mmCSM-NA

mmCSM-NA predicts the effects of single and multiple mutations on protein-nucleic acid binding affinity to quantify mutation-driven changes in binding energetics.

Key Features:

• Extensive Dataset: Uses a curated dataset of 856 single-point mutations and 141 multiple-point mutations across 155 experimentally solved protein-nucleic acid complexes.
• Graph-Based Signatures: Employs an optimized version of graph-based signatures to quantitatively predict the impact of mutations on binding affinity.
• Performance Metrics: Achieves Pearson's correlation up to 0.67 (RMSE 1.06 kcal/mol) for single-point mutations under cross-validation and up to 0.65 (RMSE 1.12 kcal/mol) for multiple-point mutations on independent non-redundant datasets.
• Scalability: Reported as the first scalable method capable of predicting effects of multiple-point mutations on nucleic acid binding affinities and shown to outperform similar tools.

Scientific Applications:

• Investigation of genetic variation effects: Predicts how mutations alter protein-nucleic acid binding affinity to study impacts on protein function.
• Identification of therapeutic targets: Identifies critical mutations affecting binding affinities that can inform potential therapeutic target selection.
• Support for drug design: Provides quantitative predictions of mutation-driven binding changes to inform drug design efforts involving protein-nucleic acid interactions.

Methodology:

Integration of the curated experimental dataset with an optimized graph-based signature computational approach to predict the effects of single and multiple mutations on binding affinity.