Explainable graph neural networks for organic cages

Design and assembly of porous organic cages

Porous organic cages (POCs) represent a notable category of porous materials, showing remarkable material properties due to their inherent porosity. Unlike extended frameworks which are constructed by strong covalent or coordination bonds, POCs are composed of discrete molecular units held together by weak intermolecular forces. Their structure and chemical traits can be systematically tailored, making them suitable for a range of applications including gas storage and separation, molecular separation and recognition, catalysis, and proton and ion conduction. This review provides a comprehensive overview of POCs, covering their synthesis methods, structure and properties, computational approaches, and applications, serving as a primer for those who are new to the domain. A special emphasis is placed on the growing role of computational methods, highlighting how advanced data-driven techniques and automation are increasingly aiding the rapid exploration and understanding of POCs. We conclude by addressing the prevailing challenges and future prospects in the field.

Fine-tuning GPT-3 for machine learning electronic and functional properties of organic molecules

We evaluate the effectiveness of fine-tuning GPT-3 for the prediction of electronic and functional properties of organic molecules. Our findings show that fine-tuned GPT-3 can successfully identify and distinguish between chemically meaningful patterns, and discern subtle differences among them, exhibiting robust predictive performance for the prediction of molecular properties. We focus on assessing the fine-tuned models' resilience to information loss, resulting from the absence of atoms or chemical groups, and to noise that we introduce via random alterations in atomic identities. We discuss the challenges and limitations inherent to the use of GPT-3 in molecular machine-learning tasks and suggest potential directions for future research and improvements to address these issues.

Deep generative design of porous organic cages via a variational autoencoder

Porous organic cages (POCs) are a class of porous molecular materials characterised by their tunable, intrinsic porosity; this functional property makes them candidates for applications including guest storage and separation. Typically formed via dynamic covalent chemistry reactions from multifunctionalised molecular precursors, there is an enormous potential chemical space for POCs due to the fact they can be formed by combining two relatively small organic molecules, which themselves have an enormous chemical space. However, identifying suitable molecular precursors for POC formation is challenging, as POCs often lack shape persistence (the cage collapses upon solvent removal with loss of its cavity), thus losing a key functional property (porosity). Generative machine learning models have potential for targeted computational design of large functional molecular systems such as POCs. Here, we present a deep-learning-enabled generative model, Cage-VAE, for the targeted generation of shape-persistent POCs. We demonstrate the capacity of Cage-VAE to propose novel, shape-persistent POCs, via integration with multiple efficient sampling methods, including Bayesian optimisation and spherical linear interpolation.

Interpretable Attribution Assignment for Octanol-Water Partition Coefficient

With the increasing development of machine learning models, their credibility has become an important issue. In chemistry, attribution assignment is gaining relevance when it comes to designing molecules and debugging models. However, attention has only been paid to which atoms are important in the prediction and not to whether the attribution is reasonable. In this study, we developed a graph neural network model, a highly interpretable attribution model in chemistry, and modified the integrated gradients method. The credibility of our approach was confirmed by predicting the octanol-water partition coefficient (logP) and evaluating the three metrics (accuracy, consistency, and stability) in the attribution assignment.

Computational modeling to assist in the discovery of supramolecular materials

Computational modeling is increasingly used to assist in the discovery of supramolecular materials. Supramolecular materials are typically primarily built from organic components that are self-assembled through noncovalent bonding and have potential applications, including in selective binding, sorption, molecular separations, catalysis, optoelectronics, sensing, and as molecular machines. In this review, the key areas where computational prediction can assist in the discovery of supramolecular materials, including in structure prediction, property prediction, and the prediction of how to synthesize a hypothetical material are discussed, before exploring the potential impact of artificial intelligence techniques on the field. Throughout, the importance of close integration with experimental materials discovery programs will be highlighted. A series of case studies from the author's work across some different supramolecular material classes will be discussed, before finishing with a discussion of the outlook for the field.

Graph neural networks for materials science and chemistry

Machine learning plays an increasingly important role in many areas of chemistry and materials science, being used to predict materials properties, accelerate simulations, design new structures, and predict synthesis routes of new materials. Graph neural networks (GNNs) are one of the fastest growing classes of machine learning models. They are of particular relevance for chemistry and materials science, as they directly work on a graph or structural representation of molecules and materials and therefore have full access to all relevant information required to characterize materials. In this Review, we provide an overview of the basic principles of GNNs, widely used datasets, and state-of-the-art architectures, followed by a discussion of a wide range of recent applications of GNNs in chemistry and materials science, and concluding with a road-map for the further development and application of GNNs.