On failure modes in molecule generation and optimization.

There has been a wave of generative models for molecules triggered by advances in the field of Deep Learning. These generative models are often used to optimize chemical compounds towards particular properties or a desired biological activity. The evaluation of generative models remains challenging and suggested performance metrics or scoring functions often do not cover all relevant aspects of drug design projects. In this work, we highlight some unintended failure modes in molecular generation and optimization and how these evade detection by current performance metrics.

[1]  Peter Ertl,et al.  Estimation of synthetic accessibility score of drug-like molecules based on molecular complexity and fragment contributions , 2009, J. Cheminformatics.

[2]  George Papadatos,et al.  The ChEMBL bioactivity database: an update , 2013, Nucleic Acids Res..

[3]  Xiaoqing Lyu,et al.  Molecular Graph Generation with Deep Reinforced Multitask Network and Adversarial Imitation Learning , 2019, 2019 IEEE International Conference on Bioinformatics and Biomedicine (BIBM).

[4]  Gerard J. P. van Westen,et al.  An exploration strategy improves the diversity of de novo ligands using deep reinforcement learning: a case for the adenosine A2A receptor , 2018, Journal of Cheminformatics.

[5]  Pascal Friederich,et al.  Self-referencing embedded strings (SELFIES): A 100% robust molecular string representation , 2019, Mach. Learn. Sci. Technol..

[6]  Sergey Nikolenko,et al.  druGAN: An Advanced Generative Adversarial Autoencoder Model for de Novo Generation of New Molecules with Desired Molecular Properties in Silico. , 2017, Molecular pharmaceutics.

[7]  Hugo Ceulemans,et al.  Large-scale comparison of machine learning methods for drug target prediction on ChEMBL , 2018, Chemical science.

[8]  Razvan Pascanu,et al.  Learning Deep Generative Models of Graphs , 2018, ICLR 2018.

[9]  Matthias Rarey,et al.  Machine Learning in Drug Discovery , 2018, J. Chem. Inf. Model..

[10]  Ah Chung Tsoi,et al.  The Graph Neural Network Model , 2009, IEEE Transactions on Neural Networks.

[11]  Frank Noé,et al.  Learning Continuous and Data-Driven Molecular Descriptors by Translating Equivalent Chemical Representations , 2018 .

[12]  Alán Aspuru-Guzik,et al.  Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules , 2016, ACS central science.

[13]  SchmidhuberJürgen.  Deep learning in neural networks , 2015 .

[14]  Gordon M. Crippen,et al.  Prediction of Physicochemical Parameters by Atomic Contributions , 1999, J. Chem. Inf. Comput. Sci..

[15]  Yoshua Bengio,et al.  On the Properties of Neural Machine Translation: Encoder–Decoder Approaches , 2014, SSST@EMNLP.

[16]  Gisbert Schneider,et al.  Tuning artificial intelligence on the de novo design of natural-product-inspired retinoid X receptor modulators , 2018, Communications Chemistry.

[17]  Risto Miikkulainen,et al.  The Surprising Creativity of Digital Evolution: A Collection of Anecdotes from the Evolutionary Computation and Artificial Life Research Communities , 2018, Artificial Life.

[18]  Alán Aspuru-Guzik,et al.  Inverse molecular design using machine learning: Generative models for matter engineering , 2018, Science.

[19]  Alán Aspuru-Guzik,et al.  Optimizing distributions over molecular space. An Objective-Reinforced Generative Adversarial Network for Inverse-design Chemistry (ORGANIC) , 2017 .

[20]  Jan H Jensen.  A graph-based genetic algorithm and generative model/Monte Carlo tree search for the exploration of chemical space† †Electronic supplementary information (ESI) available: The codes used in this study can be found on GitHub: github.com/jensengroup/GB-GA/tree/v0.0 and github.com/jensengroup/GB-GM/tree , 2019, Chemical science.

[21]  David Rogers,et al.  Extended-Connectivity Fingerprints , 2010, J. Chem. Inf. Model..

[22]  Gaël Varoquaux,et al.  Scikit-learn: Machine Learning in Python , 2011, J. Mach. Learn. Res..

