Optimizing Selection of Brown Bear Hair for Noninvasive Genetic Analysis

Abstract

Many wildlife studies use noninvasive survey methods such as barbed wire to obtain hair samples for DNA analysis. If laboratory costs preclude processing all samples, it may be important to know a priori which samples are most likely to yield useful DNA. It may also be helpful to know whether apparently poor-quality samples will yield useable DNA, and so be worth the time and cost of processing. To help resolve this issue, we initiated a field classification system for hair from brown bears (Ursus arctos) obtained using unbaited wires deployed across 6 streams in western Alaska, USA, where bears travel and feed on adult Pacific salmon (Oncorhynchus nerka). Averaged over 4 years (2014–2017), amplification success ranged from 89% for samples categorized in the field as being of highest quality, based on volume of hairs, to 37% for single or double strands of hair. Moreover, DNA analysis of poor-quality samples markedly increased the numbers of individual bears detected, and those detected multiple times—important data for genetic capture–mark–recapture abundance estimates. Thus, field classification can help cut costs and streamline laboratory analyses when hair samples are abundant. Yet, even apparently poor-quality samples are likely worth processing if they are the only ones representing a site, date, or other sampling stratum, hair capture devices are checked infrequently, or laboratory costs are not limiting. © 2020 The Wildlife Society.