Predicting the pharmacological activity, toxicity, and pharmacokinetic properties of molecules is a central task in drug discovery. Existing machine learning methods are transferred from one resource rich molecular property to another data scarce property in the same scaffold dataset. However, existing models may produce fragile and highly uncertain predictions for new scaffold molecules. And these models were tested on different benchmarks, which seriously affected the quality of their evaluation results. In this article, we introduce Meta-MolNet, a collection of data benchmark and algorithms, which is a standard benchmark platform for measuring model generalization and uncertainty quantification capabilities. Meta-MolNet manages a wide range of molecular datasets with high ratio of molecules/scaffolds, which often leads to more difficult data shift and generalization problems. Furthermore, we propose a graph attention network based on cross-domain meta-learning, Meta-GAT, which uses bilevel optimization to learn meta-knowledge from the scaffold family molecular dataset in the source domain. Meta-GAT benefits from meta-knowledge that reduces the requirement of sample complexity to enable reliable predictions of new scaffold molecules in the target domain through internal iteration of a few examples. We evaluate existing methods as baselines for the community, and the Meta-MolNet benchmark demonstrates the effectiveness of measuring the proposed algorithm in domain generalization and uncertainty quantification. Extensive experiments demonstrate that the Meta-GAT model has state-of-the-art domain generalization performance and robustly estimates uncertainty under few examples constraints. By publishing AI-ready data, evaluation frameworks, and baseline results, we hope to see the Meta-MolNet suite become a comprehensive resource for the AI-assisted drug discovery community. Meta-MolNet is freely accessible at https://github.com/lol88/Meta-MolNet.

Motivation: Advances in single-cell measurement techniques provide rich multimodal data, which helps us to explore the life state of cells more deeply. However, multimodal integration, or, learning joint embeddings from multimodal data remains a current challenge. The difficulty in integrating unpaired single-cell multimodal data is that different modalities have different feature spaces, which easily leads to information loss in joint embedding. And few existing methods have fully exploited and fused the information in single-cell multimodal data. Result: In this study, we propose CoVEL, a deep learning method for unsupervised integration of single-cell multimodal data. CoVEL learns single-cell representations from a comprehensive view, including regulatory relationships between modalities, fine-grained representations of cells, and relationships between different cells. The comprehensive view embedding enables CoVEL to remove the gap between modalities while protecting biological heterogeneity. Experimental results on multiple public datasets show that CoVEL is accurate and robust to single-cell multimodal integration. Data availability: https://github.com/shapsider/scintegration.

Inductive bias in machine learning (ML) is the set of assumptions describing how a model makes predictions. Different ML-based methods for protein-ligand binding affinity (PLA) prediction have different inductive biases, leading to different levels of generalization capability and interpretability. Intuitively, the inductive bias of an ML-based model for PLA prediction should fit in with biological mechanisms relevant for binding to achieve good predictions with meaningful reasons. To this end, we propose an interaction-based inductive bias to restrict neural networks to functions relevant for binding with two assumptions: 1) A protein-ligand complex can be naturally expressed as a heterogeneous graph with covalent and non-covalent interactions; 2) The predicted PLA is the sum of pairwise atom-atom affinities determined by non-covalent interactions. The interaction-based inductive bias is embodied by an explainable heterogeneous interaction graph neural network (EHIGN) for explicitly modeling pairwise atom-atom interactions to predict PLA from 3D structures. Extensive experiments demonstrate that EHIGN achieves better generalization capability than other state-of-the-art ML-based baselines in PLA prediction and structure-based virtual screening. More importantly, comprehensive analyses of distance-affinity, pose-affinity, and substructure-affinity relations suggest that the interaction-based inductive bias can guide the model to learn atomic interactions that are consistent with physical reality. As a case study to demonstrate practical usefulness, our method is tested for predicting the efficacy of Nirmatrelvir against SARS-CoV-2 variants. EHIGN successfully recognizes the changes in the efficacy of Nirmatrelvir for different SARS-CoV-2 variants with meaningful reasons.