[23]  Jure Leskovec,et al.  Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation , 2018, NeurIPS.

[24]  Joan Bruna,et al.  Intriguing properties of neural networks , 2013, ICLR.

[25]  Dominique Douguet,et al.  A genetic algorithm for the automated generation of small organic molecules: Drug design using an evolutionary algorithm , 2000, J. Comput. Aided Mol. Des..

[26]  Jason Yosinski,et al.  Deep neural networks are easily fooled: High confidence predictions for unrecognizable images , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[27]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[28]  Sepp Hochreiter,et al.  Fréchet ChemNet Distance: A Metric for Generative Models for Molecules in Drug Discovery , 2018, J. Chem. Inf. Model..

[29]  Regina Barzilay,et al.  Junction Tree Variational Autoencoder for Molecular Graph Generation , 2018, ICML.

[30]  Gisbert Schneider,et al.  De Novo Design of Bioactive Small Molecules by Artificial Intelligence , 2018, Molecular informatics.

[31]  Li Li,et al.  Optimization of Molecules via Deep Reinforcement Learning , 2018, Scientific Reports.

[32]  Alán Aspuru-Guzik,et al.  Deep learning enables rapid identification of potent DDR1 kinase inhibitors , 2019, Nature Biotechnology.

[33]  Alán Aspuru-Guzik,et al.  Objective-Reinforced Generative Adversarial Networks (ORGAN) for Sequence Generation Models , 2017, ArXiv.

[34]  David Weininger,et al.  SMILES, a chemical language and information system. 1. Introduction to methodology and encoding rules , 1988, J. Chem. Inf. Comput. Sci..

[35]  Noel M. O'Boyle,et al.  DeepSMILES: An Adaptation of SMILES for Use in Machine-Learning of Chemical Structures , 2018 .

[36]  Alán Aspuru-Guzik,et al.  Molecular Sets (MOSES): A Benchmarking Platform for Molecular Generation Models , 2018, Frontiers in Pharmacology.

[37]  G. V. Paolini,et al.  Quantifying the chemical beauty of drugs. , 2012, Nature chemistry.

[38]  Olexandr Isayev,et al.  Deep reinforcement learning for de novo drug design , 2017, Science Advances.

[39]  Thomas Blaschke,et al.  Molecular de-novo design through deep reinforcement learning , 2017, Journal of Cheminformatics.

[40]  Marwin H. S. Segler,et al.  GuacaMol: Benchmarking Models for De Novo Molecular Design , 2018, J. Chem. Inf. Model..

[41]  Jürgen Schmidhuber,et al.  Deep learning in neural networks: An overview , 2014, Neural Networks.

[42]  Gisbert Schneider.  De novo Molecular Design , 2013 .

[43]  Andrey Kazennov,et al.  The cornucopia of meaningful leads: Applying deep adversarial autoencoders for new molecule development in oncology , 2016, Oncotarget.

[44]  Frank Noé,et al.  Efficient multi-objective molecular optimization in a continuous latent space† †Electronic supplementary information (ESI) available: Details of the desirability scaling functions, high resolution figures and detailed results of the GuacaMol benchmark. See DOI: 10.1039/c9sc01928f , 2019, Chemical science.

[45]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[46]  Wenhao Gao,et al.  The Synthesizability of Molecules Proposed by Generative Models , 2020, J. Chem. Inf. Model..

[47]  Venkat Venkatasubramanian,et al.  Computer-aided molecular design using genetic algorithms , 1994 .

[48]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.

[49]  Matt J. Kusner,et al.  Grammar Variational Autoencoder , 2017, ICML.

[50]  Gisbert Schneider.  De novo Molecular Design: Schneider/De novo Molecular Design , 2013 .

[51]  Thierry Kogej,et al.  Generating Focused Molecule Libraries for Drug Discovery with Recurrent Neural Networks , 2017, ACS central science.

[52]  G. Pez,et al.  Electrophilic NF Fluorinating Agents. , 1996, Chemical reviews.

[53]  Zois Boukouvalas,et al.  Deep learning for molecular generation and optimization - a review of the state of the art , 2019, Molecular Systems Design & Engineering.