A single tool now infers which allele is ancestral from any genealogy.

When scientists look at DNA differences across individuals, they want to figure out which letter is original, called ancestral, and which is a newer mutation, called derived. Knowing this is key for studying evolution, natural selection, and population history. However, existing tools cannot handle all data types, especially when sites have multiple mutations or lack related species for comparison, known as outgroups. To solve this, researchers built Ancestree, a software engine that calculates probabilities for all four DNA letters at every position. It works by tracing how DNA sequences relate through family trees. It can use a single fixed tree, take changing local trees from an ancestral recombination graph, or estimate those trees directly from raw genetic data using a hidden Markov model. In tests on simulated data, the tree-based methods were more accurate and scaled better. Having outgroups still proved vital, especially for detecting mutations shared by the entire study group.

Why it matters

Population geneticists studying selection, adaptation, and demography rely on knowing which DNA variant came first. Until now, they had to use different tools for different types of data—plain variants, family trees, with or without outgroups—and none gave full uncertainty. Ancestree unifies all these inputs in one framework, returning probabilities for each possible ancestor, which improves accuracy and enables more robust downstream analyses.

Questions answered