Taxonomy of Pallas's Reed Bunting Emberiza pallasi: an integrative approach

These are sequence homologs of intronic loci extracted from fasta sequences compiled for each individual sampleby the BirdScanner 2 pipeline (Ericson et al., 2020) (https://github.com/ Naturhistoriska/birdscanner). As references we used 7,121 alignments of intronic loci of four passerines species, Acanthisitta chloris, Corvus brachyrhynchos, Geospiza fortis and Manacus vitellinus, from the data generated by Jarvis et al. (2014) and previously used by, e.g., Jiang et al. (2024). For each sample, the parsed-out sequences were placed in these fasta files compiling the orthologs of respective locus. These fasta files were then processed with ATPW – Align-and-Trees-Parallel-Workflow (Nylander, 2022, https://github.com/nylander/Align-and-trees-parallel-workflow), which first aligned all sets of orthologous genes with Mafft v.7.526 (Katoh et al., 2002; Katoh & Standley, 2013). The alignments were then filtered with BMGE v.1.12 (Criscuolo & Gribaldo, 2010) which selects regions in a multiple sequence alignment that are suited for phylogenetic inference. Phylogenetic trees for each filtered locus were inferred with ParGenes (Morel et al., 2018) and RAxML-NG, (Kozlov et al., 2019), filtered with TreeShrink (Uyen & Mirarab, 2018) and realigned with Mafft. Substitution model selection was performed using ModelTest-NG (Darriba et al., 2020, Flouri et al., 2014) and phylogenetic trees for each locus were inferred for these filtered and realigned alignments with RAxML-NG and ParGenes, subjected to 5000 bootstrap iterations. Finally, a species tree rooted by E. schoeniclus was estimated from the individual trees using ASTER (Zhang & Mirarab, 2022).