Recent advancements in predicting protein phosphorylation sites using machine learning methods

Reversible post-translational modification (PTM) plays an important role in increasing the functional diversity of proteins and has a significant influence on the regulation of protein function in prokaryotic and eukaryotic organisms. PTM acts as an important molecular regulatory mechanism used to regulate various cellular processes. Nevertheless, the most studied PTMs can mainly be named after protein types with phosphorylation and important roles in many biological processes. Disturbances in this modification can be caused by a variety of diseases, including neurological disorders and cancer. It is therefore necessary to predict the phosphorylation of target residues in uncharacterized amino acid sequences. Most experimental techniques for predicting phosphorylation are time-consuming, expensive, and error-prone. Various calculation methods have emerged. A very effective and recently developed machine learning method. In general, these methods fall into two categories: conventional machine learning methods and deep learning methods. We have given an overview of both. This review summarizes the research progress in this area. We hope the review could provide important guide for further computational study on protein phosphorylation sites.