Search

This work proposes a novel approach for drug molecule design using data-assisted techniques. This approach leverages a generation-based framework to expedite the drug discovery process, aiming to identify candidate molecules suitable for production while minimizing development timelines and regulatory hurdles. The core of the proposed method is a conditional variational autoencoder (CVAE) for molecule generation, employing NCSMILES string representation. The framework involves three key stages: (1) molecule generation using the CVAE, (2) filtering based on a scoring function, and (3) identification of the optimal molecule from the generated pool. To enhance the latent space representation, we incorporate molecule properties alongside conditional selection criteria. The performance of the proposed scheme is comprehensively evaluated on standard benchmark datasets using various metrics, including validity, diversity, usefulness, and novelty. The method demonstrates superior performance compared to existing state-of-the-art approaches, attributable to several key improvements, including intermediary optimizations and condition-based selection. [ABSTRACT FROM AUTHOR]

Abstract SMILES-based generative models are amongst the most robust and successful recent methods used to augment drug design. They are typically used for complete de novo generation, however, scaffold decoration and fragment linking applications are sometimes desirable which requires a different grammar, architecture, training dataset and therefore, re-training of a new model. In this work, we describe a simple procedure to conduct constrained molecule generation with a SMILES-based generative model to extend applicability to scaffold decoration and fragment linking by providing SMILES prompts, without the need for re-training. In combination with reinforcement learning, we show that pre-trained, decoder-only models adapt to these applications quickly and can further optimize molecule generation towards a specified objective. We compare the performance of this approach to a variety of orthogonal approaches and show that performance is comparable or better. For convenience, we provide an easy-to-use python package to facilitate model sampling which can be found on GitHub and the Python Package Index. Scientific contribution This novel method extends an autoregressive chemical language model to scaffold decoration and fragment linking scenarios. This doesn’t require re-training, the use of a bespoke grammar, or curation of a custom dataset, as commonly required by other approaches.

Abstract Generative models are undergoing rapid research and application to de novo drug design. To facilitate their application and evaluation, we present MolScore. MolScore already contains many drug-design-relevant scoring functions commonly used in benchmarks such as, molecular similarity, molecular docking, predictive models, synthesizability, and more. In addition, providing performance metrics to evaluate generative model performance based on the chemistry generated. With this unification of functionality, MolScore re-implements commonly used benchmarks in the field (such as GuacaMol, MOSES, and MolOpt). Moreover, new benchmarks can be created trivially. We demonstrate this by testing a chemical language model with reinforcement learning on three new tasks of increasing complexity related to the design of 5-HT2a ligands that utilise either molecular descriptors, 266 pre-trained QSAR models, or dual molecular docking. Lastly, MolScore can be integrated into an existing Python script with just three lines of code. This framework is a step towards unifying generative model application and evaluation as applied to drug design for both practitioners and researchers. The framework can be found on GitHub and downloaded directly from the Python Package Index. Scientific Contribution MolScore is an open-source platform to facilitate generative molecular design and evaluation thereof for application in drug design. This platform takes important steps towards unifying existing benchmarks, providing a platform to share new benchmarks, and improves customisation, flexibility and usability for practitioners over existing solutions. Graphical Abstract

SMILES-based generative models are amongst the most robust and successful recent methods used to augment drug design. They are typically used for complete de novo generation, however, scaffold decoration and fragment linking applications are sometimes desirable which requires a different grammar, architecture, training dataset and therefore, re-training of a new model. In this work, we describe a simple procedure to conduct constrained molecule generation with a SMILES-based generative model to extend applicability to scaffold decoration and fragment linking by providing SMILES prompts, without the need for re-training. In combination with reinforcement learning, we show that pre-trained, decoder-only models adapt to these applications quickly and can further optimize molecule generation towards a specified objective. We compare the performance of this approach to a variety of orthogonal approaches and show that performance is comparable or better. For convenience, we provide an easy-to-use python package to facilitate model sampling which can be found on GitHub and the Python Package Index. Scientific contribution This novel method extends an autoregressive chemical language model to scaffold decoration and fragment linking scenarios. This doesn't require re-training, the use of a bespoke grammar, or curation of a custom dataset, as commonly required by other approaches. [ABSTRACT FROM AUTHOR]

