GHIST 2024: The First Genomic History Inference Strategies Tournament.

Area of Science:

• Population genetics
• Evolutionary biology
• Computational genomics

Background:

• Evaluating population genetic inference methods is complex due to evolutionary history intricacies.
• Model misspecification and self-assessment biases challenge method evaluation.
• Standardized benchmarks are needed for reliable evolutionary history inference.

Purpose of the Study:

• To establish a community-driven competition for evaluating population genomic inference methods.
• To assess the performance of diverse inference strategies across various demographic scenarios.
• To identify strengths and weaknesses of current methods for inferring evolutionary history.

Main Methods:

• The Genomic History Inference Strategies Tournament (GHIST) competition was conducted from July to November 2024.
• Four demographic history challenges were presented: bottleneck, split with isolation, secondary contact, and archaic admixture.
• Participants submitted numerical parameter estimates from error-free VCF files, scored by relative root-mean-squared error.

Main Results:

• Approximately 60 participants competed using diverse computational approaches.
• Methods based on site frequency spectra demonstrated current dominance.
• Flexible model-building approaches showed advantages for complex demographic histories.

Conclusions:

• GHIST provides standardized benchmarks for population genomic inference.
• The competition highlighted the need for improved methods for complex evolutionary scenarios.
• The ongoing GHIST iteration aims to expand challenge diversity for more robust evaluations.