KPhogly-RAM:Prediction of Phosphoglycerylation Sites Using Residue Adjacency Matrix

Post-translational modification (PTM) is a type of covalent alteration that occurs after the biosynthesis process and is crucial for the study of cell biology. Even though a significant number of proteins have been sequenced, phosphoglycerylation recognition remains a major challenge due to factors such as cost, time, and inefficiency in experimental efforts. Lysine phosphoglycerylation is a newly discovered reversible type of PTM that changes glycolyticenzyme activity and is linked to different diseases. In this study, we developed a new computational approach KPhogly-RAM based on sequence-based information about amino acid residues to identify lysin phosphoglycerylation sites in the protein to understand the functionality and causality of phosphoglycerylation sites. We extracted the residue adjacency matrix for each lysine residue in the protein sequences from the data set and encoded it into fused feature vectors before applying recursive feature elimination to enhance prediction quality. It obtained an accuracy score of 0.8877 in repeated stratified ten-fold cross-validation, surpassing most other existing predictors.