Structure-Aware Multimodal Deep Learning for Drug–Protein Interaction Prediction

Abstract

Identifying drug−protein interactions (DPIs) is crucial in drug discovery, and a number of machine learning methods have been developed to predict DPIs. Existing methods usually use unrealistic data sets with hidden bias, which will limit the accuracy of virtual screening methods. Meanwhile, most DPI prediction methods pay more attention to molecular representation but lack effective research on protein representation and high-level associations between different instances. To this end, we present the novel structure-aware multimodal deep DPI prediction model, STAMP-DPI, which was trained on a curated industry-scale benchmark data set. We built a high-quality benchmark data set named GalaxyDB for DPI prediction. This industry-scale data set along with an unbiased training procedure resulted in a more robust benchmark study. For informative protein representation, we constructed a structure-aware graph neural network method from the protein sequence by combining predicted contact maps and graph neural networks. Through further integration of structure-based representation and high-level pretrained embeddings for molecules and proteins, our model effectively captures the feature representation of the interactions between them. As a result, STAMP-DPI outperformed state-of-the-art DPI prediction methods by decreasing 7.00% mean square error (MSE) in the Davis data set and improving 8.89% area under the curve (AUC) in the GalaxyDB data set. Moreover, our model is an interpretable model with the transformer-based interaction mechanism, which can accurately reveal the binding sites between molecules and proteins.

TANKBind: Trigonometry-Aware Neural NetworKs for Drug-Protein Binding Structure Prediction

“…We compare TankBind against two state-of-the-art sequence-based methods, TransformerCPI Chen et al [2020b] and MONN Li et al [2020], two complex-based methods, IGN Jiang et al [2021] and PIGNet Moon et al [2022] both requiring prior knowledge of the inter-molecular structure to predict affinity, and two structure-based methods, HOLOPTOT Somnath et al [2021] and STAMPDPI Wang et al [2022]. For evaluating various methods, we use four metrics – root mean squared error (RMSE), Pearson correlation coefficient, Spearman correlation coefficient and mean absolute error (MAE).…”

Alignment-free metal ion-binding site prediction from protein sequence through pretrained language model and multi-task learning

“…For example, the prediction can help narrow down potential binding sites for further wet experiment validation [47] or provide hypotheses and insights for the mechanisms of disease-causing gene mutations [48] as shown in other similar fields. The binding site prediction can also be used for druggability prediction [49] or de novo drug design [50][51][52]. In the future, we would further extend our method to predict various functional sites, including binding sites with proteins [38] and nucleic acids [39].…”

Unsupervised Protein-Ligand Binding Energy Prediction via Neural Euler's Rotation Equation

“…Energy functions for small molecules fall into two categories: supervised and unsupervised learning. Supervised models are trained on binding affinity data from PDBBind [1][2][3][12][13][14][15]. Their input is typically a protein-ligand 3D complex represented as a geometric graph, which is embedded into a latent representation by a neural network for affinity prediction.…”