Deep Learning-Based Modeling of Drug–Target Interaction Prediction Incorporating Binding Site Information of Proteins
Problem
Predicting unknown interactions between candidate drugs and protein targets in early drug discovery is slow and computationally costly.
Why it matters
Early prediction of drug-target interactions can reduce costly late-stage failures in drug development.
What the AI did
Trained a CNN model (DeepPS) on one-dimensional SMILES drug strings and motif-rich binding-pocket protein subsequences to predict ligand-target interactions.
Results
Better or similar performance on MSE and AUPR versus shallow ML methods; more computationally efficient (no numeric values stated in abstract).
Impact
Evidence supports only benchmark-level improvement over shallow baselines; no real-world drug discovery outcome stated.
Record
| Venue | Interdisciplinary Sciences Computational Life Sciences |
|---|---|
| Published | 2023 |
| Work type | article |
| Authors | Sofia D’Souza, K. V. Prema, S. Balaji, Ronak Shah |
| Citations | 36 (OpenAlex, at import time) |
| References | 46 |
| Model or method | DeepPS (CNN) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1007/s12539-023-00557-z |
| Open access | hybrid |
| Keywords | deep learning, CNN, drug-target interaction, SMILES, chemogenomics, binding site |
Importance rating
significant — thousands to millions affected
Chemogenomics is used in early drug discovery to predict off-target effects of therapeutic candidates; abstract implies computational efficiency matters but states no scale or real-world discovery outcome.