Synergizing Bayesian and Heuristic Approaches: D-BPP Uncovers Ghost Introgression in Panthera and Thuja

Area of Science:

• Evolutionary Biology
• Genomics
• Phylogenetics

Background:

• Hybridization with extinct or unsampled lineages significantly impacts species evolution.
• Detecting such "ghost" introgression is methodologically difficult for current approaches.

Purpose of the Study:

• To introduce D-BPP, a novel framework for inferring phylogenetic networks by integrating D-statistics and Bayesian phylogenomic inference.
• To efficiently and accurately detect introgression events, including those involving ghost lineages.

Main Methods:

• The D-BPP framework first uses D-statistics (ABBA-BABA test) to identify candidate introgression events.
• Bayesian phylogenomic inference (BPP) is then employed to validate candidates and build phylogenetic networks.
• The method handles ambiguous species trees by comparing introgression models within a Bayesian framework.

Main Results:

• Simulation analyses demonstrate high power for D-BPP, with reliable detection of introgression and accurate discrimination of scenarios.
• The framework successfully identifies ghost introgression, often missed by existing methods.
• Application to big cat (Panthera) and conifer (Thuja) genomic data revealed previously undetected ghost introgression events.

Conclusions:

• D-BPP offers a practical and powerful approach for reconstructing complex, reticulate evolutionary histories.
• The framework combines computational efficiency with robust statistical rigor.
• Ghost introgression plays a pervasive role across diverse taxa, as evidenced by findings in Panthera and Thuja.