Images of chemical structures as molecular representations for deep learning
Source abstract
Abstract
Abstract Implementing Artificial Intelligence for chemical applications provides a wealth of opportunity for materials discovery, healthcare and smart manufacturing. For such applications to be successful, it is necessary to translate the properties of molecules into a digital format so they can be passed to the algorithms used for smart modelling. The literature has shown a wealth of different strategies for this task, yet there remains a host of limitations. To overcome these challenges, we present two-dimensional images of chemical structures as molecular representations. This methodology was evaluated against other techniques in both classification and regression tasks. Images unlocked (1) superior augmentation strategies, (2) application of specialist network architectures and (3) transfer learning, all contributing to superior performance and without prior specialised knowledge on cheminformatics required. This work takes advantage of image feature maps which do not rely on chemical properties and so can represent multi-component systems without further property calculations. Graphical abstract
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Computational Drug Discovery Methods · Machine Learning in Materials Science · Spectroscopy and Chemometric Analyses
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (41)
Show source-provided references
1. L. David, A. Thakkar, R. Mercado, O. Engkvist, Molecular representations in AI-driven drug discovery: a review and practical guide. J. Cheminform. 12(1), 56 (2020). https://doi.org/10.1186/s13321-020-00460-510.1186/s13321-020-00460-5 ↗
2. S. Nagaprasad, D.L. Padmaja, Y. Qureshi, S.L. Bangare, M. Mishra, B.D. Mazumdar, Investigating the impact of machine learning in pharmaceutical industry. J. Pharm. Res. Int. 33, 6–14 (2021). https://doi.org/10.9734/JPRI/2021/v33i46A3283410.9734/JPRI/2021/v33i46A32834 ↗
3. K.-K. Mak, M.R. Pichika, Artificial intelligence in drug development: present status and future prospects. Drug Discov. Today 24, 773–780 (2019). https://doi.org/10.1016/j.drudis.2018.11.01410.1016/j.drudis.2018.11.014 ↗
4. L.S. Taylor, D.E. Braun, J.W. Steed, Crystals and crystallization in drug delivery design. Cryst. Growth Des. 21(3), 1375–1377 (2021). https://doi.org/10.1021/acs.cgd.0c0159210.1021/acs.cgd.0c01592 ↗
5. C.R. Gardner, C.T. Walsh, Ö. Almarsson, Drugs as materials: valuing physical form in drug discovery. Nat. Rev. Drug Discov. 3(11), 926–934 (2004). https://doi.org/10.1038/nrd155010.1038/nrd1550 ↗
6. J.K. Haleblian, Characterization of habits and crystalline modification of solids and their pharmaceutical applications. J. Pharm. Sci. 64(8), 1269–1288 (1975). https://doi.org/10.1002/jps.260064080510.1002/jps.2600640805 ↗
7. N. Pudasaini, P.P. Upadhyay, C.R. Parker, S.U. Hagen, A.D. Bond, J. Rantanen, Downstream processability of crystal habit-modified active pharmaceutical ingredient. Org. Process Res. Dev. 21(4), 571–577 (2017). https://doi.org/10.1021/acs.oprd.6b0043410.1021/acs.oprd.6b00434 ↗
8. N. Qiao, M. Li, W. Schlindwein, N. Malek, A. Davies, G. Trappitt, Pharmaceutical cocrystals: an overview. Int. J. Pharm. 419(1), 1–11 (2011). https://doi.org/10.1016/j.ijpharm.2011.07.03710.1016/j.ijpharm.2011.07.037 ↗
9. D.J. Good, N. Rodríguez-Hornedo, Solubility advantage of pharmaceutical cocrystals. Cryst. Growth Des. 9(5), 2252–2264 (2009). https://doi.org/10.1021/cg801039j10.1021/cg801039j ↗
