Sequence Capture Versus Restriction Site Associated DNA Sequencing for Phylogeography
Michael G. Harvey, Brian Tilston Smith, Travis C. Glenn, Brant C. Faircloth, Robb T. Brumfield
(Submitted on 22 Dec 2013)
Genomic datasets generated with massively parallel sequencing methods have the potential to propel systematics in new and exciting directions, but selecting appropriate markers and methods is not straightforward. We applied two approaches with particular promise for systematics, restriction site associated DNA sequencing (RAD-Seq) and sequence capture (Seq-cap) of ultraconserved elements (UCEs), to the same set of samples from a non-model, Neotropical bird. We found that both RAD-Seq and Seq-cap produced genomic datasets containing thousands of loci and SNPs and that the inferred population assignments and species trees were concordant between datasets. However, model-based estimates of demographic parameters differed between datasets, particularly when we estimated the parameters using a method based on allele frequency spectra. The differences we observed may result from differences in assembly, alignment, and filtering of sequence data between methods, and our findings suggest that caution is warranted when using allele frequencies to estimate parameters from low-coverage sequencing data. We further explored the differences between methods using simulated Seq-cap- and RAD-Seq-like datasets. Analyses of simulated data suggest that increasing the number of loci from 500 to 5000 increased phylogenetic concordance factors and the accuracy and precision of demographic parameter estimates, but increasing the number of loci past 5000 resulted in minimal gains. Increasing locus length from 64 bp to 500 bp improved phylogenetic concordance factors and minimal gains were observed with loci longer than 500 bp, but locus length did not influence the accuracy and precision of demographic parameter estimates. We discuss our results relative to the diversity of data collection methods available, and we provide advice for harnessing next-generation sequencing for systematics research.