I-Mutant2.0 is a support vector machine (SVM)-based tool for predicting protein stability changes upon single point mutations. It can predict stability changes using either the protein structure or sequence. The method was trained and tested on a dataset derived from ProTherm, a comprehensive database of thermodynamic experimental data. I-Mutant2.0 can act as a classifier to predict the sign of the stability change or as a regression estimator to predict the ΔΔG values. When used as a classifier, it correctly predicts 80% or 77% of the dataset depending on whether structural or sequence information is used. When predicting ΔΔG values, the correlation between predicted and experimental values is 0.71 (standard error 1.30 kcal/mol) for structural data and 0.62 (standard error 1.45 kcal/mol) for sequence data. The web interface allows users to select a predictive mode based on the availability of the protein structure or sequence. I-Mutant2.0 is based on SVMs and uses a radial basis function (RBF) kernel. The input vector includes information about temperature, pH, residue types, and the residue environment. The RSA value can be calculated using DSSP when the structure is available. The RI value is computed based on the SVM output. I-Mutant2.0 was trained on 2087 single mutations in 65 proteins, with 58 known at atomic resolution. The accuracy of structure-based predictions is 0.80, while sequence-based predictions achieve 0.77. The correlation between predicted and experimental ΔΔG values is 0.71 for structure-based predictions and 0.62 for sequence-based predictions. The tool allows users to predict the stability change based on the protein structure or sequence, and provides outputs including the predicted ΔΔG value, sign, temperature, pH, and reliability index (RI) or RSA. The results show that I-Mutant2.0 is a valuable tool for protein design, even when the protein structure is not known at atomic resolution. The tool is available at http://gpcr.biocomp.unibo.it/cgi/predictors/I-Mutant2.0/I-Mutant2.0.cgi.I-Mutant2.0 is a support vector machine (SVM)-based tool for predicting protein stability changes upon single point mutations. It can predict stability changes using either the protein structure or sequence. The method was trained and tested on a dataset derived from ProTherm, a comprehensive database of thermodynamic experimental data. I-Mutant2.0 can act as a classifier to predict the sign of the stability change or as a regression estimator to predict the ΔΔG values. When used as a classifier, it correctly predicts 80% or 77% of the dataset depending on whether structural or sequence information is used. When predicting ΔΔG values, the correlation between predicted and experimental values is 0.71 (standard error 1.30 kcal/mol) for structural data and 0.62 (standard error 1.45 kcal/mol) for sequence data. The web interface allows users to select a predictive mode based on the availability of the protein structure or sequence. I-Mutant2.0 is based on SVMs and uses a radial basis function (RBF) kernel. The input vector includes information about temperature, pH, residue types, and the residue environment. The RSA value can be calculated using DSSP when the structure is available. The RI value is computed based on the SVM output. I-Mutant2.0 was trained on 2087 single mutations in 65 proteins, with 58 known at atomic resolution. The accuracy of structure-based predictions is 0.80, while sequence-based predictions achieve 0.77. The correlation between predicted and experimental ΔΔG values is 0.71 for structure-based predictions and 0.62 for sequence-based predictions. The tool allows users to predict the stability change based on the protein structure or sequence, and provides outputs including the predicted ΔΔG value, sign, temperature, pH, and reliability index (RI) or RSA. The results show that I-Mutant2.0 is a valuable tool for protein design, even when the protein structure is not known at atomic resolution. The tool is available at http://gpcr.biocomp.unibo.it/cgi/predictors/I-Mutant2.0/I-Mutant2.0.cgi.