BigFoot: Bayesian alignment and phylogenetic footprinting with MCMC

Area of Science:

• Computational Biology
• Bioinformatics
• Genomics

Background:

• Previous methods combined statistical alignment and phylogenetic footprinting for functional element detection.
• These methods analyzed up to four sequences using dynamic programming and hidden Markov models.
• A probability-weighted distribution of alignments enhances accuracy and reduces sensitivity to alignment errors.

Purpose of the Study:

• To develop a novel phylogenetic footprinting approach for analyzing larger numbers of sequences.
• To implement this method in a software package named BigFoot.
• To improve the accuracy and robustness of functional element prediction.

Main Methods:

• Developed a Markov chain Monte Carlo (MCMC) approach.
• Sampled both sequence alignments and locations of slowly evolving regions.
• Extended the existing StatAlign software package.

Main Results:

• The BigFoot software package enables phylogenetic footprinting on numerous sequences.
• Tested on Drosophila even-skipped and vertebrate alpha-globin gene regions.
• Demonstrated improved accuracy in functional predictions with increased sequence data.
• Showcased BigFoot's superior performance compared to existing alignment-based techniques.

Conclusions:

• BigFoot extends alignment and phylogenetic footprinting for large-scale sequence analysis using MCMC.
• The approach is robust to alignment errors and uncertainty.
• Applicable to diverse biological datasets with publicly available source code.