BindingSite-AugmentedDTA: enabling a next-generation pipeline for interpretable prediction models in drug repurposing

Abstract

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 (DL)-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 four widely used evaluation metrics, including concordance index, mean squared error, modified squared correlation coefficient ($r^2_m$) and the area under the precision curve. We also contribute to three benchmark drug–traget interaction datasets by including additional information on 3D structure of all proteins contained in those datasets, which include the two most commonly used datasets, namely Kiba and Davis, as well as the data from IDG-DREAM drug-kinase binding prediction challenge. Furthermore, we experimentally validate the practical potential of our proposed framework through in-lab experiments. 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.

Acoustofluidic Engineering of Functional Vessel-on-a-Chip

“…In the permeability computational model, the 70 kDa FITC-dextran solute is represented as a series of discrete beads uniformly distributed on one side of the simulation domain (Figure A green beads). The Cooke-Deserno three-bead membrane representation model is implemented to simulate the vascular membrane, which offers an elegant balance between computational simplicity and biological realism. − The interaction between these entities is defined by employing the Lennard-Jones potential, allowing for the manifestation of Brownian motion among the beads. The use of the Lennard-Jones potential not only provides a mechanism for interaction between the beads but also models the diffusion of these bead-represented solutes through the membrane. , Within this framework, each lipid is depicted as three discrete beads, representing the hydrophilic headgroup (red in Figure ) and the two hydrophobic tail groups (blue in Figure ), effectively capturing the amphiphilic nature of lipids.…”

Microphysiologically Engineered Vessel-Tumor Model to Investigate Vascular Transport Dynamics of Immune Cells

“…The vascular system is a natural mechanism in the immune response that heavily depends on its proper functioning. − It enables immune cells and signaling molecules to move through the blood vessels and reach sites of cancerous cell occurrence or injury, facilitating an effective immune response. Furthermore, the efficiency of immune therapy is influenced by the internal vessel network, which exhibits significant diversity between healthy and tumor tissues and the impact on immune cell recruitment. − The endothelial cell surface is typically coated with a layer of polysaccharides, which serves as the interface between the blood and the endothelium. − The endothelial glycocalyx (VGCX) influences the immune cell behavior during vascular transport, and the inherent features of VGCX are evident from experimental and computational models. − Since most leukocyte adhesion receptors (cell adhesion molecules, CAMs) are located in postcapillary venules, such as intercellular adhesion molecule (ICAM), endothelial-leukocyte adhesion molecule (E-Selectin), and P-Selectin, shedding of the venular glycocalyx may play an important role during the immune response. ,− Thus, the ability to model the structure and transport function of blood vessels in a comprehensive setting is important for understanding the physiological processes of immunotherapy. , …”