A Better Newspaper

## Overview sstar2 is a Python software package for detecting introgressed genomic fragments from unsampled or extinct source populations using the S* statistic, updated with machine learning-based threshold estimation (bioRxiv:2026.05.31.729079). It was presented in a June 2026 preprint as an update to the original sstar implementation. ## Technical Advance The original sstar implementation relied on generalized additive models (GAMs) to smooth quantile-specific values precomputed from fixed count bins, requiring simulations with fixed numbers of segregating sites (bioRxiv:2026.05.31.729079). sstar2 replaces this procedure with quantile regression to directly estimate S* thresholds at specified null quantiles from simulated genomic windows — a more flexible and computationally tractable approach. ## Scientific Context The S* statistic is widely used for detecting Neanderthal, Denisovan, and other archaic hominin introgression in modern human genomes. As ancient DNA sequencing and population genomics scale up, robust computational tools for introgression detection become increasingly important. ## Relevance Beyond Academic Research While primarily a bioinformatics tool, sstar2 has relevance to: - **Forensic genomics**: Population stratification analysis - **Biotech IP**: Genomic analysis pipelines used in ancestry testing or population health could incorporate or cite such tools - **Regulatory science**: As ancient DNA analysis enters clinical and forensic contexts, validated open-source tools become reference standards ## Outlook sstar2 is expected to become a standard tool in population genomics pipelines analyzing archaic introgression. Its machine learning backbone positions it within the broader trend of ML integration into genomic analysis workflows.