Dense Sampling of Taxa and Genomes Untangles the Phylogenetic Backbone of a Non-model Plant Lineage Rife with Deep Hybridization and Allopolyploidy

Abstract

Phylogenetic networks, rather than purely bifurcating trees, more accurately depict the intricate evolutionary dynamics of most lineages, especially those characterized by extensive hybridization and allopolyploidization events. However, the challenges of achieving complete taxon sampling, and limited financial resources for studying non-model plant lineages, have hindered comprehensive and robust estimation of phylogenetic backbones with guidance from networks. The bellflower tribe, Campanuleae, characterized by a reticulate evolutionary history, serves as an ideal model to investigate how to diagnose nested ancient reticulation events. Here, by integrating multiple genomic data sources and a range of phylogenetic inference methods, we produced a robust phylogenetic backbone for the tribe Campanuleae. Our investigation of reticulate evolution indicates that hybridization and allopolyploidization were instrumental in shaping the diversity of the bellflower tribe, particularly during the initial diversification of the subtribe Phytematinae. Additionally, we ascertained that conflicting topologies resulting from distinct genomic datasets and inference methodologies significantly impact downstream estimates of divergence dating, ancestral area construction, and diversification rates. This study offers a universally relevant framework for deciphering how to use network-based phylogenetic structures using various genomic sources and inference methods. [Campanulaceae, Campanuleae, Cytonuclear discordance, paralog, phylogenomics, reticulate evolution]

AdvancingPyrusphylogeny: Deep genome skimming‐based inference coupled with paralogy analysis yields a robust phylogenetic backbone and an updated infrageneric classification of the pear genus (Maleae, Rosaceae)

“…We assembled CDSs for phylogenetic inference while taking into account uneven sequencing coverage (G.N. Liu & al., 2023b; Xu & al., 2023). Specifically, we retrieved plastome‐associated reads that were generated during the extending iterations stage of Getorganelle v.1.7.6.1 (J.J. Jin & al., 2020) for each sample.…”