Design of a 9K illumina BeadChip for polar bears (Ursus maritimus) from RAD and transcriptome sequencing

Abstract

Single-nucleotide polymorphisms (SNPs) offer numerous advantages over anonymous markers such as microsatellites, including improved estimation of population parameters, finer-scale resolution of population structure and more precise genomic dissection of quantitative traits. However, many SNPs are needed to equal the resolution of a single microsatellite, and reliable large-scale genotyping of SNPs remains a challenge in nonmodel species. Here, we document the creation of a 9K Illumina Infinium BeadChip for polar bears (Ursus maritimus), which will be used to investigate: (i) the fine-scale population structure among Canadian polar bears and (ii) the genomic architecture of phenotypic traits in the Western Hudson Bay subpopulation. To this end, we used restriction-site associated DNA (RAD) sequencing from 38 bears across their circumpolar range, as well as blood/fat transcriptome sequencing of 10 individuals from Western Hudson Bay. Six-thousand RAD SNPs and 3000 transcriptomic SNPs were selected for the chip, based primarily on genomic spacing and gene function respectively. Of the 9000 SNPs ordered from Illumina, 8042 were successfully printed, and - after genotyping 1450 polar bears - 5441 of these SNPs were found to be well clustered and polymorphic. Using this array, we show rapid linkage disequilibrium decay among polar bears, we demonstrate that in a subsample of 78 individuals, our SNPs detect known genetic structure more clearly than 24 microsatellites genotyped for the same individuals and that these results are not driven by the SNP ascertainment scheme. Here, we present one of the first large-scale genotyping resources designed for a threatened species.

Assessing polar bear (Ursus maritimus) population structure in the Hudson Bay region using SNPs

“…SNP genotypes were obtained using a custom‐designed (Malenfant, Coltman & Davis (2015)), 9K Illumina Infinium BeadChip (Illumina, San Diego, CA, USA) processed by Delta Genomics (Edmonton, Canada). SNP genotypes were called using GenomeStudio 2011.1 (Genotyping Module 1.9; Illumina).…”

“…Individuals with call rates of <0.9 were removed. Loci on the SNP chip were derived from both transcriptome and RAD sequencing (Malenfant, Coltman & Davis (2015)). Our focus and interest was to detect patterns of variation at likely neutral markers so only RAD loci were used for this study.…”

“…Additional loci were removed with low minor allele frequencies (<0.01), high rates of missing data (>0.05) and those that were in linkage disequilibrium (LD). LD was assessed using a window of 10 SNPs that slides over by 1 SNP to prune at a cutoff of r 2  = 0.5 as implemented in a custom version of PLINK 1.07 (Malenfant, Coltman & Davis (2015), Purcell et al., 2007). There were no known first‐degree relationships among individuals in WH (Malenfant, Coltman, Richardson, et al.…”

“…The few studies that used SNPs used a small number of markers and sampling of only two to four subpopulations (Cronin et al., 2014; Miller et al., 2012); with the exception of Malenfant, Coltman & Davis (2015), who used 5,441 SNP markers, but only 78 individuals from 13 of the 19 worldwide subpopulations, including only 18 individuals from the Hudson Bay region.…”

High genetic variability of vagrant polar bears illustrates importance of population connectivity in fragmented sea ice habitats

“…This would decrease the likelihood that a subpopulation would receive novel genetic variants from immigrants. However, consistent with the high dispersal capability of polar bears, genetic differentiation among the 19 recognized subpopulations is low (Paetkau et al, 1999;Edwards et al, 2011;Miller et al, 2012b;Campagna et al, 2013;Bidon et al, 2014;Cronin et al, 2014;Malenfant, Coltman & Davis, 2015;Peacock et al, 2015). Further, a study that included data from 18 of the recognized subpopulations detected gene flow among clusters of subpopulations (Peacock et al, 2015).…”