ABSTRACT
Cultivated maize was domesticated from teosinte (Doebley 2004), through selection against certain undesirable characteristics, such as branching pattern or glume architecture (Wang et al. 1999; Wang et al. 2005), followed by many generations of selection for improved agronomic performance, including increased yield and improved abiotic and biotic stress tolerance. Much of the selection during the past 50 or so years took place within private seed companies, which developed their own advanced germplasm, initially taking advantage of the public breeding efforts. This resulted in some decrease in the available genetic diversity of the maize gene pool (Hilton and Gaut 1998; Duvick et al. 2004a,b; Smith et al. 2004; Wright et al. 2005). Ongoing maize breeding for heterotic yield and other desirable agronomic and output traits is expected to continue to change the patterns of genetic diversity of cultivated maize. Numerous investigators evaluated genetic diversity at multiple loci (Tenaillon et al. 2001; Buckler et al. 2006; Vigouroux et al. 2008; Van Inghelandt et al. 2010), but a high-resolution description of genetic diversity as a function of genetic or physical map position has been lacking. As part of the platform development for genetic association mapping for maize (Beló et al. 2007) we genotyped a diverse collection of over 700 maize inbreds consisting of current elite materials as well as their ancestors. This genotypic data was successfully used for genomewide association study for a quantitative trait (Beló et al. 2007). Here we present an analysis of the relationship between single nucleotide polymorphism (SNP) diversity and genome location in the context of heterotic groups. We also examine changes in diversity of cultivated maize with time. Recent studies (Springer et al. 2009; Beló et al. 2010; Lai et al. 2010) highlighted the prevalence of copy number variants, including large genic deletions in maize. While these genomic sequence variants, especially those including expressed genes or regulatory regions are likely to affect trait expression, most of them are ancient relative to plant breeding of the last 100 years or so, and expected to be in linkage disequilibrium (LD) with surrounding SNPs. Therefore their inclusion in the analysis would be unlikely to change the overall results.
