AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection
Abstract Protein-ligand docking is a computational method for identifying drug leads. The method is capable of narrowing a vast library of compounds down to a tractable size for downstream simulation or experimental testing and is widely used in drug discovery. While there has been progress in accel...
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2023-02-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-023-28785-9 |
| _version_ | 1811166013890232320 |
|---|---|
| author | Austin Clyde Xuefeng Liu Thomas Brettin Hyunseung Yoo Alexander Partin Yadu Babuji Ben Blaiszik Jamaludin Mohd-Yusof Andre Merzky Matteo Turilli Shantenu Jha Arvind Ramanathan Rick Stevens |
| author_facet | Austin Clyde Xuefeng Liu Thomas Brettin Hyunseung Yoo Alexander Partin Yadu Babuji Ben Blaiszik Jamaludin Mohd-Yusof Andre Merzky Matteo Turilli Shantenu Jha Arvind Ramanathan Rick Stevens |
| author_sort | Austin Clyde |
| collection | DOAJ |
| description | Abstract Protein-ligand docking is a computational method for identifying drug leads. The method is capable of narrowing a vast library of compounds down to a tractable size for downstream simulation or experimental testing and is widely used in drug discovery. While there has been progress in accelerating scoring of compounds with artificial intelligence, few works have bridged these successes back to the virtual screening community in terms of utility and forward-looking development. We demonstrate the power of high-speed ML models by scoring 1 billion molecules in under a day (50 k predictions per GPU seconds). We showcase a workflow for docking utilizing surrogate AI-based models as a pre-filter to a standard docking workflow. Our workflow is ten times faster at screening a library of compounds than the standard technique, with an error rate less than 0.01% of detecting the underlying best scoring 0.1% of compounds. Our analysis of the speedup explains that another order of magnitude speedup must come from model accuracy rather than computing speed. In order to drive another order of magnitude of acceleration, we share a benchmark dataset consisting of 200 million 3D complex structures and 2D structure scores across a consistent set of 13 million “in-stock” molecules over 15 receptors, or binding sites, across the SARS-CoV-2 proteome. We believe this is strong evidence for the community to begin focusing on improving the accuracy of surrogate models to improve the ability to screen massive compound libraries 100 × or even 1000 × faster than current techniques and reduce missing top hits. The technique outlined aims to be a fast drop-in replacement for docking for screening billion-scale molecular libraries. |
| first_indexed | 2024-04-10T15:45:31Z |
| format | Article |
| id | doaj.art-9350cd1448aa41b5828f39539f4ae94c |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-04-10T15:45:31Z |
| publishDate | 2023-02-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-9350cd1448aa41b5828f39539f4ae94c2023-02-12T12:09:40ZengNature PortfolioScientific Reports2045-23222023-02-0113111410.1038/s41598-023-28785-9AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detectionAustin Clyde0Xuefeng Liu1Thomas Brettin2Hyunseung Yoo3Alexander Partin4Yadu Babuji5Ben Blaiszik6Jamaludin Mohd-Yusof7Andre Merzky8Matteo Turilli9Shantenu Jha10Arvind Ramanathan11Rick Stevens12Argonne National Laboratory, Data Science and Learning DivisionDepartment of Computer Science, University of ChicagoDepartment of Computer Science, University of ChicagoArgonne National Laboratory, Data Science and Learning DivisionArgonne National Laboratory, Data Science and Learning DivisionDepartment of Computer Science, University of ChicagoArgonne National Laboratory, Data Science and Learning DivisionLos Alamos National Laboratory, Computer, Computational, and Statistical SciencesDepartment of Electrical and Computer Engineering, Rutgers UniversityDepartment of Electrical and Computer Engineering, Rutgers UniversityDepartment of Electrical and Computer Engineering, Rutgers UniversityArgonne National Laboratory, Data Science and Learning DivisionDepartment of Computer Science, University of ChicagoAbstract Protein-ligand docking is a computational method for identifying drug leads. The method is capable of narrowing a vast library of compounds down to a tractable size for downstream simulation or experimental testing and is widely used in drug discovery. While there has been progress in accelerating scoring of compounds with artificial intelligence, few works have bridged these successes back to the virtual screening community in terms of utility and forward-looking development. We demonstrate the power of high-speed ML models by scoring 1 billion molecules in under a day (50 k predictions per GPU seconds). We showcase a workflow for docking utilizing surrogate AI-based models as a pre-filter to a standard docking workflow. Our workflow is ten times faster at screening a library of compounds than the standard technique, with an error rate less than 0.01% of detecting the underlying best scoring 0.1% of compounds. Our analysis of the speedup explains that another order of magnitude speedup must come from model accuracy rather than computing speed. In order to drive another order of magnitude of acceleration, we share a benchmark dataset consisting of 200 million 3D complex structures and 2D structure scores across a consistent set of 13 million “in-stock” molecules over 15 receptors, or binding sites, across the SARS-CoV-2 proteome. We believe this is strong evidence for the community to begin focusing on improving the accuracy of surrogate models to improve the ability to screen massive compound libraries 100 × or even 1000 × faster than current techniques and reduce missing top hits. The technique outlined aims to be a fast drop-in replacement for docking for screening billion-scale molecular libraries.https://doi.org/10.1038/s41598-023-28785-9 |
| spellingShingle | Austin Clyde Xuefeng Liu Thomas Brettin Hyunseung Yoo Alexander Partin Yadu Babuji Ben Blaiszik Jamaludin Mohd-Yusof Andre Merzky Matteo Turilli Shantenu Jha Arvind Ramanathan Rick Stevens AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection Scientific Reports |
| title | AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection |
| title_full | AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection |
| title_fullStr | AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection |
| title_full_unstemmed | AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection |
| title_short | AI-accelerated protein-ligand docking for SARS-CoV-2 is 100-fold faster with no significant change in detection |
| title_sort | ai accelerated protein ligand docking for sars cov 2 is 100 fold faster with no significant change in detection |
| url | https://doi.org/10.1038/s41598-023-28785-9 |
| work_keys_str_mv | AT austinclyde aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT xuefengliu aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT thomasbrettin aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT hyunseungyoo aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT alexanderpartin aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT yadubabuji aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT benblaiszik aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT jamaludinmohdyusof aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT andremerzky aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT matteoturilli aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT shantenujha aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT arvindramanathan aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection AT rickstevens aiacceleratedproteinliganddockingforsarscov2is100foldfasterwithnosignificantchangeindetection |