A study on the discrimination of thermophilic and mesophilic proteins based on dipeptide composition / 生物工程学报

ABSTRACT

In this work, the dipeptide composition of 3216 thermophilic and 4007 mesophilic protein sequences was systematically analyzed. We found that the thermophilic proteins contained more dipeptides such as EE, EK, KE, VE, EI, KI, EV, KK, VK and IE, whereas less dipeptides such as AA, LL, LA, AL, QA, QL, AQ, LT, TL and EQ. Based on this information, a statistical method for discriminating thermophilic and mesophilic proteins was developed. Our approach correctly picked up the thermophilic proteins with the accuracy of 94.0% and 89%, respectively, for the testing sets of 382 and 73 thermophilic proteins. And for the testing 325 and 73 mesophilic proteins, the accuracy was 85.2% and 89%, respectively. The influence of specific dipeptides on discrimination was also discussed.