Journal of Chemical Information and Modeling, volume 60, issue 1, pages 22-28

Graph Convolutional Neural Networks as "general-Purpose" Property Predictors: The Universality and Limits of Applicability

Korolev Vadim 1, 2
Mitrofanov Artem 1, 2
Korotcov Alexandru 1
Tkachenko Valery 1
Publication typeJournal Article
Publication date2019-12-20
Quartile SCImago
Q1
Quartile WOS
Q1
Impact factor5.6
ISSN15499596, 1549960X
General Chemistry
Computer Science Applications
General Chemical Engineering
Library and Information Sciences
Abstract
Nowadays the development of new functional materials/chemical compounds using machine learning (ML) techniques is a hot topic and includes several crucial steps, one of which is the choice of chemical structure representation. Classical approach of rigorous feature engineering in ML typically improves the performance of the predictive model, but at the same time, it narrows down the scope of applicability and decreases the physical interpretability of predicted results. In this study, we present graph convolutional neural networks (GCNN) as an architecture that allows to successfully predict the properties of compounds from diverse domains of chemical space, using a minimal set of meaningful descriptors. The applicability of GCNN models has been demonstrated by a wide range of chemical domain-specific properties. Their performance is comparable to state-of-the-art techniques; however, this architecture exempts from the need to carry out precise feature engineering.

Citations by journals

1
2
3
4
5
6
Journal of Chemical Information and Modeling
Journal of Chemical Information and Modeling, 6, 10.34%
Journal of Chemical Information and Modeling
6 publications, 10.34%
Briefings in Bioinformatics
Briefings in Bioinformatics, 6, 10.34%
Briefings in Bioinformatics
6 publications, 10.34%
International Journal of Molecular Sciences
International Journal of Molecular Sciences, 2, 3.45%
International Journal of Molecular Sciences
2 publications, 3.45%
ACS Omega
ACS Omega, 2, 3.45%
ACS Omega
2 publications, 3.45%
Journal of Cheminformatics
Journal of Cheminformatics, 2, 3.45%
Journal of Cheminformatics
2 publications, 3.45%
Angewandte Chemie
Angewandte Chemie, 2, 3.45%
Angewandte Chemie
2 publications, 3.45%
Angewandte Chemie - International Edition
Angewandte Chemie - International Edition, 2, 3.45%
Angewandte Chemie - International Edition
2 publications, 3.45%
Environmental Science & Technology
Environmental Science & Technology, 2, 3.45%
Environmental Science & Technology
2 publications, 3.45%
Chemical Science
Chemical Science, 2, 3.45%
Chemical Science
2 publications, 3.45%
Materials Horizons
Materials Horizons, 1, 1.72%
Materials Horizons
1 publication, 1.72%
Journal of Physical Chemistry Letters
Journal of Physical Chemistry Letters, 1, 1.72%
Journal of Physical Chemistry Letters
1 publication, 1.72%
Advanced Intelligent Systems
Advanced Intelligent Systems, 1, 1.72%
Advanced Intelligent Systems
1 publication, 1.72%
Frontiers in Chemistry
Frontiers in Chemistry, 1, 1.72%
Frontiers in Chemistry
1 publication, 1.72%
Nature Reviews Chemistry
Nature Reviews Chemistry, 1, 1.72%
Nature Reviews Chemistry
1 publication, 1.72%
Science China Materials
Science China Materials, 1, 1.72%
Science China Materials
1 publication, 1.72%
International Journal of Minerals, Metallurgy and Materials, 1, 1.72%
International Journal of Minerals, Metallurgy and Materials
1 publication, 1.72%
Microporous and Mesoporous Materials
Microporous and Mesoporous Materials, 1, 1.72%
Microporous and Mesoporous Materials
1 publication, 1.72%
Drug Discovery Today
Drug Discovery Today, 1, 1.72%
Drug Discovery Today
1 publication, 1.72%
Materials and Design
Materials and Design, 1, 1.72%
Materials and Design
1 publication, 1.72%
Analytica Chimica Acta
Analytica Chimica Acta, 1, 1.72%
Analytica Chimica Acta
1 publication, 1.72%
Mendeleev Communications
Mendeleev Communications, 1, 1.72%
Mendeleev Communications
1 publication, 1.72%
Langmuir
Langmuir, 1, 1.72%
Langmuir
1 publication, 1.72%
Molecular Informatics
Molecular Informatics, 1, 1.72%
Molecular Informatics
1 publication, 1.72%
AICHE Journal
AICHE Journal, 1, 1.72%
AICHE Journal
1 publication, 1.72%
ACS Central Science
ACS Central Science, 1, 1.72%
ACS Central Science
1 publication, 1.72%
Chemical Reviews
Chemical Reviews, 1, 1.72%
Chemical Reviews
1 publication, 1.72%
Journal of Medicinal Chemistry
Journal of Medicinal Chemistry, 1, 1.72%
Journal of Medicinal Chemistry
1 publication, 1.72%
ACS Polymers Au
ACS Polymers Au, 1, 1.72%
ACS Polymers Au
1 publication, 1.72%
Methods in Molecular Biology
Methods in Molecular Biology, 1, 1.72%
Methods in Molecular Biology
1 publication, 1.72%
Wireless Communications and Mobile Computing
Wireless Communications and Mobile Computing, 1, 1.72%
Wireless Communications and Mobile Computing
1 publication, 1.72%
1
2
3
4
5
6