Generative models are undergoing rapid research and application to de novo drug design. To facilitate their application and evaluation, we present MolScore. MolScore already contains many drug-design-relevant scoring functions commonly used in benchmarks such as, molecular similarity, molecular docking, predictive models, synthesizability, and more. In addition, providing performance metrics to evaluate generative model performance based on the chemistry generated. With this unification of functionality, MolScore re-implements commonly used benchmarks in the field (such as GuacaMol, MOSES, and MolOpt). Moreover, new benchmarks can be created trivially. We demonstrate this by testing a chemical language model with reinforcement learning on three new tasks of increasing complexity related to the design of 5-HT2a ligands that utilise either molecular descriptors, 266 pre-trained QSAR models, or dual molecular docking. Lastly, MolScore can be integrated into an existing Python script with just three lines of code. This framework is a step towards unifying generative model application and evaluation as applied to drug design for both practitioners and researchers. The framework can be found on GitHub and downloaded directly from the Python Package Index. Scientific Contribution MolScore is an open-source platform to facilitate generative molecular design and evaluation thereof for application in drug design. This platform takes important steps towards unifying existing benchmarks, providing a platform to share new benchmarks, and improves customisation, flexibility and usability for practitioners over existing solutions. [ABSTRACT FROM AUTHOR]

Designing functional molecules is the prerogative of experts who have advanced knowledge and experience in their fields. To democratize automatic molecular design for both experts and nonexperts, we introduce a generic open‐sourced framework, ChemTSv2, to design molecules based on a de novo molecule generator equipped with an easy‐to‐use interface. Besides, ChemTSv2 can easily be integrated with various simulation packages, such as Gaussian 16 package, and supports a massively parallel exploration that accelerates molecular designs. We exhibit the potential of molecular design with ChemTSv2, including previous work, such as chromophores, fluorophores, drugs, and so forth. ChemTSv2 contributes to democratizing inverse molecule design in various disciplines relevant to chemistry. This article is categorized under:Data Science > Databases and Expert SystemsData Science > Artificial Intelligence/Machine LearningData Science > Computer Algorithms and Programming [ABSTRACT FROM AUTHOR]

Abstract Computational molecular design can yield chemically unreasonable compounds when performed carelessly. A popular strategy to mitigate this risk is mimicking reference chemistry. This is commonly achieved by restricting the way in which molecules are constructed or modified. While it is well established that such an approach helps in designing chemically appealing molecules, concerns about these restrictions impacting chemical space exploration negatively linger. In this work we present a software library for constrained graph-based molecule manipulation and showcase its functionality by developing a molecule generator. Said generator designs molecules mimicking reference chemical features of differing granularity. We find that restricting molecular construction lightly, beyond the usual positive effects on drug-likeness and synthesizability of designed molecules, provides guidance to optimization algorithms navigating chemical space. Nonetheless, restricting molecular construction excessively can indeed hinder effective chemical space exploration.

Computational molecular design can yield chemically unreasonable compounds when performed carelessly. A popular strategy to mitigate this risk is mimicking reference chemistry. This is commonly achieved by restricting the way in which molecules are constructed or modified. While it is well established that such an approach helps in designing chemically appealing molecules, concerns about these restrictions impacting chemical space exploration negatively linger. In this work we present a software library for constrained graph-based molecule manipulation and showcase its functionality by developing a molecule generator. Said generator designs molecules mimicking reference chemical features of differing granularity. We find that restricting molecular construction lightly, beyond the usual positive effects on drug-likeness and synthesizability of designed molecules, provides guidance to optimization algorithms navigating chemical space. Nonetheless, restricting molecular construction excessively can indeed hinder effective chemical space exploration. [ABSTRACT FROM AUTHOR]

Recently, artificial intelligence (AI)-enabled de novo molecular generators (DNMGs) have automated molecular design based on data-driven or simulation-based property estimates. In some domains like the game of Go where AI surpassed human intelligence, humans are trying to learn from AI about the best strategy of the game. To understand DNMG’s strategy of molecule optimization, we propose an algorithm called characteristic functional group monitoring (CFGM). Given a time series of generated molecules, CFGM monitors statistically enriched functional groups in comparison to the training data. In the task of absorption wavelength maximization of pure organic molecules (consisting of H, C, N, and O), we successfully identified a strategic change from diketone and aniline derivatives to quinone derivatives. In addition, CFGM led us to a hypothesis that 1,2-quinone is an unconventional chromophore, which was verified with chemical synthesis. This study shows the possibility that human experts can learn from DNMGs to expand their ability to discover functional molecules.

