Genome Wide Association Mapping in Watermelon

Genotyping by sequencing was carried out following the protocol standardized by the institute of genomic diversity at Cornell University. Briefly, genome complexity was reduced by digesting total genomic DNA from 280 watermelon samples with ApeKI methylation sensitive restriction enzyme. Digested products were used to make GBS “library” for sequencing on the Illumina® HiSeq 2500. Chromosomal assignment and position for 14526 SNPs was deduced from the Whole Genome Sequence (WGS) draft of watermelon archived at http://www.icugi.org. Mapped SNPs were further filtered by successful call rate and Hardy-Weinberg equilibrium (HWE). Only SNPs with a minimum minor allele frequency > 0.1 were considered for LD studies and candidate gene association mapping. LD was estimated as the r² between all the pairs of SNPs from 280 diverse watermelon cultivated accessions using the SNP & Variation Suite v7.7.6 (SVS). Using the default settings of SVS, 362 SNP Haplotype were identified. Adjacent and pairwise measurements of LD for GBS data were calculated separately for SNPs in each individual chromosome. Principal components was calculated from 5,231 LD-pruned SNPs by the genotypic principal component analysis using SVS and used to control population structure while performing Genome Wide Association Study (GWAS). GWAS with 5,231 SNPs was performed using Linear Regression, EMMAX and MLMM methods as implemented in the SVS. Manhattan plots for associated SNPs were visualized using GenomeBrowseTM v1 by Golden Helix, Inc. The SNP P-values obtained from GWAS were subjected to sequential Bonferroni correction as well FDR. GWAS results for fruit length, width, rind thickness and soluble solids obtained for two consecutive seasons will be discussed.