DAPCy: a Python package for the discriminant analysis of principal components method for population genetic analyses.

Area of Science:

• Population genetics
• Bioinformatics
• Computational biology

Background:

• Discriminant Analysis of Principal Components (DAPC) is crucial for population genetics.
• Existing R implementations of DAPC struggle with large genomic datasets.

Purpose of the Study:

• Introduce DAPCy, a Python package for enhanced DAPC analysis.
• Improve scalability and efficiency for large-scale genomic data.

Main Methods:

• Utilize scikit-learn library in Python.
• Employ compressed sparse matrices and truncated SVD for dimensionality reduction.
• Implement training-test cross-validation for model evaluation.

Main Results:

• DAPCy processes large genomic datasets (thousands of samples/features) faster and with less memory.
• Demonstrated computational advantages using Plasmodium falciparum and 1000 Genomes Project datasets.
• DAPCy includes de novo clustering and comprehensive visualization tools.

Conclusions:

• DAPCy provides a scalable and efficient solution for DAPC analysis.
• Facilitates genetic structure assessment in large population genomics studies.
• Offers advanced features for clustering, visualization, and reporting.