FastCodeML

FastCodeML accelerates likelihood-based detection of positive selection in codon-based phylogenetic analyses to enable large-scale studies of gene and genome evolution.

Key Features:

• Computational Efficiency: Implements optimization techniques that improve likelihood-estimation efficiency relative to CodeML from the PAML package and SlimCodeML for codon models.
• Scalability: Handles large numbers of codons and species to support extensive phylogenetic datasets.
• Performance Enhancements: Demonstrates average sequential speedups up to 5.8× over CodeML, multicore speedups up to 36.9× on 12 CPU cores (single-node, shared memory), and distributed speedups up to 170.9× across eight nodes (96 CPU cores) versus CodeML.
• Applicability: Illustrated using the branch-site model of codon evolution and applicable to other likelihood-based phylogeny software.
• Likelihood Estimation Optimization: Optimizes estimation of likelihood functions for codon-based phylogenetic analyses to reduce computational overhead.

Scientific Applications:

• Detection of Positive Selection: Enables likelihood-based tests to identify positive selection across phylogenetic trees using codon models.
• Gene and Genome Evolution: Supports studies of gene function evolution, adaptive changes, and genomic diversity across species.
• Large-scale Evolutionary Analyses: Facilitates analysis of large phylogenomic datasets to address complex evolutionary questions.

Methodology:

Applies a series of computational optimization techniques tailored to codon-model phylogenetic likelihood estimation, illustrated on the branch-site model.