Classification of signaling proteins based on molecular star graph descriptors using Machine Learning models

TL;DR

This paper develops a machine learning model using molecular star graph descriptors to accurately classify signaling proteins, aiding drug development by predicting signaling activity from protein structures.

Contribution

It introduces a novel topological descriptor-based classification model for signaling proteins, achieving high accuracy with SVM-RFE and Laplacian kernel.

Findings

Achieved an AUROC of 0.961 with the best model.

Successfully predicted signaling activity for proteins of unknown function.

Identified important signaling pathways with high confidence.

Abstract

Signaling proteins are an important topic in drug development due to the increased importance of finding fast, accurate and cheap methods to evaluate new molecular targets involved in specific diseases. The complexity of the protein structure hinders the direct association of the signaling activity with the molecular structure. Therefore, the proposed solution involves the use of protein star graphs for the peptide sequence information encoding into specific topological indices calculated with S2SNet tool. The Quantitative Structure-Activity Relationship classification model obtained with Machine Learning techniques is able to predict new signaling peptides. The best classification model is the first signaling prediction model, which is based on eleven descriptors and it was obtained using the Support Vector Machines - Recursive Feature Elimination (SVM-RFE) technique with the Laplacian kernel (RFE-LAP) and an AUROC of 0.961. Testing a set of 3114 proteins of unknown function from the PDB database assessed the prediction performance of the model. Important signaling pathways are presented for three UniprotIDs (34 PDBs) with a signaling prediction greater than 98.0%.