Population genetic inference software for the genomic age

MIGRATE is a widely used open-source tool for estimating population size, migration rates, and divergence times from genetic data. This project will modernize MIGRATE and build capacity to analyze whole-genome datasets and incorporate an AI-powered, Retrieval-Augmented Generation (RAG) assistant to help users with software setup, analysis, and interpretation. By combining advanced genomics with AI-enabled scientific computing, the project will broaden access to population-genetic inference and accelerate research across biotechnology, agriculture and the life sciences. Open-source distribution, cross-platform executables, tutorials, workshops, and community support will foster workforce development, open science, and reusable cyberinfrastructure aligned with national priorities in AI and Biotechnology. The technical effort will rewrite MIGRATE in a modular C/C++ architecture. The new design separates data I/O, model specification, and inference engines. It enables seamless integration with high‑performance computing clusters via MPI. New parsers will accept standard genomic formats such as VCF and reference‑sequence files. These parsers will allow analysis of thousands of loci across hundreds of individuals. They will avoid the memory bottlenecks of the legacy code. An API will be provided for Python and R. The API will facilitate interoperability with existing bioinformatics pipelines and statistical environments. A locally hosted AI‑based RAG system will be grounded in the updated documentation. It will offer context‑aware suggestions, error diagnosis, and best‑practice recommendations. These features will lower the learning curve for new users. The re-engineered software will retain MIGRATE’s finite‑mutation‑model capabilities. It will offer a robust alternative to site‑frequency‑spectrum methods that rely on restrictive infinite‑sites assumptions. Performance gains will be shown through benchmarking on large high-performance computing clusters. All source code, binaries, manuals, tutorials, and workshop materials will be released publicly under the MIT license. This ensures long‑term sustainability beyond the grant period. By delivering an extensible, user‑friendly, and AI‑enhanced platform, the project addresses NSF priorities in artificial intelligence, advanced computational methods, and biotechnology. It will serve as a foundational resource for the next generation of population‑genomic research. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2549631 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Peter Beerli | Institution: Florida State University, TALLAHASSEE, FL | Award Amount: $499,882 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2549631 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2549631.html