pygenstrat: a Python package for EIGENSTRAT data processing

Area of Science:

• Population Genetics
• Bioinformatics
• Computational Biology

Background:

• Ancient DNA studies utilize the EIGENSTRAT genotype format for population genetic analyses.
• Limited software exists for processing EIGENSTRAT data, posing a challenge for researchers.

Purpose of the Study:

• Introduce pygenstrat, a Python package designed for comprehensive EIGENSTRAT data processing.
• Provide efficient tools for filtering, subsetting, and converting ancient DNA datasets.

Main Methods:

• Developed pygenstrat as a command-line interface with memory-efficient, chunked processing algorithms.
• Implemented functionalities for updating files, subsetting datasets, filtering by MAF and missingness, pseudo-haploidisation, and allele polarization.
• Enabled conversion between EIGENSTRAT and ANCESTRYMAP formats.

Main Results:

• Benchmarking demonstrated 2×-15× speedups and 90%-95% memory reduction compared to existing tools like convertf.
• pygenstrat produces equivalent outputs to standard operations, ensuring data integrity.
• The package facilitates reproducible processing and reduces turnaround time in ancient DNA workflows.

Conclusions:

• pygenstrat offers a powerful, memory-efficient solution for processing large ancient DNA datasets in EIGENSTRAT format.
• Its modular design allows for integration into custom pipelines and future extensions.
• The open-source availability promotes accessibility and collaborative development in the field.