Login / Signup

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

Michelle ViengkoneAndrew Edward DerocherEvan Shaun RichardsonRené Michael MalenfantJoshua Moses MillerMartyn E ObbardMarkus G DyckNick J LunnVicki SahanatienCorey S Davis
Published in: Ecology and evolution (2016)
Defining subpopulations using genetics has traditionally used data from microsatellite markers to investigate population structure; however, single-nucleotide polymorphisms (SNPs) have emerged as a tool for detection of fine-scale structure. In Hudson Bay, Canada, three polar bear (Ursus maritimus) subpopulations (Foxe Basin (FB), Southern Hudson Bay (SH), and Western Hudson Bay (WH)) have been delineated based on mark-recapture studies, radiotelemetry and satellite telemetry, return of marked animals in the subsistence harvest, and population genetics using microsatellites. We used SNPs to detect fine-scale population structure in polar bears from the Hudson Bay region and compared our results to the current designations using 414 individuals genotyped at 2,603 SNPs. Analyses based on discriminant analysis of principal components (DAPC) and STRUCTURE support the presence of four genetic clusters: (i) Western-including individuals sampled in WH, SH (excluding Akimiski Island in James Bay), and southern FB (south of Southampton Island); (ii) Northern-individuals sampled in northern FB (Baffin Island) and Davis Strait (DS) (Labrador coast); (iii) Southeast-individuals from SH (Akimiski Island in James Bay); and (iv) Northeast-individuals from DS (Baffin Island). Population structure differed from microsatellite studies and current management designations demonstrating the value of using SNPs for fine-scale population delineation in polar bears.
Keyphrases
  • genome wide
  • water quality
  • air pollution
  • ionic liquid
  • south africa
  • dna methylation
  • electronic health record
  • case control
  • machine learning
  • gene expression
  • big data
  • sensitive detection
  • quantum dots
  • deep learning