Finding candidate molecules with favorable pharmacological activity, low toxicity, and proper pharmacokinetic properties is an important task in drug discovery. Deep neural networks have made impressive progress in accelerating and improving drug discovery. However, these techniques rely on a large amount of label data to form accurate predictions of molecular properties. At each stage of the drug discovery pipeline, usually, only a few biological data of candidate molecules and derivatives are available, indicating that the application of deep neural networks for low-data drug discovery is still a formidable challenge. Here, we propose a meta learning architecture with graph attention network, Meta-GAT, to predict molecular properties in low-data drug discovery. The GAT captures the local effects of atomic groups at the atom level through the triple attentional mechanism and implicitly captures the interactions between different atomic groups at the molecular level. GAT is used to perceive molecular chemical environment and connectivity, thereby effectively reducing sample complexity. Meta-GAT further develops a meta learning strategy based on bilevel optimization, which transfers meta knowledge from other attribute prediction tasks to low-data target tasks. In summary, our work demonstrates how meta learning can reduce the amount of data required to make meaningful predictions of molecules in low-data scenarios. Meta learning is likely to become the new learning paradigm in low-data drug discovery. The source code is publicly available at: https://github.com/lol88/Meta-GAT.

Drug-drug interactions (DDIs) trigger unexpected pharmacological effects in vivo, often with unknown causal mechanisms. Deep learning methods have been developed to better understand DDI. However, learning domain-invariant representations for DDI remains a challenge. Generalizable DDI predictions are closer to reality than source domain predictions. For existing methods, it is difficult to achieve out-of-distribution (OOD) predictions. In this article, focusing on substructure interaction, we propose DSIL-DDI, a pluggable substructure interaction module that can learn domain-invariant representations of DDIs from source domain. We evaluate DSIL-DDI on three scenarios: the transductive setting (all drugs in test set appear in training set), the inductive setting (test set contains new drugs that were not present in training set), and OOD generalization setting (training set and test set belong to two different datasets). The results demonstrate that DSIL-DDI improve the generalization and interpretability of DDI prediction modeling and provides valuable insights for OOD DDI predictions. DSIL-DDI can help doctors ensuring the safety of drug administration and reducing the harm caused by drug abuse.

MOTIVATION: Large-scale prediction of drug-target affinity (DTA) plays an important role in drug discovery. In recent years, machine learning algorithms have made great progress in DTA prediction by utilizing sequence or structural information of both drugs and proteins. However, sequence-based algorithms ignore the structural information of molecules and proteins, while graph-based algorithms are insufficient in feature extraction and information interaction. RESULTS: In this article, we propose NHGNN-DTA, a node-adaptive hybrid neural network for interpretable DTA prediction. It can adaptively acquire feature representations of drugs and proteins and allow information to interact at the graph level, effectively combining the advantages of both sequence-based and graph-based approaches. Experimental results have shown that NHGNN-DTA achieved new state-of-the-art performance. It achieved the mean squared error (MSE) of 0.196 on the Davis dataset (below 0.2 for the first time) and 0.124 on the KIBA dataset (3% improvement). Meanwhile, in the case of cold start scenario, NHGNN-DTA proved to be more robust and more effective with unseen inputs than baseline methods. Furthermore, the multi-head self-attention mechanism endows the model with interpretability, providing new exploratory insights for drug discovery. The case study on Omicron variants of SARS-CoV-2 illustrates the efficient utilization of drug repurposing in COVID-19. AVAILABILITY AND IMPLEMENTATION: The source code and data are available at https://github.com/hehh77/NHGNN-DTA.

