Discrimination of thermophilic and mesophilic proteins via artificial neural networks

Discrimination of thermophilic and mesophilic proteins is still a challenge in the research of the protein pattern recognition. That is because that the discrimination in the level of amino acid sequence would be helpful to knowing better the folding mechanism and the function of protein. What is more, it would be an important assistant to reconstruct the stability of protein. In this paper, a feature extraction method came to be true which considered the fusion of amino acids models and chem-composition models. Meanwhile, the feedforward artificial neural network which was optimized by particle swarm optimization (PSO-NN) was introduced to discriminate protein thermostsbility. Compared with the previous research, the discrimination accuracy had been improved to some extent. The result reflect that on the basis of amino acids models, chem-composition models plays an important part in improving the accuracy, and PSO-NN should be regard as an effective tool for recognition of mesophilic and thermophilic proteins.