Disagreement among Genomic Markers Profoundly Influences Phylogenetic Inference in Squamates

Phylogenomic-scale studies of the same clades using different markers and methods often support highly confident yet different species trees, obscuring many phylogenetic relationships and hampering comparative studies. Among reptiles, the >11,000 extant species of squamates (Order Squamata: lizards, snakes, and amphisbaenians) comprise a highly diverse and well-studied clade, yet many unresolved questions about squamate origins remain, including the root of the squamate phylogeny and the relationships of snakes to other toxicoferans. To understand biological, molecular, and methodological sources of phylogenomic heterogeneity in squamates, we analyzed four genome-scale marker datasets, including thousands of protein-coding genes, anchored hybrid enrichment loci, and ultra-conserved elements. We applied standardized alignment, filtering, and tree-building methods across marker sets, analyzed patterns in substitutional saturation and codon positions, and measured gene tree-species tree discordance across all data partitions. We found that many contentious relationships in squamate phylogenetics are driven by conflicts stemming from input data quality, incorrect model fit, and sampling bias, and that biological drivers of heterogeneity include both incomplete lineage sorting and introgression. We account for these sources of heterogeneity and strengthen resolution of the squamate phylogeny. Using a simulation approach, we find that current ultra-conserved element datasets for squamates deviate the most from phylogenetic expectations among marker types. Our work demonstrates that identifying sources of phylogenomic heterogeneity while accounting for input data limitations can resolve phylogenetic conflicts.