Viral Simulations Expose a Hidden Bias in Evolutionary Reconstructions
A new study reveals a critical flaw in phylodynamic methods used to track viral evolution within a host. Researchers developed an agent-based simulation tool, virolution, to model HIV-1 evolution across anatomical compartments under both neutral and selective pressures. They discovered that when purifying selection is present—a common real-world scenario—standard Bayesian models (like those in BEAST2) significantly overestimate migration rates between compartments. This bias arises because these widely used methods assume neutral evolution, an assumption that fails when parts of the genome are under selective pressure, leading to potentially inaccurate reconstructions of viral population dynamics and spread.
Why it might matter to you: For professionals in evolutionary biology and phylogenetics, this work underscores a fundamental methodological pitfall. If your research involves inferring migration, gene flow, or population history from genetic sequences—especially in pathogens—this finding mandates a re-evaluation of model assumptions. It highlights the necessity of incorporating selection into phylogeographic frameworks to avoid biased estimates that could misinform our understanding of evolutionary processes like adaptation and spread within hosts.
Source →Stay curious. Stay informed — with Science Briefing.
Always double check the original article for accuracy.