Citations by publishers

5
10
15
20
25
American Chemical Society (ACS)
American Chemical Society (ACS), 21, 36.21%
American Chemical Society (ACS)
21 publications, 36.21%
Wiley
Wiley, 8, 13.79%
Wiley
8 publications, 13.79%
Elsevier
Elsevier, 8, 13.79%
Elsevier
8 publications, 13.79%
Oxford University Press
Oxford University Press, 6, 10.34%
Oxford University Press
6 publications, 10.34%
Springer Nature
Springer Nature, 5, 8.62%
Springer Nature
5 publications, 8.62%
Royal Society of Chemistry (RSC)
Royal Society of Chemistry (RSC), 3, 5.17%
Royal Society of Chemistry (RSC)
3 publications, 5.17%
Multidisciplinary Digital Publishing Institute (MDPI)
Multidisciplinary Digital Publishing Institute (MDPI), 2, 3.45%
Multidisciplinary Digital Publishing Institute (MDPI)
2 publications, 3.45%
Frontiers Media S.A.
Frontiers Media S.A., 1, 1.72%
Frontiers Media S.A.
1 publication, 1.72%
University of Science and Technology Beijing, 1, 1.72%
University of Science and Technology Beijing
1 publication, 1.72%
Hindawi Limited
Hindawi Limited, 1, 1.72%
Hindawi Limited
1 publication, 1.72%
Annual Reviews
Annual Reviews, 1, 1.72%
Annual Reviews
1 publication, 1.72%
5
10
15
20
25
  • We do not take into account publications that without a DOI.
  • Statistics recalculated only for publications connected to researchers, organizations and labs registered on the platform.
  • Statistics recalculated weekly.
Metrics
Share
Cite this
GOST |
Cite this
GOST Copy
Korolev V. et al. Graph Convolutional Neural Networks as "general-Purpose" Property Predictors: The Universality and Limits of Applicability // Journal of Chemical Information and Modeling. 2019. Vol. 60. No. 1. pp. 22-28.
GOST all authors (up to 50) Copy
Korolev V., Mitrofanov A., Korotcov A., Tkachenko V. Graph Convolutional Neural Networks as "general-Purpose" Property Predictors: The Universality and Limits of Applicability // Journal of Chemical Information and Modeling. 2019. Vol. 60. No. 1. pp. 22-28.
RIS |
Cite this
RIS Copy
TY - JOUR
DO - 10.1021/acs.jcim.9b00587
UR - https://doi.org/10.1021%2Facs.jcim.9b00587
TI - Graph Convolutional Neural Networks as "general-Purpose" Property Predictors: The Universality and Limits of Applicability
T2 - Journal of Chemical Information and Modeling
AU - Korotcov, Alexandru
AU - Tkachenko, Valery
AU - Korolev, Vadim
AU - Mitrofanov, Artem
PY - 2019
DA - 2019/12/20 00:00:00
PB - American Chemical Society (ACS)
SP - 22-28
IS - 1
VL - 60
SN - 1549-9596
SN - 1549-960X
ER -
BibTex |
Cite this
BibTex Copy
@article{2019_Korolev,
author = {Alexandru Korotcov and Valery Tkachenko and Vadim Korolev and Artem Mitrofanov},
title = {Graph Convolutional Neural Networks as "general-Purpose" Property Predictors: The Universality and Limits of Applicability},
journal = {Journal of Chemical Information and Modeling},
year = {2019},
volume = {60},
publisher = {American Chemical Society (ACS)},
month = {dec},
url = {https://doi.org/10.1021%2Facs.jcim.9b00587},
number = {1},
pages = {22--28},
doi = {10.1021/acs.jcim.9b00587}
}
MLA
Cite this
MLA Copy
Korolev, Vadim, et al. “Graph Convolutional Neural Networks as "general-Purpose" Property Predictors: The Universality and Limits of Applicability.” Journal of Chemical Information and Modeling, vol. 60, no. 1, Dec. 2019, pp. 22-28. https://doi.org/10.1021%2Facs.jcim.9b00587.
Found error?