Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules
By Rafael Gómez‐Bombarelli, David Kristjanson Duvenaud, José Miguel Hernández-Lobato, Jorge Aguilera-Iparraguirre, Timothy D. Hirzel, Ryan P. Adams, Alán Aspuru-Guzik. 2016
3,556 citations recorded by the source corpus.
We report a method to convert discrete representations of molecules to and from a multidimensional continuous representation. This model allows us to generate new molecules for efficient exploration and optimization through open-ended spaces of chemical compounds. A deep neural network was trained on hundreds of thousands of existing chemical structures to construct three coupled functions: an encoder, a decoder, and a predictor. The encoder converts the discrete representation of a molecule into a real-valued continuous vector, and the decoder converts these continuous vectors back to discrete molecular representations. The predictor estimates chemical properties from the latent continuous vector representation of the molecule. Continuous representations of molecules allow us to automatically generate novel chemical structures by performing simple operations in the latent space, such as decoding random vectors, perturbing known chemical structures, or interpolating between molecules. Continuous representations also allow the use of powerful gradient-based optimization to efficiently guide the search for optimized functional compounds. We demonstrate our method in the domain of drug-like molecules and also in a set of molecules with fewer that nine heavy atoms.
Subfields: Computational Theory and Mathematics, Environmental Chemistry, Materials Chemistry, Generative Models, 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 Computational Theory and Mathematics
- Tommi S. Jaakkola
- Regina Barzilay
- Shitong Luo
- Jure Leskovec
- José Miguel Hernández-Lobato
- Wengong Jin
- Gisbert Schneider
- William L. Hamilton
In Environmental Chemistry
- José Miguel Hernández-Lobato
- Shitong Luo
- Jian Tang
- Minkai Xu
- Tommi S. Jaakkola
- Ryan-Rhys Griffiths
- Erik A. Daxberger
- Chence Shi
In Materials Chemistry
- Tommi S. Jaakkola
- Regina Barzilay
- Shitong Luo
- William L. Hamilton
- Jure Leskovec
- Wengong Jin
- José Miguel Hernández-Lobato
- Gisbert Schneider
In Generative Models
- Stefano Ermon
- Arash Vahdat
- Tommi S. Jaakkola
- Diederik P. Kingma
- Jiaming Song
- William L. Hamilton
- Regina Barzilay
- Max Welling
In Computational Chemistry & Drug Discovery
- Tommi S. Jaakkola
- Regina Barzilay
- José Miguel Hernández-Lobato
- Sergey I. Nikolenko
- Wengong Jin
- Gisbert Schneider
- Philippe Schwaller
- Jure Leskovec
Citations in the corpus
- Sequence Tutor: Conservative Fine-Tuning of Sequence Generation Models with KL-control (2016, methodology): “However, generating valid molecules can prove difficult, as it is hard for probabilistic models to learn all the constraints that define physically realizable molecules directly from data (G´omez-Bombarelli et al., 2016).”
- Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation (2018, background): “JT-VAE has been shown to outperform previous deep generative models for molecule generation, including Character-VAE [9], Grammar-VAE [22], SD-VAE [4] and GraphVAE [39].”
- Opportunities and obstacles for deep learning in biology and medicine (2017, methodology): “In other cases, researchers limited the size of their neural network [27] or the total number of training instances [314].”
- Learning Multimodal Graph-to-Graph Translation for Molecular Optimization (2018, methodology): “• VSeq2Seq: Our second baseline is a variational sequence-to-sequence translation model that uses SMILES strings to represent molecules and has been successfully applied to other molecule generation tasks (Gómez-Bombarelli et al., 2016).”
- Junction Tree Variational Autoencoder for Molecular Graph Generation (2018, methodology): “Baselines We compare our approach with SMILES-based baselines: 1) Character VAE (CVAE) (Gómez-Bombarelli et al., 2016) which generates SMILES strings character by character; 2) Grammar VAE (GVAE) (Kusner et al., 2017) that generates SMILES following syntactic constraints given by a context-free…”
- Generating functional protein variants with variational autoencoders (2020, methodology): “311 VAEs are able to capture higher-order dependencies, can be adapted to straightforwardly and 312 flexibly incorporate conditioning information, and learn continuous latent spaces which offer various 313 possibilities for controlled generation, including the local sampling strategy employed here, ”
- Learning Deep Generative Models of Graphs (2018, methodology): “For instance, sampling from the graph model can lead to the discovery of new configurations that share same global properties as is, for example, required in drug discovery (Gómez-Bombarelli et al., 2016).”
- GANS for Sequences of Discrete Elements with the Gumbel-softmax Distribution (2016, background): “For example, text sentences [1], molecules encoded in the SMILE language [4], etc.”