DNA Sequencing
We obtained mitochondrial genome sequences from 83 mountain treeshrew individuals and 10 sequences from three congeners including the large treeshrew (n = 6), the small treeshrew (n = 2) and the ruddy treeshrew (n = 2). Mitogenomes were sequenced to an average depth of 50×. Sequences were submitted to GenBank (Accession numbers pending).
We sequenced UCEs from 80 mountain treeshrews (NCBI SRA number pending). Each UCE-enriched library was sequenced with a mean of 2.3 million reads (914,104–7,011,836), yielding a mean of 3,344 UCE loci (2,137–3,489) per sample. The total number of UCE alignments that we used to generate the pseudo-reference was 4,106, and the mean length was 495 base pairs (149–2167 bps). After aligning reads to the pseudo-reference and quality filtering, there were 7,861 SNPs including multiple SNPs per locus. After removing loci with more than 10% missing data across individuals, 3,168 SNPs remained. The unlinked SNP dataset included 1,794 independent SNPs. Removing loci with missing data left 684 unlinked loci. In the phased pseudo-haplotype sequence alignment dataset, after removing loci with less than one (114 loci) and more than 10 PIS (16 loci), 1,664 UCE alignments remained (Figure S1).