Phylogenetic and co-ancestry analyses
Phylogenetic analyses based on GBS data were conducted with the main objective of identifying the source lineage from which the UCSD population derived. For this test we excluded once again the UCSD n.b.s. individuals from the ‘Neutral Dataset’. A maximum likelihood phylogeny and a neighbor-net were produced using the programs IQ-TREE (Nguyenet al. 2015) and SplitsTree (Bryant & Moulton 2004), respectively. For the former, the ascertainment bias correction and the generalized time-reversible (GTR) model were implemented.
We further tested the origin of the UCSD junco population by exploring pairwise levels of co-ancestry among Oregon junco forms with the program RADpainter (Malinsky et al. 2018). Starting from the ‘Full Dataset’, UCSD n.b.s. individuals were excluded. To detect subtler signs of differentiation, and because this program relies on linkage disequilibrium to achieve more refined assignment of nearest- neighbor relationships, we did not apply neutrality and linkage filters in this analysis. We did, however, increase the per-site missing data threshold to 0.75 as recommended by the authors, and applied the same MAF and HWE filters as in the PCA and STRUCTURE analyses, resulting in a matrix of 41,231 SNPs. We ran the analysis with 100K burn-in and 100K sampling iterations, and a thinning interval of 1,000 to produce a matrix of pairwise individual co-ancestry estimates, which were averaged over forms. For estimating 95% confidence intervals around pairwise co-ancestry values between UCSD and the different Oregon junco forms, we applied a non-parametric bootstrap procedure. Bootstrapping was carried out by splitting the SNP matrix in 1,000 blocks and randomly combining them for each one of the repetitions. A total of 100 analyses were run for confidence interval estimation.
Breeding latitude of UCSD juncos captured during the non-breeding season Hydrogen isotopic ratios in bird feathers correspond to those found in the localities where the feathers were grown, and since juncos molt their feathers at the end of the breeding season, isotopic ratios can be used to determine relative differences in breeding latitude (Chamberlainet al. 1996). To determine whether UCSD birds sampled in the non-breeding season (UCSD n.b.s.) were local breeders or winter visitors breeding elsewhere, we tested for correlation between deuterium isotope ratio (δD) and posterior likelihoods of genetic population assignment. We obtained isotopic data from the 28 individuals sampled in UCSD over winter, to compare to two confirmed year-round residents used as controls. A single secondary feather was collected per bird and washed using a 2:1 chloroform: methanol solution. Approximately 0.50 mg of vane material was clipped from the proximal end of each feather and the hydrogen isotope composition was measured by continuous flow isotope ratio mass spectrometry using a Finnigan TC/EA interfaced to a Finnigan DeltaPlus XL mass spectrometer (Thermo Scientific, Bremen, Germany), as described in Wunder et al. (2012). δD values are reported in per mil notation (‰) relative to V-SMOW, using internal standards (-78 ‰ and -172 ‰ respectively) calibrated to CFS-CHS-BWB (Wassenaar & Hobson 2003). Benzoic acid, (δD = −61‰) and IAEA-CH-7 (δ2H = −100‰) were also analyzed within analytical sequences with a precision of < +/-4‰. A SNP subset of the ‘Neutral Dataset’ was generated for 26 of the same 30 individuals (four failed to pass missing data filters, Table 1), and assignment likelihoods were computed with STRUCTURE under the same conditions, in this case only for K = 2. Correlation was tested by means of a linear regression between STRUCTURE scores and δD abundance.