0) and final (C4) populations. Among the more than 100 loci tested, 10% were found to deviate from expectations attributable to drift, suggesting the effect of selection. Effects of selection could be related to NCLB response but could also be due to acknowledged selection for yield and rust resistance or to unrecognized factors. As an initial effort to identify loci likely to be under selection for NLCB resistance, we compared the results of our analysis with those of a summary of all NCLB QTLs, all disease QTLs reported in maize, and major genes for NCLB resistance (Ht genes). Several loci showing evidence of selection coincided with genomic regions previously associated with quantitative or qualitative resistance or both.