Abstract:Identification of drug-target interactions (DTIs) plays a key role in drug discovery. The high cost and labor-intensive nature of in vitro and in vivo experiments have highlighted the importance of in silico-based DTI prediction approaches. In several computational models, conventional protein descriptors have been shown to not be sufficiently informative to predict accurate DTIs. Thus, in this study, we propose a deep learning based DTI prediction model capturing local residue patterns of proteins participating in DTIs. When we employ a convolutional neural network (CNN) on raw protein sequences, we perform convolution on various lengths of amino acids subsequences to capture local residue patterns of generalized protein classes. We train our model with large-scale DTI information and demonstrate the performance of the proposed model using an independent dataset that is not seen during the training phase. As a result, our model performs better than previous protein descriptor-based models. Also, our model performs better than the recently developed deep learning models for massive prediction of DTIs. By examining pooled convolution results, we confirmed that our model can detect binding sites of proteins for DTIs. In conclusion, our prediction model for detecting local residue patterns of target proteins successfully enriches the protein features of a raw protein sequence, yielding better prediction results than previous approaches. Our code is available at <a href="https://github.com/GIST-CSBL/DeepConv-DTI">https://github.com/GIST-CSBL/DeepConv-DTI</a>.Drugs work by interacting with target proteins to activate or inhibit a target's biological process. Therefore, identification of DTIs is a crucial step in drug discovery. However, identifying drug candidates via biological assays is very time and cost consuming, which introduces the need for a computational prediction approach for the identification of DTIs. In this work, we constructed a novel DTI prediction model to extract local residue patterns of target protein sequences using a CNN-based deep learning approach. As a result, the detected local features of protein sequences perform better than other protein descriptors for DTI prediction and previous models for predicting PubChem independent test datasets. That is, our approach of capturing local residue patterns with CNN successfully enriches protein features from a raw sequence.

PeSTo-Carbs: Geometric Deep Learning for Prediction of Protein-Carbohydrate Binding Interfaces

PeSTo: parameter-free geometric deep learning for accurate prediction of protein binding interfaces

Geometric Transformers for Protein Interface Contact Prediction

Prediction of protein-carbohydrate binding sites from protein primary sequence

Protein interaction interface region prediction by geometric deep learning

GraphsformerCPI: Graph Transformer for Compound–Protein Interaction Prediction

EGGNet, a Generalizable Geometric Deep Learning Framework for Protein Complex Pose Scoring

Protein–protein and protein–nucleic acid binding site prediction via interpretable hierarchical geometric deep learning

A Point Cloud-Based Deep Learning Strategy for Protein-Ligand Binding Affinity Prediction

Structure-based neural network protein-carbohydrate interaction predictions at the residue level

Inferring Protein-Protein Interface from Geometry of Unbound Structures

Protein-Protein Interaction Interface Residue Pair Prediction Based on Deep Learning Architecture

Improving Protein-peptide Interface Predictions in the Low Data Regime

MulinforCPI: enhancing precision of compound–protein interaction prediction through novel perspectives on multi-level information integration

DeepConv-DTI: Prediction of drug-target interactions via deep learning with convolution on protein sequences

Geometric deep learning of protein-DNA binding specificity

Co-evolution Transformer for Protein Contact Prediction.

CarsiDock: a deep learning paradigm for accurate protein-ligand docking and screening based on large-scale pre-training

PepCA: Unveiling protein-peptide interaction sites with a multi-input neural network model

Highly accurate carbohydrate-binding site prediction with DeepGlycanSite

A Point Cloud Graph Neural Network for Protein–Ligand Binding Site Prediction