Designing an optimal and desired drug molecule structure is a challenging problem. Most of the existing solutions/representations reported in the literature for this problem are complex and time consuming. This is due to larger datasets with longer training periods and long learning dependencies. Deep learning's generative model can be used to enable chemical modelling to generate molecules without explicit complex molecular rules. However, Deep Learning models (LSTM based VAE) suffer from posterior collapse, larger vocabulary of datasets and sub-optimal latent space searching mechanisms. Motivated by this, we propose a recently researched idea of Normalized Reparameterized conditional Variational Autoencoder with applied beam search in latent space (NRC-VABS). The resulting model with normalized vocabulary, conditionally augmented dataset and revised/reparameterized loss function addresses posterior collapse and constructs continuous and consistent latent space for exploitation by beam search during generation stages. The conditions/properties of desirable molecules are specified through a condition vector and is used while training as well as during generation of drug molecules. Beam search is coined on improved normalized SMILES representation. The idea entails by creating samples with beam search and filtering them depending on their condition and identifying the optimal molecules with desired properties. Normalization also improves the information and reduces complexity in latent space. To address the diversity of the generated molecules, a tunable parameter (D) is also used. Various performance evaluation metrics, such as validity, uniqueness, novelty, accuracy, and Frechet ChemNet Distance are used to evaluate the NRC-VABS on benchmark data sets such as GDB13, MOSES and subset of 250k ZINC molecules. The performance of the NRC-VABS is compared with state-of-the-art peer techniques. NRC-VABS generates molecules at validity range from 92% to 84%, Accuracy 89% to 97% at varied level of diversities (D = 1, D = 2 and D = 3). An application of the proposal in terms interpolation and controlling other (2 of 3) properties by varying one (1 of 3) property at a time. Generating only target molecules with desired properties and maintaining diversity improves novel molecules while greatly reducing time complexity as only novel and desired molecules can be generated. • NRC-VABS improved SMILES notation for reducing complexity in the SMILES. • A solution to posterior collapse is proposed that enhance the NRC-VABS performance. • Latent Space is explored by using beam search algorithm. [ABSTRACT FROM AUTHOR]

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. [ABSTRACT FROM AUTHOR]

Introduction: Deep learning approaches have become popular in recent years in de novo drug design. Generative models for molecule generation and optimization have shown promising results. Molecules trained on different chemical data could regenerate molecules that were similar to the query molecule, thus supporting lead optimization. Recurrent neural network-based generative models have demonstrated application in low-data drug discovery, fragment-based drug design and in lead optimization., Areas Covered: In this review, we have provided an overview of recurrent neural network models and their variants for molecule generation with recent examples. The input representation of molecules as SMILES and molecular graphs have been discussed. The evaluation benchmarks and metrics used in generative neural network models are also highlighted. For this, ScienceDirect, Web of Science, and Google Scholar databases were searched with the article's keywords and their combinations to retrieve the most relevant and up-to-date information., Expert Opinion: The simplicity of SMILES notation makes it suitable for training a sequence-based model such as a recurrent neural network. However, models that could be trained on molecular graphs to generate molecular structures which could be synthesized could open new possibility for valid molecule generation and synthetic feasibility.

Recently, artificial intelligence (AI)-enabled de novo molecular generators (DNMGs) have automated molecular design based on data-driven or simulation-based property estimates. In some domains like the game of Go where AI surpassed human intelligence, humans are trying to learn from AI about the best strategy of the game. To understand DNMG's strategy of molecule optimization, we propose an algorithm called characteristic functional group monitoring (CFGM). Given a time series of generated molecules, CFGM monitors statistically enriched functional groups in comparison to the training data. In the task of absorption wavelength maximization of pure organic molecules (consisting of H, C, N, and O), we successfully identified a strategic change from diketone and aniline derivatives to quinone derivatives. In addition, CFGM led us to a hypothesis that 1,2-quinone is an unconventional chromophore, which was verified with chemical synthesis. This study shows the possibility that human experts can learn from DNMGs to expand their ability to discover functional molecules., Competing Interests: No potential conflict of interest was reported by the author(s)., (© 2022 The Author(s). Published by National Institute for Materials Science in partnership with Taylor & Francis Group.)

Artificial intelligence (AI) has undergone rapid development in recent years and has been successfully applied to real-world problems such as drug design. In this chapter, we review recent applications of AI to problems in drug design including virtual screening, computer-aided synthesis planning, and de novo molecule generation, with a focus on the limitations of the application of AI therein and opportunities for improvement. Furthermore, we discuss the broader challenges imposed by AI in translating theoretical practice to real-world drug design; including quantifying prediction uncertainty and explaining model behavior., (© 2022. The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature.)

The de novo molecule generation problem involves generating novel or modified molecular structures with desirable properties. Taking advantage of the great representation learning ability of deep learning models, deep generative models, which differ from discriminative models in their traditional machine learning approach, provide the possibility of generation of desirable molecules directly. Although deep generative models have been extensively discussed in the machine learning community, a specific investigation of the computational issues related to deep generative models for de novo molecule generation is needed. A concise and insightful discussion of recent advances in applying deep generative models for de novo molecule generation is presented, with particularly emphasizing the most important challenges for successful application of deep generative models in this specific area. This article is categorized under: Computer and Information Science > ChemoinformaticsComputer and Information Science > Computer Algorithms and Programming Distinctive deep generative models open a new path for efficient de novo molecule generation in computational molecular science. The framework of five distinctive deep generative models, for example, VAE‐based, AAE‐based, GAN‐based, RNN‐based, and hybrid models coupling with reinforcement learning are briefly described, providing a visual and general understanding of the underlying rationales of these models for molecule design. [ABSTRACT FROM AUTHOR]