Meanwhile, genomic data from 27 diverse maize inbred lines showed that the genome consists of highly divergent haplotypes with 10- to 30-fold variations in recombination rates [27]. This reinforces the concept that maize is a highly polymorphic species. However, it also shows that there are often large genomic regions that have little or no variation [28]. Much valuable LDE225 in vivo information likely underlies the genome

signature due to selection that can be exploited for breeding. The objectives of this study were to (i) confirm the genetic locus for cob glume color using a genome wide association study (GWAS) with high resolution SNPs, (ii) reveal the genome pattern surrounding it, and (iii) find find more evidence of the effects of selection across the target region. The results reported here may provide insights as to the manner by which breeding efforts have affected and will affect genome evolution.

A set of 283 diverse inbred lines, representing the modern temperate maize elite inbred lines in China [29] and [30], was used for genotyping with 55,000 SNPs and GWAS. Forty of the lines from this association panel and 47 tropical lines with white cob glumes were re-sequenced 10 × through an international collaboration (Xu et al., in preparation). These plant materials were grown in Sanya, Hainan province (18°45°N, 109°30°E), during the winter of 2011–2012. Each line was planted in a plot with 20 plants in a 4.5 m row with 0.6 m spacing between rows. Normal agronomic practices were used in field management. After harvest, cob glume color was

scored for each line as “0” for white and “1” for other colors. The scores were used for GWAS. Based on the B73 reference sequence, 56,110 evenly spaced SNPs were featured on the MaizeSNP50 BeadChip (Illumina, Inc.). These were selected from several public and private sources and included 984 negative controls. DNA was extracted from the 283 temperate lines by a modified CTAB procedure Bcl-w [31]. Before genotyping, each DNA sample was evaluated using gel-electrophoresis and spectrophotometry (NanoDrop 2000, Thermo Scientific). As controls, four lines (Qi 319, Huangzao 4, Ye 478 and Dan 340) were added to each of the six independent BeadChips. SNP genotyping was performed using the MaizeSNP50 BeadChip by Emei Tongde (Beijing, China). SNP calling for the 283 samples was implemented according to the Infinium HD Assay Ultra Protocol Guide (Illumina, Inc.). After filtering out monomorphic and non-specific SNPs, a subset of 44,235 SNPs with known physical positions was generated, with an average heterozygosity of 0.5%. Within the four controls, the mean reproducibility between replicates across all data points was 99.9%, which is consistent with high-quality data for replicates of the B73 maize line using the same chip (Illumina, Inc.). The error rate (ER) for genotyping was 0.0353%.