MOTIVATION: Drug-drug interactions (DDIs) occur during the combination of drugs. Identifying potential DDI helps us to study the mechanism behind the combination medication or adverse reactions so as to avoid the side effects. Although many artificial intelligence methods predict and mine potential DDI, they ignore the 3D structure information of drug molecules and do not fully consider the contribution of molecular substructure in DDI. RESULTS: We proposed a new deep learning architecture, 3DGT-DDI, a model composed of a 3D graph neural network and pre-trained text attention mechanism. We used 3D molecular graph structure and position information to enhance the prediction ability of the model for DDI, which enabled us to deeply explore the effect of drug substructure on DDI relationship. The results showed that 3DGT-DDI outperforms other state-of-the-art baselines. It achieved an 84.48% macro F1 score in the DDIExtraction 2013 shared task dataset. Also, our 3D graph model proves its performance and explainability through weight visualization on the DrugBank dataset. 3DGT-DDI can help us better understand and identify potential DDI, thereby helping to avoid the side effects of drug mixing. AVAILABILITY: The source code and data are available at https://github.com/hehh77/3DGT-DDI.

The prediction of drug-target affinity (DTA) plays an increasingly important role in drug discovery. Nowadays, lots of prediction methods focus on feature encoding of drugs and proteins, but ignore the importance of feature aggregation. However, the increasingly complex encoder networks lead to the loss of implicit information and excessive model size. To this end, we propose a deep-learning-based approach namely FusionDTA. For the loss of implicit information, a novel muti-head linear attention mechanism was utilized to replace the rough pooling method. This allows FusionDTA aggregates global information based on attention weights, instead of selecting the largest one as max-pooling does. To solve the redundancy issue of parameters, we applied knowledge distillation in FusionDTA by transfering learnable information from teacher model to student. Results show that FusionDTA performs better than existing models for the test domain on all evaluation metrics. We obtained concordance index (CI) index of 0.913 and 0.906 in Davis and KIBA dataset respectively, compared with 0.893 and 0.891 of previous state-of-art model. Under the cold-start constrain, our model proved to be more robust and more effective with unseen inputs than baseline methods. In addition, the knowledge distillation did save half of the parameters of the model, with only 0.006 reduction in CI index. Even FusionDTA with half the parameters could easily exceed the baseline on all metrics. In general, our model has superior performance and improves the effect of drug-target interaction (DTI) prediction. The visualization of DTI can effectively help predict the binding region of proteins during structure-based drug design.

With the rapid development of proteomics and the rapid increase of target molecules for drug action, computer-aided drug design (CADD) has become a basic task in drug discovery. One of the key challenges in CADD is molecular representation. High-quality molecular expression with chemical intuition helps to promote many boundary problems of drug discovery. At present, molecular representation still faces several urgent problems, such as the polysemy of substructures and unsmooth information flow between atomic groups. In this research, we propose a deep contextualized Bi-LSTM architecture, Mol2Context-vec, which can integrate different levels of internal states to bring dynamic representations of molecular substructures. And the obtained molecular context representation can capture the interactions between any atomic groups, especially a pair of atomic groups that are topologically distant. Experiments show that Mol2Context-vec achieves state-of-the-art performance on multiple benchmark datasets. In addition, the visual interpretation of Mol2Context-vec is very close to the structural properties of chemical molecules as understood by humans. These advantages indicate that Mol2Context-vec can be used as a reliable and effective tool for molecular expression. Availability: The source code is available for download in https://github.com/lol88/Mol2Context-vec.

Although there is a growing recognition of the differences, not diminished abilities, of facial affective expressivity between Typically Developing (TD) and Autism Spectrum Disorder (ASD) individuals, which might lead to the varied recognizability of conveyed emotion by both TD and ASD individuals, little is explored on the ecological validity of these findings; that is, whether spontaneous affective facial expressions can better be produced and recognized by both populations. We aimed to address these issues in the present study, using children's cartoon clips to assess two aspects of spontaneous emotion production and recognition in a context closer to real-life children's cartoon movie watching (at home or a classroom). Based on the facial landmark data and a teacher/parent's manual emotion tags (happy), we performed a computational analysis to compare the happy emotion labels generated by the automated algorithm and the human TD rater. Two pilot studies of six ASD children revealed the potential as well as challenges of such an approach.