Property-Controllable Generation of Quaternary Ammonium Compounds

Designing molecules with desired biological properties remains an outstanding challenge both in the wet and dry laboratories. Meeting this challenge promises great translational impacts across drug discovery, material sciences, biotechnology, and more. Recent momentum in deep learning promises to advance our computational capabilities on molecule generation. In particular, deep graph generative models which treat molecule design as a graph generation problem are allowing us to directly learn from existing databases of small molecules and generate novel, valid molecules. Currently, these models have many shortcomings, including poor controllability of desired molecular properties, especially in practical application where the training data is usually small, noisy, and incomplete. This paper focuses on equipping graph variational autoencoders with the ability to control for desired properties and its practical application in a practical application which is the generation of Quaternary Ammonium Compounds (QAC). Several controllable graph generation mechanisms are investigated for their effectiveness. A general framework is then proposed to extend these mechanisms by our newly proposed objective function to handle the challenges in practical applications where the property value annotations are usually censored and not fully available in all training samples. The experimental evaluation considers an experimentally-characterized dataset of antimicrobial small molecules with wet-lab characterized activity against antibiotic-resistant bacteria. Extensive experiments demonstrate the superiority of the proposed models and control of desired properties.

[1]  Amarda Shehu,et al.  Small molecule generation via disentangled representation learning , 2022, Bioinform..

[2]  Shengchao Liu,et al.  MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design , 2022, ArXiv.

[3]  Amarda Shehu,et al.  Interpretable Molecular Graph Generation via Monotonic Constraints , 2022, SDM.

[4]  Amarda Shehu,et al.  Deep Latent-Variable Models for Controllable Molecule Generation , 2021, 2021 IEEE International Conference on Bioinformatics and Biomedicine (BIBM).

[5]  Daniel Rothchild,et al.  C5T5: Controllable Generation of Organic Molecules with Transformers , 2021, ArXiv.

[6]  Alon Shoshan,et al.  GAN-Control: Explicitly Controllable GANs , 2021, 2021 IEEE/CVF International Conference on Computer Vision (ICCV).

[7]  Xiaojie Guo,et al.  Interpretable Molecule Generation via Disentanglement Learning , 2020 .

[8]  Alan Aspuru-Guzik,et al.  Graph Deconvolutional Generation , 2020, ArXiv.

[9]  David Duvenaud,et al.  Invertible Residual Networks , 2018, ICML.

[10]  Qi Liu,et al.  Constrained Graph Variational Autoencoders for Molecule Design , 2018, NeurIPS.

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

[12]  Nikos Komodakis,et al.  GraphVAE: Towards Generation of Small Graphs Using Variational Autoencoders , 2018, ICANN.

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

[14]  Carl Doersch,et al.  Tutorial on Variational Autoencoders , 2016, ArXiv.

[15]  Petra Schneider,et al.  De Novo Design at the Edge of Chaos. , 2016, Journal of medicinal chemistry.

[16]  Shakir Mohamed,et al.  Variational Inference with Normalizing Flows , 2015, ICML.

[17]  Yoshua Bengio,et al.  Generative Adversarial Nets , 2014, NIPS.

[18]  Pavlo O. Dral,et al.  Quantum chemistry structures and properties of 134 kilo molecules , 2014, Scientific Data.

[19]  Jean-Louis Reymond,et al.  Enumeration of 166 Billion Organic Small Molecules in the Chemical Universe Database GDB-17 , 2012, J. Chem. Inf. Model..

[20]  Jürgen Bajorath,et al.  Similarity searching , 2011 .

[21]  Yuanqi Du,et al.  Property Controllable Variational Autoencoder via Invertible Mutual Dependence , 2021, ICLR.

[22]  George M Whitesides,et al.  Reinventing chemistry. , 2015, Angewandte Chemie.