Measuring deviation from a deeply conserved consensus in protein multiple sequence alignments

摘要

Proteins across species show variable degrees of conservation. Different patterns of conservation in the columns of an alignment indicate different evolutionary pressures on sequences. Protein conservation analysis is useful for a wide variety of applications, including disease mutation assessment, pseudogene analysis and functional residue prediction. This study describes a novel measure of column conservation in protein multiple sequence alignments ('MSA'), and the application of this measure to calculate statistical deviation from alignment consensus ('SDAC'). We have assessed SDAC for two case studies of sequences: (a) putative pseudogenes in Mycobacteria, and (b) young lineage-specific retrotransposed sequences in the human and mouse genomes. In the procedure, we rank residue positions for deep conservation, and evaluate statistically significant violations from MSA consensus.;Novel conservation measure clearly indicated a variable degree of physiochemical conservation for a given column entropy. That, in turn, enabled us to detect deviations from physiochemical consensus in a protein MSA, which are not found by entropy measures.