10. A. Ghosh, L. Louis, K.K. Arora, B.C. Hancock, J.F. Krzyzaniak, P. Meenan, S. Nakhmanson, G.P.F. Wood, Assessment of machine learning approaches for predicting the crystallization propensity of active pharmaceutical ingredients. CrystEngComm 21(8), 1215–1223 (2019). https://doi.org/10.1039/C8CE01589A10.1039/C8CE01589A ↗
11. J.G.P. Wicker, R.I. Cooper, Will it crystallise? Predicting crystallinity of molecular materials. CrystEngComm 17(9), 1927–1934 (2015). https://doi.org/10.1039/C4CE01912A10.1039/C4CE01912A ↗
12. A. Alhalaweh, A. Alzghoul, W. Kaialy, D. Mahlin, C.A.S. Bergström, Computational predictions of glass-forming ability and crystallization tendency of drug molecules. Mol. Pharm. 11(9), 3123–3132 (2014). https://doi.org/10.1021/mp500303a10.1021/mp500303a ↗
13. J.G.P. Wicker, L.M. Crowley, O. Robshaw, E.J. Little, S.P. Stokes, R.I. Cooper, S.E. Lawrence, Will they co-crystallize? CrystEngComm 19(36), 5336–5340 (2017). https://doi.org/10.1039/C7CE00587C10.1039/C7CE00587C ↗
14. J.R. Gröls, B. Castro-Dominguez, Mechanochemical co-crystallization: insights and predictions. Comput. Chem. Eng. (2021). https://doi.org/10.1016/j.compchemeng.2021.10741610.1016/j.compchemeng.2021.107416 ↗
15. D.S. Palmer, N.M. O’Boyle, R.C. Glen, J.B.O. Mitchell, Random forest models to predict aqueous solubility. J. Chem. Inf. Model. 47(1), 150–158 (2007). https://doi.org/10.1021/ci060164k10.1021/ci060164k ↗
16. R.M. Bhardwaj, A. Johnston, B.F. Johnston, A.J. Florence, A random forest model for predicting the crystallisability of organic molecules. CrystEngComm 17(23), 4272–4275 (2015). https://doi.org/10.1039/C4CE02403F10.1039/C4CE02403F ↗
17. T. Barnard, H. Hagan, S. Tseng, G.C. Sosso, Less may be more: an informed reflection on molecular descriptors for drug design and discovery. Mol. Syst. Des. Eng. 5(1), 317–329 (2020). https://doi.org/10.1039/C9ME00109C10.1039/C9ME00109C ↗
18. H. Moriwaki, Y.-S. Tian, N. Kawashita, T. Takagi, Mordred: a molecular descriptor calculator. J. Cheminform. 10(1), 4 (2018). https://doi.org/10.1186/s13321-018-0258-y10.1186/s13321-018-0258-y ↗
19. R. Gladysz, F.M.D. Santos, W. Langenaeker, G. Thijs, K. Augustyns, H.D. Winter, Spectrophores as one-dimensional descriptors calculated from three-dimensional atomic properties: applications ranging from scaffold hopping to multi-target virtual screening. J. Cheminform. 10, 9 (2018). https://doi.org/10.1186/s13321-018-0268-910.1186/s13321-018-0268-9 ↗
20. D. Rogers, M. Hahn, Extended-connectivity fingerprints. J. Chem. Inf. Model. 50, 742–754 (2010). https://doi.org/10.1021/ci100050t10.1021/ci100050t ↗
21. S. Jaeger, S. Fulle, S. Turk, Mol2vec: unsupervised machine learning approach with chemical intuition. J. Chem. Inf. Model. 58, 27–35 (2018). https://doi.org/10.1021/acs.jcim.7b0061610.1021/acs.jcim.7b00616 ↗
22. A. Bender, How similar are those molecules after all? Use two descriptors and you will have three different answers. Expert Opin. Drug Discov. 5(12), 1141–1151 (2010). https://doi.org/10.1517/17460441.2010.51783210.1517/17460441.2010.517832 ↗
23. D. Weininger, SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules. J. Chem. Inf. Comput. Sci. 28(1), 31–36 (1988)10.1021/ci00057a005 ↗
24. N.M. O’Boyle, Towards a Universal SMILES representation—a standard method to generate canonical SMILES based on the InChI. J. Cheminform. 4(1), 22 (2012). https://doi.org/10.1186/1758-2946-4-2210.1186/1758-2946-4-22 ↗
