SPICE, A Dataset of Drug-like Molecules and Peptides for Training Machine Learning Potentials

Simple item page
dc.contributor.authorEastman, Peteren
dc.contributor.authorBehara, Pavan Kumaren
dc.contributor.authorDotson, David L.en
dc.contributor.authorGalvelis, Raimondasen
dc.contributor.authorHerr, John E.en
dc.contributor.authorHorton, Josh T.en
dc.contributor.authorMao, Yuezhien
dc.contributor.authorChodera, John D.en
dc.contributor.authorPritchard, Benjamin P.en
dc.contributor.authorWang, Yuanqingen
dc.contributor.authorDe Fabritiis, Giannien
dc.contributor.authorMarkland, Thomas E.en
dc.date.accessioned2023-04-04T15:06:25Zen
dc.date.available2023-04-04T15:06:25Zen
dc.date.issued2023-01-04en
dc.description.abstractMachine learning potentials are an important tool for molecular simulation, but their development is held back by a shortage of high quality datasets to train them on. We describe the SPICE dataset, a new quantum chemistry dataset for training potentials relevant to simulating drug-like small molecules interacting with proteins. It contains over 1.1 million conformations for a diverse set of small molecules, dimers, dipeptides, and solvated amino acids. It includes 15 elements, charged and uncharged molecules, and a wide range of covalent and non-covalent interactions. It provides both forces and energies calculated at the omega B97M-D3(BJ)/def2-TZVPPD level of theory, along with other useful quantities such as multipole moments and bond orders. We train a set of machine learning potentials on it and demonstrate that they can achieve chemical accuracy across a broad region of chemical space. It can serve as a valuable resource for the creation of transferable, ready to use potential functions for use in molecular simulations.en
dc.description.notesResearch reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under award number R01GM140090 (JDC, TEM, PE, GdF) and R01GM132386 (JDC, PKB, YW). BPP acknowledges support from the National Science Foundation under award number CHE-2136142.en
dc.description.sponsorshipNational Institute of General Medical Sciences of the National Institutes of Health [R01GM140090, R01GM132386]; National Science Foundation [CHE-2136142]en
dc.description.versionPublished versionen
dc.format.mimetypeapplication/pdfen
dc.identifier.doihttps://doi.org/10.1038/s41597-022-01882-6en
dc.identifier.eissn2052-4463en
dc.identifier.issue1en
dc.identifier.other11en
dc.identifier.pmid36599873en
dc.identifier.urihttp://hdl.handle.net/10919/114246en
dc.identifier.volume10en
dc.language.isoenen
dc.publisherNature Portfolioen
dc.rightsCreative Commons Attribution 4.0 Internationalen
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/en
dc.subjectProtein-ligand bindingen
dc.subjectaccuracyen
dc.subjectdatabaseen
dc.titleSPICE, A Dataset of Drug-like Molecules and Peptides for Training Machine Learning Potentialsen
dc.title.serialScientific Dataen
dc.typeArticle - Refereeden
dc.type.dcmitypeTexten

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
s41597-022-01882-6.pdf
Size:
1.14 MB
Format:
Adobe Portable Document Format
Description:
Published version
Download

Collections

Scholarly Works, Chemistry