Artifactual orthologs and the need for diligent data exploration in complex phylogenomic datasets: A museomic case study from the Andean flora

The Andes mountains of western South America are a globally important biodiversity hotspot, yet there is a paucity of resolved phylogenies for plant clades from this region. Filling an important gap to our understanding of the World’s richest flora, we present the first phylogeny of Freziera (Pentaphylacaceae), an Andean-centered, cloud forest radiation. Our dataset was obtained via hybrid-enriched target sequence capture of Angiosperms353 universal loci for 50 of the ca. 75 spp., obtained almost entirely from herbarium specimens. We identify high phylogenomic complexity in Freziera, including a significant proportion of paralogous loci and a high degree of gene tree discordance. Via gene tree filtering, by-eye observation of gene trees, and detailed examination of warnings from recently improved assembly pipelines, we identified that cryptic paralogs (i.e., the presence of only one copy of a multi-copy gene due to assembly errors) were a major source of gene tree heterogeneity that had..., , , # Strong phylogenetic signal despite high phylogenomic complexity in an Andean plant radiation (Freziera, Pentaphylacaceae)

## Description of the data and file structure

Data files are organized into three main directories: alignments, gene trees and species trees. Below is a summary of the file structure in which each folder is noted by a bullet. Names of folders are bolded; descriptions of the folders are provided after the folder name.

* **alignments**: .nexus files, DNA sequence data
* **Hybpiper_v1.3.1**: alignments of supercontigs assembled by HybPiper v1.3.1 for loci without paralog warnings (318 loci). Sequences identified by TreeShrink as producing abnormally long branch lengths have been removed from alignments (\"treeshrunk\").
* **paralogs**: alignments of supercontigs assembled by HybPiper v1.3.1 for loci with paralog warnings (31 loci).
* **Hybpiper_v2.0.1**: alignments of supercontigs assembled by HybPiper v2.0.1 for loci without long paralog war...