Phylogenomic subsampling and upsampling for efficient evolutionary analyses of big data

Area of Science:

• Computational Biology
• Evolutionary Biology
• Bioinformatics

Background:

• Long phylogenetic analyses face computational challenges like high runtime and memory usage.
• Current methods often require high-performance computing for large phylogenomic datasets.

Purpose of the Study:

• To introduce and evaluate a scalable framework, phylogenomic subsampling and upsampling (PSU), for efficient evolutionary analysis.
• To demonstrate how PSU can approximate full-data analyses with reduced computational cost.

Main Methods:

• PSU involves analyzing small subsamples of sites from large alignments, extended by upsampling, and aggregating results.
• This approach separates computational burden from inferential power by reducing distinct site patterns while retaining evolutionary information.

Main Results:

• PSU accurately estimates key phylogenetic metrics including bootstrap support, model selection, and branch lengths.
• The framework significantly reduces runtime and memory requirements, often by orders of magnitude.
• PSU facilitates the detection of concordant and conflicting phylogenetic signals.

Conclusions:

• PSU is a general strategy for scalable phylogenomic inference across various statistical methods.
• It democratizes access to computationally intensive evolutionary methods by enabling analyses on standard hardware.
• PSU reduces environmental and infrastructural costs associated with big-data phylogenomics.