Barlow Twins Deep Neural Network for Advanced 1D Drug-Target Interaction Prediction
This work addresses the problem of accelerating drug discovery for researchers and pharmaceutical developers, though it appears incremental as it builds on existing architectures.
The paper tackles drug-target interaction prediction by introducing BarlowDTI, a method using the Barlow Twins architecture for feature extraction from 1D input, which achieves state-of-the-art performance on established benchmarks.
Accurate prediction of drug-target interactions is critical for advancing drug discovery. By reducing time and cost, machine learning and deep learning can accelerate this laborious discovery process. In a novel approach, BarlowDTI, we utilise the powerful Barlow Twins architecture for feature-extraction while considering the structure of the target protein. Our method achieves state-of-the-art predictive performance against multiple established benchmarks using only one-dimensional input. The use of gradient boosting machine as the underlying predictor ensures fast and efficient predictions without the need for substantial computational resources. We also investigate how the model reaches its decision based on individual training samples. By comparing co-crystal structures, we find that BarlowDTI effectively exploits catalytically active and stabilising residues, highlighting the model's ability to generalise from one-dimensional input data. In addition, we further benchmark new baselines against existing methods. Together, these innovations improve the efficiency and effectiveness of drug-target interaction predictions, providing robust tools for accelerating drug development and deepening the understanding of molecular interactions. Therefore, we provide an easy-to-use web interface that can be freely accessed at https://www.bio.nat.tum.de/oc2/barlowdti .