Automatic chemical design using a data-driven continuous representation of molecules
| Autor: | Timothy D. Hirzel, David Duvenaud, Alán Aspuru-Guzik, Dennis Sheberla, Benjamin Sanchez-Lengeling, Jennifer N. Wei, Ryan P. Adams, Jorge Aguilera-Iparraguirre, José Miguel Hernández-Lobato, Rafael Gómez-Bombarelli |
|---|---|
| Přispěvatelé: | Gómez-Bombarelli, Rafael [0000-0002-9495-8599], Wei, Jennifer N [0000-0003-3567-9511], Sheberla, Dennis [0000-0002-5239-9151], Aspuru-Guzik, Alán [0000-0002-8277-4434], Apollo - University of Cambridge Repository |
| Rok vydání: | 2016 |
| Předmět: |
0301 basic medicine
FOS: Computer and information sciences Computer science General Chemical Engineering cs.LG physics.chem-ph FOS: Physical sciences Domain (mathematical analysis) Data-driven Machine Learning (cs.LG) Set (abstract data type) 03 medical and health sciences Physics - Chemical Physics Representation (mathematics) QD1-999 Chemical Physics (physics.chem-ph) Artificial neural network General Chemistry Construct (python library) Computer Science - Learning Chemistry 030104 developmental biology Algorithm Encoder Decoding methods Research Article |
| Zdroj: | ACS Central Science ACS Central Science, Vol 4, Iss 2, Pp 268-276 (2018) |
| DOI: | 10.48550/arxiv.1610.02415 |
| Popis: | 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. To solve the inverse design challenge in chemistry, we convert molecules into continuous vector representations using neural networks. We demonstrate gradient-based property optimization of molecules. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|