ABSTRACT
Non-invasive genomic research on free-ranging mammals typically relies on the use of fecal DNA. This requires the isolation and enrichment of endogenous DNA, given its small proportion compared to bacterial DNA. Current approaches for acquiring large-scale genomic data from feces rely on bait-and-capture techniques. While this technique has greatly improved our understanding of mammalian population genomics, it is limited by biases inherent to the capture process, including allele dropout, low mapping rates, PCR duplication artifacts, and structural biases. We report here a new method for generating whole mammalian genomes from feces using fluorescence-activated cell sorting (FACS). Instead of enriching endogenous DNA from extracted fecal DNA, we isolated mammalian cells directly from feces. We then built fragment libraries with low input material from commercially available kits, which we sequenced at high and low coverage. We validated this method on feces collected from primates stored in RNAlater for up to three years. We sequenced one fecal genome at high coverage (12X) and 15 additional fecal genomes at low coverage (0.1X - 4X). For comparative purposes, we also sequenced DNA from nine blood or tissue samples opportunistically collected from capuchin monkeys that died of natural causes or were treated in a local rehabilitation center. Across all fecal samples, we achieved median mapping and duplication rates of 82% and 6%, respectively. Our high-depth fecal genome did not differ in the distribution of coverage, heterozygosity, or GC content from those derived from blood or tissue origin. As a practical application of our new approach with low coverage fecal genomes, we were able to resolve the population genetic structure of capuchin monkeys from four sites in Costa Rica.