A Novel Statistical Method for Thermostable Proteins Discrimination

In this study, we used features that can be extracted from protein sequences to discriminate mesophilic, thermophilic and hyper-thermophilic proteins. Amino acid frequency, dipeptide amino acid frequency and physical-chemical features are used in this study. The effect of mentioned features on proposed discrimination algorithm was evaluated both separately and in combination. Statistical methods are used in the proposed algorithm. The results of implementing the algorithm on a dataset containing 239 mesophilic proteins, 69 thermophilic proteins and 59 hyper-thermophilic proteins show the effect of each bunch of features on the evaluation measures.