A comparative study of conservation and variation scores

Area of Science:

• Bioinformatics
• Computational Biology
• Structural Biology

Background:

• Conservation and variation scores are vital for analyzing multiple sequence alignments.
• These scores help identify residues critical for protein structure and function.
• The relationship between different available scores remains unclear.

Purpose of the Study:

• To compare and classify various conservation and variation scores.
• To evaluate the performance of different scores in predicting catalytic sites.

Main Methods:

• Applied 25 conservation and variation scores to alignments from the Catalytic Site Atlas (CSA).
• Calculated score distances using correlation coefficients.
• Constructed a dendrogram via average linking cluster analysis.

Main Results:

• Most scores clustered into two main groups: substitution matrix-based and frequency-based.
• Frequency-based scores generally demonstrated superior performance in predicting catalytic sites.
• The study successfully classified conservation and variation scores into two major categories.

Conclusions:

• Conservation and variation scores can be broadly categorized into two primary groups.
• Frequency-based scores, incorporating background distributions, are most effective for predicting catalytic sites.