DeepDrug3D: Classification of ligand-binding pockets in proteins with a convolutional neural network

Comprehensive characterization of ligand-binding sites is invaluable to infer molecular functions of hypothetical proteins, trace evolutionary relationships between proteins, engineer enzymes to achieve a desired substrate specificity, and develop drugs with improved selectivity profiles. These research efforts pose significant challenges owing to the fact that similar pockets are commonly observed across different folds, leading to the high degree of promiscuity of ligand-protein interactions at the system-level. On that account, novel algorithms to accurately classify binding sites are needed. Deep learning is attracting a significant attention due to its successful applications in a wide range of disciplines. In this communication, we present DeepDrug3D, a new approach to characterize and classify binding pockets in proteins with deep learning. It employs a state-of-the-art convolutional neural network in which biomolecular structures are represented as voxels assigned interaction energy-based attributes. The current implementation of DeepDrug3D, trained to detect and classify nucleotide- and heme-binding sites, not only achieves a high accuracy of 95%, but also has the ability to generalize to unseen data as demonstrated for steroid-binding proteins and peptidase enzymes. Interestingly, the analysis of strongly discriminative regions of binding pockets reveals that this high classification accuracy arises from learning the patterns of specific molecular interactions, such as hydrogen bonds, aromatic and hydrophobic contacts. DeepDrug3D is available as an open-source program at https://github.com/pulimeng/DeepDrug3D with the accompanying TOUGH-C1 benchmarking dataset accessible from https://osf.io/enz69/.

Everyone Is a Protagonist: Residue Conformational Preferences in High-Resolution Protein Structures

In many structural bioinformatics problems, there is a broad range of unanswered questions about protein dynamics and amino acid properties. Proteins are not strictly static objects, but rather populate ensembles of conformations. One way to understand these particularities is to analyze the information available in experimental databases. The Ramachandran plot, despite being more than half a century old, remains an utterly useful tool in the study of protein conformation. Based on its assumptions, we inspected a large data set (11,130 protein structures, amounting to 5,255,768 residues) and discriminated the conformational preferences of each residue type regarding their secondary structure participation. These data were studied for phi [Formula: see text], psi [Formula: see text], and side chain chi [Formula: see text] angles, being presented in non-Ramachandranian plots. In the largest analysis of protein conformation made so far, we propose an original plot to depict conformational preferences in relation to different secondary structure elements. Despite confirming previous observations, our results strongly support a unique character for each residue type, whereas also reinforcing the observation that side chains have a major contribution to secondary structure and, by consequence, on protein conformation. This information can be further used in the development of more robust methods and computational strategies for structural bioinformatics problems.