WHOLE GENOME RESEQUENCING OF AN ALMOND GERMPLASM COLLECTION ALLOWS HIGH-RESOLUTION GWAS AND THE IDENTIFICATION OF CANDIDATE GENES FOR KEY ECONOMIC TRAITS

Almond [Prunus dulcis Miller (D.A. Webb)] is the main tree nut species worldwide for cultivated area and economic importance. Recently, three almond genome assemblies (approximate size 240 Mbps) were released, thereby prompting genome-wide association studies (GWASs) aiming to the identify marker-trait associations. Notably, all of these studies were performed using low to medium single nucleotide polymorphism (SNP) marker densities obtained by chip arrays or genotyping by sequencing (GBS), which can be insufficient to reveal associations in a species, such as almond, displaying an extremely fast linkage disequilibrium (LD) decay. Here, we carried out a massive effort to perform whole-genome resequencing of 104 almond cultivars belonging to the ex-situ collection of the Italian Council for Agricultural Research and Analysis of Agricultural Economics—Section Agriculture and Environment (CREA-AA). Mapping of reads against the cv. NonPareil reference genome yielded as many as 2,224,934 high-quality SNPs. GWAS was performed merging SNP information with multi-year phenotypic data collected for main economically important traits (kernel, nut and shell weight, kernel yield, twin/shrivelled kernel percentage and flowering time). An extremely fast average genomic LD decay was estimated (70 bps), indicating that marker loci identify by our GWAS reside within or in close proximity of genes underlying phenotypic variation. Based on functional information retrieved in literature, genes encoding two aspartyl proteases, an AGAMOUS homolog and a RICESLEEPER 2 homolog were deemed as obvious candidate for kernel weight and flowering time. Overall, this study contributes to significant advances in almond genetics and provides a basis for the implementation of molecular breeding approaches.