{ "access": { "embargo": { "active": false, "reason": null }, "files": "public", "record": "public", "status": "open" }, "created": "2019-11-12T21:11:42.764482+00:00", "custom_fields": {}, "deletion_status": { "is_deleted": false, "status": "P" }, "files": { "count": 2, "enabled": true, "entries": { "OFFwebinar-YuanqingWang-2019.pdf": { "checksum": "md5:32c12160c406a5cc6d8c16ad86b6304f", "ext": "pdf", "id": "f75232af-52cd-4a90-84e6-59c6b8fb7b8c", "key": "OFFwebinar-YuanqingWang-2019.pdf", "metadata": null, "mimetype": "application/pdf", "size": 17349317 }, "YuanqingWang_GraphNets.mp4": { "checksum": "md5:3b08fd630a07af3c8ad38b0b12fce9a1", "ext": "mp4", "id": "9b97b834-d46f-45a5-90ac-490596de6aee", "key": "YuanqingWang_GraphNets.mp4", "metadata": null, "mimetype": "video/mp4", "size": 59859318 } }, "order": [], "total_bytes": 77208635 }, "id": "3539635", "is_draft": false, "is_published": true, "links": { "access": "https://zenodo.org/api/records/3539635/access", "access_links": "https://zenodo.org/api/records/3539635/access/links", "access_request": "https://zenodo.org/api/records/3539635/access/request", "access_users": "https://zenodo.org/api/records/3539635/access/users", "archive": "https://zenodo.org/api/records/3539635/files-archive", "archive_media": "https://zenodo.org/api/records/3539635/media-files-archive", "communities": "https://zenodo.org/api/records/3539635/communities", "communities-suggestions": "https://zenodo.org/api/records/3539635/communities-suggestions", "doi": "https://doi.org/10.5281/zenodo.3539635", "draft": "https://zenodo.org/api/records/3539635/draft", "files": "https://zenodo.org/api/records/3539635/files", "latest": "https://zenodo.org/api/records/3539635/versions/latest", "latest_html": "https://zenodo.org/records/3539635/latest", "media_files": "https://zenodo.org/api/records/3539635/media-files", "parent": "https://zenodo.org/api/records/3539634", "parent_doi": "https://zenodo.org/doi/10.5281/zenodo.3539634", "parent_html": "https://zenodo.org/records/3539634", "requests": "https://zenodo.org/api/records/3539635/requests", "reserve_doi": "https://zenodo.org/api/records/3539635/draft/pids/doi", "self": "https://zenodo.org/api/records/3539635", "self_doi": "https://zenodo.org/doi/10.5281/zenodo.3539635", "self_html": "https://zenodo.org/records/3539635", "self_iiif_manifest": "https://zenodo.org/api/iiif/record:3539635/manifest", "self_iiif_sequence": "https://zenodo.org/api/iiif/record:3539635/sequence/default", "versions": "https://zenodo.org/api/records/3539635/versions" }, "media_files": { "count": 0, "enabled": false, "entries": {}, "order": [], "total_bytes": 0 }, "metadata": { "creators": [ { "affiliations": [ { "name": "Memorial Sloan Kettering Cancer Center / Weill Cornell Medicine" } ], "person_or_org": { "family_name": "Wang", "given_name": "Yuanqing", "identifiers": [ { "identifier": "0000-0003-4403-2015", "scheme": "orcid" } ], "name": "Wang, Yuanqing", "type": "personal" } } ], "description": "
Yuanqing Wang (MSKCC) gave a talk about using Graph Nets for fast prediction of atomic partial charges as a part of OFF webinar series. The preprint is available here.
\n\n
\nAbstract: Here we show that Graph Nets — a set of update and aggregate functions that operate on molecular topologies and propagate information thereon — are capable of predicting properties of atoms and molecules which would otherwise require expensive Quantum Mechanics (QM) calculations. This architecture is applied to predict atomic partial charges, which are crucial parameters for Molecular Dynamics (MD) simulations, molecular mechanics calculations, and virtual screening, as they determine the electrostatic contributions to interaction energies. We will also briefly introduce our work-in-progress on using this model for inter-hierarchical multitask learning and on Hypergraph Functional Potential (HGFP) — a fast and flexible machine learning force field.
Any materials and data produced as a part of Open Force Field Initiative efforts or application studies using Open Force Field toolkits and force fields.
\r\n", "page": "This community contains materials created within Open Force Field Initiative or associated with it. Open Force Field Initiative is an academic-industry partnership to develop a new generation of molecular mechanics force fields using open software infrastructure and open datasets, with a focus on general small molecule force fields. More information about Open Force Field efforts and the available code and data can be found at our website or GitHub repository.
\r\n\r\nMolecular modeling is widely used in diverse pharmaceutical discovery applications, but its utility and predictive power is limited by the accuracy of the underlying molecular mechanics force field used to compute the energetics of biomolecular systems. The Open Force Field Consortium (OpenFF) is an open industry-funded pre-competitive effort to build iteratively more accurate force fields to improve predictive design, along with the necessary infrastructure to make these force fields easier to build and use.
\r\n\r\n\r\n\r\n
The Open Force Field Initiative aims to:
\r\n\r\n\r\n\r\n
Engineer a modern, open, sustainable, extensible, and well-supported framework for automated force field improvement and application;
\r\n\tApply the above software infrastructure to rapidly develop iteratively improved versions of small molecule force fields (compatible with AMBER biopolymer force fields, such as AMBER ff14SB) that uses standard SMARTS chemical perception language pattern matching to assign parameters;
\r\n\tProduce new comprehensive force fields that achieve dramatically improved accuracy while maintaining compatibility with existing simulation software, providing improved predictive power for diverse applications ranging from predictions of binding affinity, selectivity, and drug resistance, to partitioning, solubility, kinetics, and other properties;
\r\n\tWork closely with industry partners to ensure the development path follows that most relevant to R&D needs;
\r\n\tOpen source, open data, open science: All software, code, data, and force fields will be open and freely available under Open Source Initiative and Creative Commons approved licenses, providing a foundation for further science beyond the scope and timescale of the formal initiative. Industry partners will be enabled to extend the force fields by adding proprietary data using open tools developed as part of this initiative, or build their own workflows based on these tools providing a pathway to sustainability.
\r\n\t
More details about the project plan can be found here.
", "title": "Open Force Field Initiative" }, "revision_id": 0, "slug": "openforcefield", "updated": "2020-04-28T02:04:06.470451+00:00" } ], "ids": [ "41490a69-af82-4a19-9584-8972f3a055e1" ] }, "id": "3539634", "pids": { "doi": { "client": "datacite", "identifier": "10.5281/zenodo.3539634", "provider": "datacite" } } }, "pids": { "doi": { "client": "datacite", "identifier": "10.5281/zenodo.3539635", "provider": "datacite" }, "oai": { "identifier": "oai:zenodo.org:3539635", "provider": "oai" } }, "revision_id": 2, "stats": { "all_versions": { "data_volume": 1243971075.0, "downloads": 57, "unique_downloads": 56, "unique_views": 84, "views": 86 }, "this_version": { "data_volume": 1243971075.0, "downloads": 57, "unique_downloads": 56, "unique_views": 83, "views": 85 } }, "status": "published", "updated": "2020-01-20T16:50:44.020023+00:00", "versions": { "index": 1, "is_latest": true } }