Search

Graph neural networks (GNNs) constitute a class of deep learning methods for graph data. They have wide applications in chemistry and biology, such as molecular property prediction, reaction prediction, and drug-target interaction prediction. Despite the interest, GNN-based modeling is challenging as it requires graph data preprocessing and modeling in addition to programming and deep learning. Here, we present Deep Graph Library (DGL)-LifeSci, an open-source package for deep learning on graphs in life science. Deep Graph Library (DGL)-LifeSci is a python toolkit based on RDKit, PyTorch, and Deep Graph Library (DGL). DGL-LifeSci allows GNN-based modeling on custom datasets for molecular property prediction, reaction prediction, and molecule generation. With its command-line interfaces, users can perform modeling without any background in programming and deep learning. We test the command-line interfaces using standard benchmarks MoleculeNet, USPTO, and ZINC. Compared with previous implementations, DGL-LifeSci achieves a speed up by up to 6×. For modeling flexibility, DGL-LifeSci provides well-optimized modules for various stages of the modeling pipeline. In addition, DGL-LifeSci provides pretrained models for reproducing the test experiment results and applying models without training. The code is distributed under an Apache-2.0 License and is freely accessible at https://github.com/awslabs/dgl-lifesci.

Alpha-synuclein (α-syn), as a highly soluble presynaptic protein expressed in the brain, plays an important role in recycling synaptic vesicles and regulating the synthesis, storage, and release of neurotransmitters. Accumulation of α-syn in Lewy bodies and Lewy neurites is the pathological hallmark of Parkinson’s disease (PD), so inhibition of α-syn aggregation may provide a novel approach for treating PD. In this study, the 3D structure of α-syn was downloaded from Protein Data Bank (PDB ID: 2N0A). A ligand-based pharmacophore model was conducted on a set of 43 diverse α-syn ligands, and the results suggested that two hydrogen-bond acceptors, one hydrophobic group, and two aromatic rings were significant to the inhibition of α-syn aggregation. A ligand-based 3D-QSAR model was also established with good statistical significance (R2 = 0.920) and excellent predictive ability (Q2 = 0.752). Novel indolinone derivatives were designed and synthesized based on the pharmacophore model. Subsequently, the 3D-QSAR model was used to predict the inhibitory activities towards α-syn aggregation, and the actual inhibitory activities were evaluated by thioflavin-T assay in vitro with the best inhibitory activity reaching 45.08%. The fitting results indicated that the built pharmacophore and 3D-QSAR models provided better reliability and accuracy for compound modification and prediction of the activity thereof. A ligand-based pharmacophore modeling and 3D-QSAR study have been performed on a set of 43 diverse ligands for α-synuclein for the first time. Based on the best pharmacophore modeling, novel indolinone derivatives were designed and synthesized, and the inhibitory activities for α-synuclein aggregation were evaluated by thioflavin-T assay in vitro, which preliminary indicated that five pharmacophore sites (two hydrogen bond acceptors (A), a hydrophobic group (H), and two aromatic rings (R)) in compounds contribute to the inhibitory activities. In the study, the built pharmacophore modeling and 3D-QSAR provided better reliability and accuracy for compound modification and prediction of the activity thereof.

As a result of the advent of high-throughput technologies, there has been rapid progress in our understanding of the genetics underlying biological processes. However, despite such advances, the genetic landscape of human diseases has only marginally been disclosed. Exploiting the present availability of large amounts of biological and phenotypic data, we can use our current understanding of disease genetics to train machine learning models to predict novel genetic factors associated with the disease. To this end, we developed DGLinker, a webserver for the prediction of novel candidate genes for human diseases given a set of known disease genes. DGLinker has a user-friendly interface that allows non-expert users to exploit biomedical information from a wide range of biological and phenotypic databases, and/or to upload their own data, to generate a knowledge-graph and use machine learning to predict new disease-associated genes. The webserver includes tools to explore and interpret the results and generates publication-ready figures. DGLinker is available at https://dglinker.rosalind.kcl.ac.uk. The webserver is free and open to all users without the need for registration., Graphical Abstract Graphical AbstractDGLinker: flexible knowledge-graph prediction of disease-gene associations.

Epoxidation is one of the reactions in drug metabolism. Since epoxide metabolites would bind with proteins or DNA covalently, drugs should avoid epoxidation metabolism in the body. Due to the instability of epoxide, it is difficult to determine epoxidation experimentally. In silico models based on big data and machine learning methods are hence valuable approaches to predict whether a compound would undergo epoxidation. In this study, we collected 884 epoxidation data manually from various sources, and finally got 829 unique sites of epoxidation. Three types of molecular fingerprints with different lengths (1024, 2048 or 4096 bits) were used to describe the reaction sites. Six machine learning methods were used to build the classification models. The training set and test set were randomly divided into 8 : 2, and 54 models were constructed and evaluated. Four best models were selected for feature selection. The features were then chosen and verified by external validation set. The resulted optimal model had the accuracy and AUC (area under the curve) values at 0.873 and 0.944 for the test set, 0.838 and 0.987 for the external validation set, respectively. The models built in this study could accurately predict whether a compound will undergo epoxidation and which part is most susceptible to epoxidation, which is of great significance for drug design.