I-MUTANT Suite

I-MUTANT Suite predicts the effects of single point mutations on protein stability using support vector machine (SVM) algorithms, providing both sign predictions (increase/decrease) and quantitative ΔΔG estimates from protein sequence or structure.

Key Features:

• Core component: I-Mutant2.0 implements support vector machine (SVM) models to predict stability changes upon single point mutations.
• Dual input capability: Predictions can be generated from either protein sequence or protein structure input.
• Predictive modes: Operates as a classifier to predict stability sign (increase/decrease) and as a regression estimator to predict ΔΔG values.
• Training and validation: Models were trained and validated using ProTherm, the thermodynamic experimental database of protein stability changes due to mutations.
• Accuracy metrics: Classifier accuracy is 80% with structural information and 77% with sequence information; ΔΔG prediction correlation coefficients are 0.71 (structure, standard error 1.30 kcal/mol) and 0.62 (sequence, standard error 1.45 kcal/mol).

Scientific Applications:

• Protein engineering and design: Provides stability impact data for rational design of proteins with altered stability profiles.
• Therapeutic and industrial protein development: Supports selection and prioritization of single point mutations for proteins intended for therapeutic or industrial applications.

Methodology:

Predictions are produced by SVM models (I‑Mutant2.0) trained and validated on ProTherm, accepting sequence or structure input and outputting classification of stability sign and regression estimates of ΔΔG (kcal/mol).