QMaker

QMaker estimates empirical amino acid substitution models by maximum likelihood, producing general time-reversible Q matrices from multiple sequence alignments for phylogenetic analyses.

Key Features:

• Efficient maximum likelihood estimation: Estimates general time-reversible (GTR) Q matrices from large collections of protein multiple sequence alignments using maximum likelihood.
• Efficient ML tree search algorithm: Implements an efficient maximum-likelihood tree search to accelerate Q-matrix estimation.
• Model heterogeneity handling: Performs model selection across different alignments to accommodate heterogeneity among datasets.
• Rate mixture models: Incorporates rate mixture models across sites to model among-site rate variation in protein sequences.
• Parallel processing: Supports parallel computation across multiple CPU cores to increase computational throughput.

Scientific Applications:

• Pfam-derived empirical models: Derives new empirical general amino acid substitution models from the Pfam database.
• Clade-specific model development: Facilitates generation of clade-specific amino acid models for mammals, birds, insects, yeasts, and plants.
• Model fit and topology effects: Improves model–data fit and can influence inferred phylogenetic tree topologies.

Methodology:

Maximum-likelihood estimation of general time-reversible Q matrices from multiple sequence alignments, efficient ML tree search algorithms, model selection for alignment heterogeneity, and inclusion of rate mixture models across sites, with computations parallelizable across CPU cores.