RESUMO
While research into Drug-Target Interaction (DTI) prediction is fairly mature, generalizability and interpretability are not always addressed in the existing works in this field. In this paper, we propose a deep learning-based framework, called BindingSite-AugmentedDTA, which improves Drug-Target Affinity (DTA) predictions by reducing the search space of potential binding sites of the protein, thus making the binding affinity prediction more efficient and accurate. Our BindingSite-AugmentedDTA is highly generalizable as it can be integrated with any DL-based regression model, while it significantly improves their prediction performance. Also, unlike many existing models, our model is highly interpretable due to its architecture and self-attention mechanism, which can provide a deeper understanding of its underlying prediction mechanism by mapping attention weights back to protein binding sites. The computational results confirm that our framework can enhance the prediction performance of seven state-of-the-art DTA prediction algorithms in terms of 4 widely used evaluation metrics, including Concordance Index (CI), Mean Squared Error (MSE), modified squared correlation coefficient [Formula], and the Area Under the Precision Curve (AUPC). We also contribute to the two most commonly used DTA benchmark datasets, namely Kiba and Davis, by including additional information on 3D structure of all proteins contained in these two datasets. We manually extracted this information from Protein Data Bank (PDB) files of proteins available at https://www.uniprot.org/. Furthermore, we experimentally validate the practical potential of our proposed framework through in-lab experiments. We measure the binding interaction between several drug candidate compounds for the inhibition of binding between (SARS-CoV-2 S-protein RBD) Spike and ACE-2 (host cell binding target) proteins. We then compare the computationally-predicted results against the ones experimentally-observed in the laboratory. The relatively high agreement between computationally-predicted and experimentally-observed binding interactions supports the potential of our framework as the next-generation pipeline for prediction models in drug repurposing.
