دورية أكاديمية
Informing geometric deep learning with electronic interactions to accelerate quantum chemistry.
| العنوان: | Informing geometric deep learning with electronic interactions to accelerate quantum chemistry. |
|---|---|
| المؤلفون: | Qiao Z; Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125., Christensen AS; Entos, Inc., Los Angeles, CA 90027., Welborn M; Entos, Inc., Los Angeles, CA 90027., Manby FR; Entos, Inc., Los Angeles, CA 90027., Anandkumar A; Division of Engineering and Applied Science, California Institute of Technology, Pasadena, CA 91125.; Nvidia Corporation, Santa Clara, CA 95051., Miller TF 3rd; Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA 91125.; Entos, Inc., Los Angeles, CA 90027. |
| المصدر: | Proceedings of the National Academy of Sciences of the United States of America [Proc Natl Acad Sci U S A] 2022 Aug 02; Vol. 119 (31), pp. e2205221119. Date of Electronic Publication: 2022 Jul 28. |
| نوع المنشور: | Journal Article; Research Support, Non-U.S. Gov't |
| اللغة: | English |
| بيانات الدورية: | Publisher: National Academy of Sciences Country of Publication: United States NLM ID: 7505876 Publication Model: Print-Electronic Cited Medium: Internet ISSN: 1091-6490 (Electronic) Linking ISSN: 00278424 NLM ISO Abbreviation: Proc Natl Acad Sci U S A Subsets: MEDLINE |
| أسماء مطبوعة: | Original Publication: Washington, DC : National Academy of Sciences |
| مواضيع طبية MeSH: | Deep Learning*, Electronics ; Machine Learning ; Neural Networks, Computer ; Small Molecule Libraries |
| مستخلص: | Predicting electronic energies, densities, and related chemical properties can facilitate the discovery of novel catalysts, medicines, and battery materials. However, existing machine learning techniques are challenged by the scarcity of training data when exploring unknown chemical spaces. We overcome this barrier by systematically incorporating knowledge of molecular electronic structure into deep learning. By developing a physics-inspired equivariant neural network, we introduce a method to learn molecular representations based on the electronic interactions among atomic orbitals. Our method, OrbNet-Equi, leverages efficient tight-binding simulations and learned mappings to recover high-fidelity physical quantities. OrbNet-Equi accurately models a wide spectrum of target properties while being several orders of magnitude faster than density functional theory. Despite only using training samples collected from readily available small-molecule libraries, OrbNet-Equi outperforms traditional semiempirical and machine learning-based methods on comprehensive downstream benchmarks that encompass diverse main-group chemical processes. Our method also describes interactions in challenging charge-transfer complexes and open-shell systems. We anticipate that the strategy presented here will help to expand opportunities for studies in chemistry and materials science, where the acquisition of experimental or reference training data is costly. |
| References: | Nat Chem. 2020 Oct;12(10):945-951. (PMID: 32929248) J Am Chem Soc. 2022 Feb 2;144(4):1971-1985. (PMID: 35049279) J Chem Theory Comput. 2022 Mar 8;18(3):1607-1618. (PMID: 35175045) J Chem Theory Comput. 2017 May 9;13(5):1989-2009. (PMID: 28418654) J Chem Inf Model. 2022 Feb 28;62(4):785-800. (PMID: 35119861) J Chem Theory Comput. 2020 Jul 14;16(7):4226-4237. (PMID: 32441933) Phys Chem Chem Phys. 2022 May 11;24(18):10775-10783. (PMID: 35470831) Proc Natl Acad Sci U S A. 2022 Aug 2;119(31):e2205221119. (PMID: 35901215) Proc Natl Acad Sci U S A. 2020 Aug 25;117(34):20520-20529. (PMID: 32796105) Science. 2021 Dec 10;374(6573):1385-1389. (PMID: 34882476) Chem Sci. 2019 Sep 9;10(41):9424-9432. (PMID: 32055318) J Chem Theory Comput. 