A model for predicting post-translational modification cross-talk based on the Multilayer Network

Abstract

Post-translational modifications (PTMs) are pivotal in controlling protein function, signaling pathways, and cellular processes, underscoring their importance in biological systems. PTMs not only regulate various signaling pathways by modifying individual residues but also regulate signaling pathways through the interaction of different modified residues within proteins or between proteins, which is known as PTM cross-talk. An in-depth study of the interactions between PTMs can lead to a clearer understanding of the regulatory mechanisms mediated by PTMs. Therefore, accurately identifying potential PTM cross-talk within proteins (Intra PTM cross-talk) or between proteins (Inter PTM cross-talk) is of utmost importance in biological research. In this work, we introduce an innovative approach called WPTCMN/PTCMN for simultaneous prediction of Intra/Inter PTM cross-talk using an integrated deep neural network, which is based on a Multilayer Network. Comprehensive experimental analysis demonstrates that using the Multilayer Network to capture the complex associations between Intra/Inter PTM cross-talk exhibits remarkable superiority in predicting PTM cross-talk. Specifically, the AUC value achieved on Intra PTM cross-talk is 0.924, while on Inter PTM cross-talk it reaches 0.872, surpassing existing methods. Therefore, WPTCMN/PTCMN represents an effective tool for simultaneous prediction of Intra/Inter PTM cross-talk.