Convolutional Embedding of Attributed Molecular Graphs for Physical Property Prediction

By Connor W. Coley, Regina Barzilay, William H. Green Jr., Tommi S. Jaakkola, Klavs F. Jensen. 2017

440 citations recorded by the source corpus.

The task of learning an expressive molecular representation is central to developing quantitative structure-activity and property relationships. Traditional approaches rely on group additivity rules, empirical measurements or parameters, or generation of thousands of descriptors. In this paper, we employ a convolutional neural network for this embedding task by treating molecules as undirected graphs with attributed nodes and edges. Simple atom and bond attributes are used to construct atom-specific feature vectors that take into account the local chemical environment using different neighborhood radii. By working directly with the full molecular graph, there is a greater opportunity for models to identify important features relevant to a prediction task. Unlike other graph-based approaches, our atom featurization preserves molecule-level spatial information that significantly enhances model performance. Our models learn to identify important features of atom clusters for the prediction of aqueous solubility, octanol solubility, melting point, and toxicity. Extensions and limitations of this strategy are discussed.

Subfields: Graph Learning, Computational Chemistry & Drug Discovery

How researchers see this paper

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

In Graph Learning

In Computational Chemistry & Drug Discovery

Citations in the corpus

Mol2vec: Unsupervised Machine Learning Approach with Chemical Intuition (2018, neutral): “However, the best reported results on the ESOL data set were so far obtained by two molecular graph convolution 9,37 39 on this data set 26 but the former two achieved higher sensitivity values (Table 2). existing literature results (Table 3, Table S6).”
Graph Networks as a Universal Machine Learning Framework for Molecules and Crystals (2018, methodology): “30,31 Recently, Faber et al. 32 have benchmarked diﬀerent features in combination with models extensively on the QM9 data set.”
Deep Learning on Graphs: A Survey (2018, neutral): “Besides standard graph inference tasks such as node or graph classification, graph-based deep learning methods have also been applied to a wide range of disciplines, such as modeling social influence [103], recommendation [51], [78], [96], chemistry [37], [41], [50], [98], [99], physics [104], [105]”
A Transdisciplinary Review of Deep Learning Research and Its Relevance for Water Resources Scientists (2017, positive): “Other authors extracted CNN’s hidden layer representation as chemical fingerprints to predict different properties (Coley et al., 2017; Schütt et al., 2017; more discussion in section 3.2.4).”
Constrained Bayesian Optimization for Automatic Chemical Design (2017, background): “[2017], Coley et al. [2019b], Schwaller et al.”
The rise of deep learning in drug discovery. (2018, background): “Recently, researchers from Google [54] reformulated several existing graph convolution algorithms [49,50,53,55,56] into a common framework known as a message passing neural network (MPNN) and used the MPNNs to predict quantum chemical properties.”
The rise of artificial intelligence in healthcare applications (2020, background): “In this way, molecular properties including octanol, solubility melting point, and biological activity can be evaluated as demonstrated by Coley et al. and others and be used to predict new features of the drug molecules [18].”
Deep Evidential Regression (2019, methodology): “This method is widely applicable across regression tasks including temporal forecasting [17], property prediction [8], and control learning [1, 30].”
Geometry-enhanced molecular representation learning for property prediction (2021, methodology): “Molecular fingerprints, such as ECFP [39] and MACCS [12], are com- 296 monly used for molecular representations by traditional machine learning methods [6, 8, 13, 24], 297 which encode a molecule into a sequence of bits according to the molecules’ topological substructures.”