2019 Mar 12;15(3):1652-1671. (PMID: 30741547) J Chem Theory Comput. 2013 Jan 8;9(1):263-72. (PMID: 26589028) J Chem Phys. 2022 Mar 21;156(11):114101. (PMID: 35317562) Nat Rev Drug Discov. 2019 Jun;18(6):463-477. (PMID: 30976107) Proc Natl Acad Sci U S A. 1987 Dec;84(24):8898-902. (PMID: 3122208) J Comput Chem. 2014 Jul 30;35(20):1509-16. (PMID: 24931471) J Chem Theory Comput. 2019 Jun 11;15(6):3678-3693. (PMID: 31042390) Nat Commun. 2020 Jul 14;11(1):3509. (PMID: 32665540) Phys Rev B Condens Matter. 1988 Jan 15;37(2):785-789. (PMID: 9944570) J Chem Phys. 2019 Feb 14;150(6):064105. (PMID: 30769998) Sci Adv. 2017 Dec 13;3(12):e1701816. (PMID: 29242828) Phys Rev Lett. 2021 Jan 22;126(3):036401. (PMID: 33543980) Phys Chem Chem Phys. 2006 May 7;8(17):1985-93. (PMID: 16633685) Nature. 2004 Dec 16;432(7019):862-5. (PMID: 15602552) J Chem Phys. 2020 Jul 14;153(2):024113. (PMID: 32668949) Phys Rev Lett. 2007 Apr 6;98(14):146401. (PMID: 17501293) Science. 2021 Aug 27;373(6558):1047-1051. (PMID: 34446608) Int J High Perform Comput Appl. 2023 Jan;37(1):28-44. (PMID: 36647365) J Comput Chem. 2004 Oct;25(13):1584-604. (PMID: 15264253) J Chem Theory Comput. 2020 Jul 14;16(7):4192-4202. (PMID: 32543858) J Am Chem Soc. 1986 Sep 1;108(19):5708-11. (PMID: 22175316) J Chem Phys. 2021 Feb 14;154(6):064103. (PMID: 33588555) Nat Commun. 2022 May 4;13(1):2453. (PMID: 35508450) Nature. 2020 Sep;585(7824):217-220. (PMID: 32908269) J Chem Inf Model. 2021 Mar 22;61(3):1066-1082. (PMID: 33629839) J Chem Phys. 2017 Oct 28;147(16):161727. (PMID: 29096505) J Chem Phys. 2020 Sep 28;153(12):124111. (PMID: 33003742) J Chem Phys. 2018 Jun 28;148(24):241717. (PMID: 29960351) J Chem Phys. 2016 Nov 7;145(17):170901. (PMID: 27825224) J Chem Phys. 2018 Feb 14;148(6):064104. (PMID: 29448802) J Comput Chem. 2008 Jan 30;29(2):167-75. (PMID: 17568435) Sci Adv. 2017 May 05;3(5):e1603015. (PMID: 28508076) Nat Commun. 2021 Dec 2;12(1):7022. (PMID: 34857738) J Phys Chem Lett. 2016 Jun 16;7(12):2197-203. (PMID: 27203625) J Phys Chem B. 2021 Apr 15;125(14):3598-3612. (PMID: 33798336) J Chem Phys. 2019 Apr 7;150(13):131103. (PMID: 30954042) Nucleic Acids Res. 2017 Jan 4;45(D1):D945-D954. (PMID: 27899562) J Am Chem Soc. 2007 May 2;129(17):5710-8. (PMID: 17411048) J Phys Chem A. 2013 Mar 14;117(10):2269-77. (PMID: 23379303) J Cheminform. 2019 Feb 14;11(1):14. (PMID: 30767086) J Chem Theory Comput. 2018 Sep 11;14(9):4772-4779. (PMID: 30040892) Phys Chem Chem Phys. 2017 Dec 13;19(48):32184-32215. (PMID: 29110012) Phys Chem Chem Phys. 2005 Sep 21;7(18):3297-305. (PMID: 16240044) Sci Data. 2017 Dec 19;4:170193. (PMID: 29257127) J Chem Theory Comput. 2015 Jun 9;11(6):2473-86. (PMID: 26575547) J Phys Chem A. 2020 Sep 3;124(35):7155-7165. (PMID: 32786985) J Cheminform. 2020 Sep 1;12(1):51. (PMID: 33431044) Phys Rev Lett. 2018 Apr 6;120(14):143001. (PMID: 29694129) ACS Cent Sci. 2020 Apr 22;6(4):513-524. (PMID: 32342001) Nat Chem. 2014 Sep;6(9):765-73. (PMID: 25143210) Sci Data. 2014 Aug 05;1:140022. (PMID: 25977779) J Chem Phys. 2020 Jan 31;152(4):044107. (PMID: 32007071) J Chem Theory Comput. 2018 Nov 13;14(11):5764-5776. (PMID: 30351008) Nat Commun. 2019 Jul 1;10(1):2903. (PMID: 31263102) J Chem Theory Comput. 2015 May 12;11(5):2087-96. (PMID: 26574412) |
| فهرسة مساهمة: | Keywords: equivariance; machine learning; quantum chemistry |
| المشرفين على المادة: | 0 (Small Molecule Libraries) |
| تواريخ الأحداث: | Date Created: 20220728 Date Completed: 20220801 Latest Revision: 20240902 |
| رمز التحديث: | 20240902 |
| مُعرف محوري في PubMed: | PMC9351474 |
| DOI: | 10.1073/pnas.2205221119 |
| PMID: | 35901215 |
| قاعدة البيانات: | MEDLINE |
| تدمد: | 1091-6490 |
|---|---|
| DOI: | 10.1073/pnas.2205221119 |