DNA extracts were sent to the Institute of Genomic Diversity at Cornell University for genotyping-by-sequencing. High quality DNA extracts (>20 ng/μL) were digested with the restriction enzyme PstI and a unique barcode/adapter combination was ligated to each sample prior to pooling and amplification by PCR. Fragments were pooled with amplicons from other bird species to create three sequencing libraries, with 95 barcoded individuals per library, and sequenced on multiple lanes of a Hi-Seq 2000 Illumina sequencer (Illumina, San Digeo, CA). Each library resulted in ~200 Gbp of raw data. The 100 base pair reads were filtered using the Universal Network Enabled Analysis Kit (UNEAK) pipeline implemented in TASSEL 3.0. Adapter dimers and sequences with an ambiguous base were removed and all reads were trimmed to 64 bp. Identical reads were merged into “tags” within each barcoded individual. Using a pairwise alignment, tags with 1 bp mismatch were retained as candidate SNPs. Candidate SNP loci that were present in fewer than 75% of individuals were removed, and then individuals possessing fewer than 50% of candidate SNPs were removed. To construct genotypes based on sequencing coverage, a maximum-likelihood method with an updated version of the script from White et al. (2013) was implemented; this method retained SNPs with an AIC ≥ 4 units lower than the next best reconstructed genotype. To remove paralogs (duplicated regions within the genome), SNPs with observed heterozygosities ≥ 0.75 were removed. Finally, SNPs present in fewer than 25% of individuals and individuals with fewer than 70% of SNPs were excluded. Following filtering, we retained a dataset of 5633 SNPs from 72 birds, including 37 Bicknell’s Thrushes from locations spanning the species’ entire breeding range, 23 Gray-cheeked Thrushes representing both subspecies and including individuals from Siberia, Alaska, and eastern Canada, a probable late migrating Gray-cheeked Thrush, and a small sample of eastern and western Veeries (n = 11).

The first column lists the sample identification number; all samples are museum accession numbers; date of collection, location of collection, and species type are listed in FitzGerald et al. (2019) The Auk. Each individual has two rows. The first row is the SNP name. A '0' indicates an uncalled allele.