Summary Tests of Introgression Are Highly Sensitive to Rate Variation Across Lineages.

Area of Science:

• Evolutionary biology
• Genomics
• Phylogenetics

Background:

• Hybridization and introgression are increasingly recognized as significant evolutionary forces.
• Current methods for detecting hybridization often rely on summary statistics from small taxon subsets.
• These methods frequently assume constant evolutionary rates, a condition often unmet in real data.

Purpose of the Study:

• To quantify the impact of evolutionary rate variation on widely used hybridization detection tests.
• To evaluate the robustness of the D test (ABBA-BABA), D3 test, and HyDe under rate variation.
• To assess how lineage-specific and gene-specific rate variations affect the accuracy of these tests.

Main Methods:

• Simulated species networks using a birth-death-hybridization model to represent realistic phylogenetic scenarios.
• Quantified the effects of rate variation across species lineages, genes, and gene-tree edges.
• Assessed type-1 error rates (false discovery of reticulation) and power to detect hybridization.

Main Results:

• Rate variation across species lineages markedly increased false discovery rates for all tested methods.
• The D3 test exhibited high sensitivity to rate variation, showing ~80% type-1 error, suggesting it detects rate departures more than reticulation.
• Increased hybridization events reduced the power of all tests, indicating potential for obscuring signals in complex scenarios.

Conclusions:

• Widely used summary statistics tests for hybridization are sensitive to evolutionary rate variation, leading to inflated false positive rates.
• The D3 test is particularly susceptible to rate variation, potentially misinterpreting clock departures as hybridization.
• Future studies should consider methods robust to rate variation or explicitly account for it to accurately infer hybridization events.