Pushing the boundaries of molecular representation for drug discovery with graph attention mechanism.

By Zhaoping Xiong, Dingyan Wang, Xiaohong Liu, Feisheng Zhong, X. Wan, Xutong Li, Zhaojun Li, Xiaomin Luo, Kaixian Chen, Hualiang Jiang and 1 more. Journal of Medicinal Chemistry, 2020

987 citations recorded by the source corpus.

Hunting for chemicals with favourable pharmacological, toxicological and pharmacokinetic properties remains a formidable challenge for drug discovery. Deep learning provides us with powerful tools to build predictive models that are appropriate for the rising amounts of data, but the gap between what these neural networks learn and what human beings can comprehend is growing. Moreover, this gap may induce distrust and restrict deep learning applications in practice. Here, we introduce a new graph neural network architecture called Attentive FP for molecular representation that uses a graph attention mechanism to learn from relevant drug discovery datasets. We demonstrate that Attentive FP achieves state-of-the-art predictive performances on a variety of datasets and that what it learns is interpretable. The feature visualization for Attentive FP suggests that it automatically learns non-local intramolecular interactions from specified tasks, which can help us gain chemical insights directly from data beyond human perception.

Subfields: Computational Theory and Mathematics, Materials Chemistry, Molecular Biology

View Paper: https://www.semanticscholar.org/paper/fc56572b7524d0b5b0e4770f57185b20d9294136

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Materials Chemistry

In Molecular Biology

Citations in the corpus

Self-Supervised Graph Transformer on Large-Scale Molecular Data (2020, methodology): “AttentiveFP [61] is an extension of the graph attention network.”
Uni-Mol: A Universal 3D Molecular Representation Learning Framework (2023, methodology): “D-MPNN [51] and AttentiveFP [52] are supervised GNNs methods.”
Could graph neural networks learn better molecular representation for drug discovery? A comparison study of descriptor-based and graph-based models (2020, methodology): “Followed by previous studies [27], almost all the easily accessible atom/ bond-level features were exhausted to comprehensively squeeze chemical information into molecular graph for graph-based models, where include nine kinds of atomic features (i.”
Everything is Connected: Graph Neural Networks (2023, background): “(cid:76) A canonical example is classifying molecules for their quantum-chemical properties (Gilmer et al., 2017), estimating pharmacological properties like toxicity or solubility (Duvenaud et al., 2015; Xiong et al., 2019; Jiang et al., 2021) or virtual drug screening (Stokes et al., 2020).”
Identification of antimicrobial peptides from the human gut microbiome using deep learning (2022, background): “The recent surge in applying artificial intelligence approaches in bio-medical research has brought up many new opportunities, including image-processing-related tools for clinical diagnostics of eye 39 , lung 40 , cancer tissue pathologies 41 and antibiotics 19,42 .”
MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design (2022, methodology): “To adapt it to molecular graphs, recent works utilize the attention to better capture the correlation at node, subgraph, and graph-level, including GAT [40], AttentiveFP [41], GTransformer [42], AWARE [43], and Graphormer [44].”
Geometry-enhanced molecular representation learning for property prediction (2021, methodology): “For example, Attentive FP [54] proposes to extend graph attention mechanism 309 in order to learn aggregation weights.”
A high quality, industrial data set for binding affinity prediction: performance comparison in different early drug discovery scenarios (2022, methodology): “The attentive fingerprints (AttentiveFP) method is included in the comparison as a model that learns a representation solely from the small molecule’s 2D graph and has no awareness of the protein’s binding site [19].”
Large-scale chemical language representations capture molecular structure and properties (2021, background): “Baseline performances are taken from references 28,40 .”
Hopfield Networks is All You Need (2020, neutral): “, 2017), and Attentive FP (Xiong et al., 2020).”
A review on machine learning approaches and trends in drug discovery (2021, background): “To avoid this fact, in [64] develop a new graphical neural network architecture for molecular representation that uses a graphical attention mechanism to learn from relevant drug discovery data sets.”
Do Large Scale Molecular Language Representations Capture Important Structural Information? (2021, methodology): “Here we consider the MPNN Gilmer et al. (2017), Attentive-FP Xiong et al. (2019) and baseline GNN models that are more powerful than 1-WL, including 123-gnn Morris et al. (2019) and PPGN Maron et al. (2020).”
Therapeutics Data Commons: Machine Learning Datasets and Tasks for Drug Discovery and Development (2021, methodology): “2015), graph convolutional network (GCN) (Kipf & Welling 2017), and attentive fingerprint (AttentiveFP) (Xiong et al. 2019), three powerful Graph neural network (GNN) models.”