Heterogeneous graph convolutional neural network for protein-ligand scoring
Aim: Drug discovery is a long process, often taking decades of research endeavors. It is still an active area of research in both academic and industrial sectors with efforts on reducing time and cost. Computational simulations like molecular docking enable fast exploration of large databases of com...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Open Exploration
2023-04-01
|
| Series: | Exploration of Drug Science |
| Subjects: | |
| Online Access: | https://www.explorationpub.com/uploads/Article/A100810/100810.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1825199623615021056 |
|---|---|
| author | Kevin Crampon Alexis Giorkallos Xavier Vigouroux Stephanie Baud Luiz Angelo Steffenel |
| author_facet | Kevin Crampon Alexis Giorkallos Xavier Vigouroux Stephanie Baud Luiz Angelo Steffenel |
| author_sort | Kevin Crampon |
| collection | DOAJ |
| description | Aim: Drug discovery is a long process, often taking decades of research endeavors. It is still an active area of research in both academic and industrial sectors with efforts on reducing time and cost. Computational simulations like molecular docking enable fast exploration of large databases of compounds and extract the most promising molecule candidates for further in vitro and in vivo tests. Structure-based molecular docking is a complex process mixing both surface exploration and energy estimation to find the minimal free energy of binding corresponding to the best interaction location. Methods: Hereafter, heterogeneous graph score (HGScore), a new scoring function is proposed and is developed in the context of a protein-small compound-complex. Each complex is represented by a heterogeneous graph allowing to separate edges according to their class (inter- or intra-molecular). Then a heterogeneous graph convolutional network (HGCN) is used allowing the discrimination of the information according to the edge crossed. In the end, the model produces the affinity score of the complex. Results: HGScore has been tested on the comparative assessment of scoring functions (CASF) 2013 and 2016 benchmarks for scoring, ranking, and docking powers. It has achieved good performances by outperforming classical methods and being among the best artificial intelligence (AI) methods. Conclusions: Thus, HGScore brings a new way to represent protein-ligand interactions. Using a representation that involves classical graph neural networks (GNNs) and splitting the learning process regarding the edge type makes the proposed model to be the best adapted for future transfer learning on other (protein-DNA, protein-sugar, protein-protein, etc.) biological complexes. |
| format | Article |
| id | doaj-art-f724812cd019461b84b1b733bd1cfd22 |
| institution | Kabale University |
| issn | 2836-7677 |
| language | English |
| publishDate | 2023-04-01 |
| publisher | Open Exploration |
| record_format | Article |
| series | Exploration of Drug Science |
| spelling | doaj-art-f724812cd019461b84b1b733bd1cfd222025-02-08T04:48:16ZengOpen ExplorationExploration of Drug Science2836-76772023-04-01112613910.37349/eds.2023.00010Heterogeneous graph convolutional neural network for protein-ligand scoringKevin Crampon0https://orcid.org/0000-0001-6124-0719Alexis Giorkallos1Xavier Vigouroux2Stephanie Baud3https://orcid.org/0000-0002-4436-0652Luiz Angelo Steffenel4https://orcid.org/0000-0003-3670-4088Eviden, 38130 Echirolles, France; UMR CNRS/URCA 7369 MEDyC, Université de Reims Champagne-Ardenne, 51687 Reims, France; LICIIS, Université de Reims Champagne-Ardenne, 51687 Reims, FranceEviden, 38130 Echirolles, FranceEviden, 38130 Echirolles, FranceUMR CNRS/URCA 7369 MEDyC, Université de Reims Champagne-Ardenne, 51687 Reims, FranceLICIIS, Université de Reims Champagne-Ardenne, 51687 Reims, FranceAim: Drug discovery is a long process, often taking decades of research endeavors. It is still an active area of research in both academic and industrial sectors with efforts on reducing time and cost. Computational simulations like molecular docking enable fast exploration of large databases of compounds and extract the most promising molecule candidates for further in vitro and in vivo tests. Structure-based molecular docking is a complex process mixing both surface exploration and energy estimation to find the minimal free energy of binding corresponding to the best interaction location. Methods: Hereafter, heterogeneous graph score (HGScore), a new scoring function is proposed and is developed in the context of a protein-small compound-complex. Each complex is represented by a heterogeneous graph allowing to separate edges according to their class (inter- or intra-molecular). Then a heterogeneous graph convolutional network (HGCN) is used allowing the discrimination of the information according to the edge crossed. In the end, the model produces the affinity score of the complex. Results: HGScore has been tested on the comparative assessment of scoring functions (CASF) 2013 and 2016 benchmarks for scoring, ranking, and docking powers. It has achieved good performances by outperforming classical methods and being among the best artificial intelligence (AI) methods. Conclusions: Thus, HGScore brings a new way to represent protein-ligand interactions. Using a representation that involves classical graph neural networks (GNNs) and splitting the learning process regarding the edge type makes the proposed model to be the best adapted for future transfer learning on other (protein-DNA, protein-sugar, protein-protein, etc.) biological complexes.https://www.explorationpub.com/uploads/Article/A100810/100810.pdfmolecular dockingligand-proteinscoring functiondeep learninggraph neural network |
| spellingShingle | Kevin Crampon Alexis Giorkallos Xavier Vigouroux Stephanie Baud Luiz Angelo Steffenel Heterogeneous graph convolutional neural network for protein-ligand scoring Exploration of Drug Science molecular docking ligand-protein scoring function deep learning graph neural network |
| title | Heterogeneous graph convolutional neural network for protein-ligand scoring |
| title_full | Heterogeneous graph convolutional neural network for protein-ligand scoring |
| title_fullStr | Heterogeneous graph convolutional neural network for protein-ligand scoring |
| title_full_unstemmed | Heterogeneous graph convolutional neural network for protein-ligand scoring |
| title_short | Heterogeneous graph convolutional neural network for protein-ligand scoring |
| title_sort | heterogeneous graph convolutional neural network for protein ligand scoring |
| topic | molecular docking ligand-protein scoring function deep learning graph neural network |
| url | https://www.explorationpub.com/uploads/Article/A100810/100810.pdf |
| work_keys_str_mv | AT kevincrampon heterogeneousgraphconvolutionalneuralnetworkforproteinligandscoring AT alexisgiorkallos heterogeneousgraphconvolutionalneuralnetworkforproteinligandscoring AT xaviervigouroux heterogeneousgraphconvolutionalneuralnetworkforproteinligandscoring AT stephaniebaud heterogeneousgraphconvolutionalneuralnetworkforproteinligandscoring AT luizangelosteffenel heterogeneousgraphconvolutionalneuralnetworkforproteinligandscoring |