The thermodynamic data for proteins can be useful for understanding the mechanism of protein folding and stability, and for designing stable mutants. Due to the advancement in molecular biology and biochemistry, a large number of proteins have been characterized, and hence the accumulation of thermodynamic data has been increasing. On the other hand, databases of thermodynamic datasets along with the sequence and structural information would be a valuable resource for developing new algorithms (or methods) to elucidate the mechanism of protein folding and stability and to predict the stability change upon mutations.

Predictions of the change on Gibbs free energy for different amino acids mutations in proteins using machine learning approach.

Figure: Predictions of the change on Gibbs free energy for different amino acids mutations in proteins. The source code of the software can be downloaded following this link: BSANN