Identification of Protein Methylation Sites Based on Convolutional Neural Network

Protein is an important component of all human cells and tissues. Protein posttranslational modification (PTM) refers to the chemical modification of protein after translation, which changes the biochemical characteristics of protein by adding chemical groups on one or more amino acid residues under the catalysis of enzymes. Protein methylation is a common post-translational modification. Protein methylation modification refers to the process of methyl transfer to specific amino acid residues under the catalysis of methyltransferase. Protein methylation is involved in a variety of biological regulation, in-depth understanding can help to understand its molecular mechanism and various functional roles in cells. Abnormal protein translation can lead to changes in protein structure and function, which is related to the occurrence and development of human diseases. The traditional experimental methods are time-consuming and laborious. In this paper, the characteristics of protein methylation sites of six species are extracted, and the convolutional neural network is used for classification. The appropriate learning rate is selected in the training network to inhibit over-fitting. Under sufficient iterations, a good classification structure is finally obtained. The AUC value calculated by this experiment: BS: 0.945, CG: 0.665, GK: 0.952, MT: 0.957, ST: 1.0. It provides theoretical guidance for the subsequent research on protein methylation site recognition.