mCSM_NA

mCSM_NA predicts the effects of missense mutations on protein-nucleic acid binding affinities using graph-based structural signatures to quantify mutation-induced changes in binding.

Key Features:

• Graph-Based Signature Enhancement: Incorporates graph-based signatures that integrate pharmacophore modeling and nucleic acid properties to represent the local structural environment around mutations.
• Reverse Mutation Consideration: Performs reverse mutation analysis to evaluate mutational effects in both directions.
• Refined Data Set Utilization: Trains and evaluates models using an updated release of the ProNIT database.
• Performance Metrics: Reports correlation coefficients up to 0.70 in cross-validation and 0.68 in blind tests.

Scientific Applications:

• Mechanistic Interpretation: Elucidates molecular mechanisms by quantifying how mutations alter nucleic acid binding and linking those changes to phenotypic variation.
• Protein–Nucleic Acid Interaction Analysis: Predicts impacts of mutations on protein–DNA and protein–RNA interactions relevant to gene regulation.
• Genomics and Personalized Medicine: Supports genomics, molecular biology, and personalized medicine studies by prioritizing mutations that affect nucleic acid binding.

Methodology:

Uses graph-based signatures integrated with pharmacophore modeling and nucleic acid properties, includes reverse mutation analysis, and is trained/validated on an updated ProNIT dataset with cross-validation and blind-test evaluations reporting correlations up to 0.70 and 0.68.