Genotyping-by-sequencing
Genomic DNA was extracted from blood and tissue samples using a Qiagen DNeasy kit (QiagenTM, Valencia, CA). We used genotyping-by-sequencing (GBS; Elshire et al. 2011) to obtain individual genotypes from a total of 174 juncos. Of these, 46 individuals were sampled at UCSD, 18 of them corresponding to confirmed resident breeders (referred to as ‘UCSD breeders’ hereafter), and 28 were sampled during the non-breeding season (referred to as ‘UCSD n.b.s.’ hereafter). In addition, we sequenced 128 Oregon juncos belonging to the following taxa (with sample sizes in parentheses): townsendi (16),pontilis (13), pinosus (14), thurberi (41; 23 of them from Laguna Mountains, and 18 from Tahoe), montanus (16),shufeldti (12) and oreganus (16) (Fig. 1, Table 1, Table S1 from Supporting Information). GBS libraries were prepared and sequenced at Cornell University’s Institute for Genomic Diversity, using restriction enzyme PstI for digestion. Sequencing of the 174 individually-barcoded libraries was carried out in five different lanes (along with another 301 junco samples intended for other analyses) of an Illumina HiSeq 2000, resulting in an average of 243.2 million good barcoded single-end reads 100 bp in length per lane.