25. E.J. Bjerrum, SMILES enumeration as data augmentation for neural network modeling of molecules. CoRR (2017). arXiv:1703.07076
26. E. Torabian, M.A. Sobati, New structure-based models for the prediction of flash point of multi-component organic mixtures. Thermochim. Acta 672, 162–172 (2019). https://doi.org/10.1016/j.tca.2018.11.01210.1016/j.tca.2018.11.012 ↗
27. J. Howard, S. Gugger, Deep Learning for Coders with Fastai and PyTorch (O’Reilly Media, Sebastopol, 2020), pp. 36–39
28. O. Russakovsky, J. Deng, H. Su, J. Krause, S. Satheesh, S. Ma, Z. Huang, A. Karpathy, A. Khosla, M. Bernstein, ImageNet large scale visual recognition challenge. Int. J. Comput. Vis. 115(3), 211–252 (2015)10.1007/s11263-015-0816-y ↗
29. T. Shi, Y. Yang, S. Huang, L. Chen, Z. Kuang, Y. Heng, H. Mei, Molecular image-based convolutional neural network for the prediction of ADMET properties. Chemometr. Intell. Lab. Syst. 194, 103853 (2019). https://doi.org/10.1016/J.CHEMOLAB.2019.10385310.1016/J.CHEMOLAB.2019.103853 ↗
30. Y. Matsuzaka, Y. Uesawa, Optimization of a deep-learning method based on the classification of images generated by parameterized deep snap a novel molecular-image-input technique for quantitative structure-activity relationship (QSAR) analysis. Front. Bioeng. Biotechnol. 7, 65 (2019)10.3389/fbioe.2019.00065 ↗
31. C. Shorten, T.M. Khoshgoftaar, A survey on image data augmentation for deep learning. J. Big Data 6(1), 60 (2019). https://doi.org/10.1186/s40537-019-0197-010.1186/s40537-019-0197-0 ↗
32. M. Fernández-Delgado, E. Cernadas, S. Barro, D. Amorim, Do we need hundreds of classifiers to solve real world classification problems? J. Mach. Learn. Res. 15, 3133–3181 (2014)
33. S. Wang, C. Aggarwal, H. Liu, Random-forest-inspired neural networks. ACM Trans. Intell. Syst. Technol. (2018). https://doi.org/10.1145/323223010.1145/3232230 ↗
34. S. Albawi, T.A. Mohammed, S. Al-Zawi, Understanding of a convolutional neural network, in 2017 International Conference on Engineering and Technology (ICET) (2017), pp. 1–6. https://doi.org/10.1109/ICEngTechnol.2017.830818610.1109/ICEngTechnol.2017.8308186 ↗
35. H.O. Brinkhaus, K. Rajan, A. Zielesny, C. Steinbeck, Randepict—random chemical structure depiction generator. ChemRxiv (2022). https://doi.org/10.26434/chemrxiv-2022-t1kbb10.26434/chemrxiv-2022-t1kbb ↗
36. M.C. Sorkun, A. Khetan, S. Er, AqSolDB, a curated reference set of aqueous solubility and 2D descriptors for a diverse set of compounds. Sci. Data 6(1), 143 (2019). https://doi.org/10.1038/s41597-019-0151-110.1038/s41597-019-0151-1 ↗
37. Q. Cui, S. Lu, B. Ni, X. Zeng, Y. Tan, Y.D. Chen, H. Zhao, Improved prediction of aqueous solubility of novel compounds by going deeper with deep learning. Front. Oncol. (2020). https://doi.org/10.3389/fonc.2020.0012110.3389/fonc.2020.00121 ↗
38. S. Kim, Exploring chemical information in PubChem. Curr. Protoc. (2021). https://doi.org/10.1002/cpz1.21710.1002/cpz1.217 ↗
39. J.L. Durant, B.A. Leland, D.R. Henry, J.G. Nourse, Reoptimization of MDL keys for use in drug discovery. J. Chem. Inf. Comput. Sci. 42, 1273–1280 (2002). https://doi.org/10.1021/ci010132r10.1021/ci010132r ↗
40. B. Ramsundar, P. Eastman, P. Walters, V. Pande, K. Leswing, Z. Wu, textitDeep Learning for the Life Sciences (O’Reilly Media, Sebastopol, 2019). https://www.amazon.com/Deep-Learning-Life-Sciences-Microscopy/dp/1492039837
41. K. He, X. Zhang, S. Ren, J. Sun, Deep residual learning for image recognition, in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (2016), pp. 770–77810.1109/CVPR.2016.90 ↗