Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials
State of the art computer-aided synthesis planning models are naturally biased toward commonly reported chemical reactions, thus reducing the usefulness of those models for the unusual chemistry relevant to shock physics. To address this problem, a neural network was trained to recognize reaction te...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
AIP Publishing
2021
|
| Online Access: | https://hdl.handle.net/1721.1/137158 |
| _version_ | 1826215223280795648 |
|---|---|
| author | Fortunato, ME Coley, CW Barnes, BC Jensen, KF |
| author_facet | Fortunato, ME Coley, CW Barnes, BC Jensen, KF |
| author_sort | Fortunato, ME |
| collection | MIT |
| description | State of the art computer-aided synthesis planning models are naturally biased toward commonly reported chemical reactions, thus reducing the usefulness of those models for the unusual chemistry relevant to shock physics. To address this problem, a neural network was trained to recognize reaction template applicability for small organic molecules to supplement the rare reaction examples of relevance to energetic materials. The training data for the neural network was generated by brute force determination of template subgraph matching for product molecules from a database of reactions in U.S. patent literature. This data generation strategy successfully augmented the information about template applicability for rare reaction mechanisms in the reaction database. The increased ability to recognize rare reaction templates was demonstrated for reaction templates of interest for energetic material synthesis such as heterocycle ring formation. |
| first_indexed | 2024-09-23T16:19:23Z |
| format | Article |
| id | mit-1721.1/137158 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T16:19:23Z |
| publishDate | 2021 |
| publisher | AIP Publishing |
| record_format | dspace |
| spelling | mit-1721.1/1371582021-11-03T03:02:21Z Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials Fortunato, ME Coley, CW Barnes, BC Jensen, KF State of the art computer-aided synthesis planning models are naturally biased toward commonly reported chemical reactions, thus reducing the usefulness of those models for the unusual chemistry relevant to shock physics. To address this problem, a neural network was trained to recognize reaction template applicability for small organic molecules to supplement the rare reaction examples of relevance to energetic materials. The training data for the neural network was generated by brute force determination of template subgraph matching for product molecules from a database of reactions in U.S. patent literature. This data generation strategy successfully augmented the information about template applicability for rare reaction mechanisms in the reaction database. The increased ability to recognize rare reaction templates was demonstrated for reaction templates of interest for energetic material synthesis such as heterocycle ring formation. 2021-11-02T18:20:03Z 2021-11-02T18:20:03Z 2020 2021-06-09T16:36:39Z Article http://purl.org/eprint/type/ConferencePaper https://hdl.handle.net/1721.1/137158 Fortunato, ME, Coley, CW, Barnes, BC and Jensen, KF. 2020. "Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials." AIP Conference Proceedings, 2272. en 10.1063/12.0000850 AIP Conference Proceedings Creative Commons Attribution-Noncommercial-Share Alike http://creativecommons.org/licenses/by-nc-sa/4.0/ application/pdf AIP Publishing Other repository |
| spellingShingle | Fortunato, ME Coley, CW Barnes, BC Jensen, KF Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials |
| title | Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials |
| title_full | Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials |
| title_fullStr | Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials |
| title_full_unstemmed | Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials |
| title_short | Machine learned prediction of reaction template applicability for data-driven retrosynthetic predictions of energetic materials |
| title_sort | machine learned prediction of reaction template applicability for data driven retrosynthetic predictions of energetic materials |
| url | https://hdl.handle.net/1721.1/137158 |
| work_keys_str_mv | AT fortunatome machinelearnedpredictionofreactiontemplateapplicabilityfordatadrivenretrosyntheticpredictionsofenergeticmaterials AT coleycw machinelearnedpredictionofreactiontemplateapplicabilityfordatadrivenretrosyntheticpredictionsofenergeticmaterials AT barnesbc machinelearnedpredictionofreactiontemplateapplicabilityfordatadrivenretrosyntheticpredictionsofenergeticmaterials AT jensenkf machinelearnedpredictionofreactiontemplateapplicabilityfordatadrivenretrosyntheticpredictionsofenergeticmaterials |