The ANI-1ccx and ANI-1x data sets, coupled-cluster and density functional theory properties for molecules

Justin S Smith^#^{1

2}, Roman Zubatyuk^#^{2

3}, Benjamin Nebgen², Nicholas Lubbers⁴, Kipton Barros⁴, Adrian E Roitberg⁵, Olexandr Isayev⁶, Sergei Tretiak⁷

Affiliations

¹ Center for Nonlinear Studies, Los Alamos National Laboratory, Los Alamos, NM, USA.
² Theoretical Division, Los Alamos National Laboratory, Los Alamos, NM, USA.
³ Department of Chemistry, Carnegie Mellon University, Pittsburgh, PA, USA.
⁴ Computer, Computational, and Statistical Sciences Division, Los Alamos National Laboratory, Los Alamos, NM, USA.
⁵ University of Florida, Department of Chemistry, PO Box 117200, 32611-7200, Gainesville, USA. roitberg@ufl.edu.
⁶ Department of Chemistry, Carnegie Mellon University, Pittsburgh, PA, USA. olexandr@olexandrisayev.com.
⁷ Theoretical Division, Los Alamos National Laboratory, Los Alamos, NM, USA. serg@lanl.gov.

^# Contributed equally.

PMID: 32358545
PMCID: PMC7195467
DOI: 10.1038/s41597-020-0473-z

The ANI-1ccx and ANI-1x data sets, coupled-cluster and density functional theory properties for molecules

Justin S Smith et al. Sci Data. 2020.

. 2020 May 1;7(1):134.

doi: 10.1038/s41597-020-0473-z.

Authors

Justin S Smith^#^{1

2}, Roman Zubatyuk^#^{2

3}, Benjamin Nebgen², Nicholas Lubbers⁴, Kipton Barros⁴, Adrian E Roitberg⁵, Olexandr Isayev⁶, Sergei Tretiak⁷

Affiliations

¹ Center for Nonlinear Studies, Los Alamos National Laboratory, Los Alamos, NM, USA.
² Theoretical Division, Los Alamos National Laboratory, Los Alamos, NM, USA.
³ Department of Chemistry, Carnegie Mellon University, Pittsburgh, PA, USA.
⁴ Computer, Computational, and Statistical Sciences Division, Los Alamos National Laboratory, Los Alamos, NM, USA.
⁵ University of Florida, Department of Chemistry, PO Box 117200, 32611-7200, Gainesville, USA. roitberg@ufl.edu.
⁶ Department of Chemistry, Carnegie Mellon University, Pittsburgh, PA, USA. olexandr@olexandrisayev.com.
⁷ Theoretical Division, Los Alamos National Laboratory, Los Alamos, NM, USA. serg@lanl.gov.

^# Contributed equally.

PMID: 32358545
PMCID: PMC7195467
DOI: 10.1038/s41597-020-0473-z

Abstract

Maximum diversification of data is a central theme in building generalized and accurate machine learning (ML) models. In chemistry, ML has been used to develop models for predicting molecular properties, for example quantum mechanics (QM) calculated potential energy surfaces and atomic charge models. The ANI-1x and ANI-1ccx ML-based general-purpose potentials for organic molecules were developed through active learning; an automated data diversification process. Here, we describe the ANI-1x and ANI-1ccx data sets. To demonstrate data diversity, we visualize it with a dimensionality reduction scheme, and contrast against existing data sets. The ANI-1x data set contains multiple QM properties from 5 M density functional theory calculations, while the ANI-1ccx data set contains 500 k data points obtained with an accurate CCSD(T)/CBS extrapolation. Approximately 14 million CPU core-hours were expended to generate this data. Multiple QM calculated properties for the chemical elements C, H, N, and O are provided: energies, atomic forces, multipole moments, atomic charges, etc. We provide this data to the community to aid research and development of ML models for chemistry.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

**Fig. 1**
Active learning schemes for building ANI data sets. (a) The active learning algorithm employed during the construction of the ANI-1x data set. (b) The ANI-1ccx selection and data generation scheme.

**Fig. 2**
2D parametric t-SNE embeddings. These embeddings are for the 1st layer of activations of the ANI-1x model for the complete QM9 data set and random subsets of the ANI-1, ANI-1x and ANI-1ccx data sets. The same number of atoms are compared for each element. The different colors correspond to the number and type of bonded neighbors.

**Fig. 3**
ANI data set energy and size distribution. (a) A histogram of the potential energies in the ANI-1x and ANI-1ccx data sets with a linear fit per atomic element E_s removed. The bin width is 1 millihartree. (b) A histogram of the total number of atoms (including C, H, N, and O atoms) per molecule in the ANI-1x and ANI-1ccx data sets. The bin width is one.

See this image and copyright information in PMC

References

1. Gandhi, D., Pinto, L. & Gupta, A. Learning to fly by crashing. In 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 3948–3955 (IEEE, 2017).
1. Settles B. Active learning. Synthesis Lectures on Artificial Intelligence and Machine Learning. 2012;18:1–111. doi: 10.2200/S00429ED1V01Y201207AIM018. - DOI
1. Reker, D. & Schneider, G. Active-learning strategies in computer-assisted drug discovery, vol. 20 (Elsevier Current Trends, 2015). - PubMed
1. Podryabinkin EV, Shapeev AV. Active learning of linearly parametrized interatomic potentials. Computational Materials Science. 2017;140:171–180. doi: 10.1016/j.commatsci.2017.08.031. - DOI
1. Smith JS, Nebgen B, Lubbers N, Isayev O, Roitberg AE. Less is more: sampling chemical space with active learning. The Journal of Chemical Physics. 2018;148:241733. doi: 10.1063/1.5023802. - DOI - PubMed

Save citation to file

Email citation

Add to Collections

Add to My Bibliography

Your saved search

Create a file for external citation management software

Your RSS Feed

The ANI-1ccx and ANI-1x data sets, coupled-cluster and density functional theory properties for molecules

Affiliations

The ANI-1ccx and ANI-1x data sets, coupled-cluster and density functional theory properties for molecules

Authors

Affiliations

Abstract

Conflict of interest statement

Figures

References

Publication types

Grants and funding

LinkOut - more resources

Full Text Sources

Other Literature Sources