Detecting groups of coevolving positions in a molecule: a clustering approach

Area of Science:

• Evolutionary Biology
• Biochemistry
• Bioinformatics

Background:

• Co-substitution patterns in RNA are understood, but detecting coevolving protein sites remains difficult.
• Weak signals in protein coevolution arise from diverse amino acid biochemical properties and complex compensatory mutation mechanisms.

Purpose of the Study:

• To develop a novel computational method for accurately detecting coevolving positions in protein sequences.
• To address the challenges of biochemical diversity and multi-site compensation in protein coevolution analysis.

Main Methods:

• Phylogenetic substitution mapping incorporating weighted amino acid changes based on biochemical properties (volume, polarity, charge).
• Application of a clustering approach to identify groups of coevolving sites of varying sizes.
• Differentiation between biochemical compensation and other coevolutionary processes.

Main Results:

• Successfully detected groups of protein sites with statistically significant correlated substitutions, deviating from independence.
• Identified coevolving site groups ranging in size from pairs to approximately ten sites.
• Demonstrated the method's applicability on diverse datasets including bacterial ribosomal RNA and vertebrate myoglobin.

Conclusions:

• The developed method effectively identifies coevolving protein sites, providing a valuable tool for evolutionary and structural biology.
• The findings highlight the structural and functional significance of identified coevolving site groups.
• The study discusses various evolutionary processes that may lead to correlated substitution patterns in proteins.