Machine-enabled inverse design of inorganic solid materials: promises and challenges

Developing high-performance advanced materials requires a deeper insight and search into the chemical space. Until recently, exploration of materials space using chemical intuitions built upon existing materials has been the general strategy, but this direct design approach is often time and resource consuming and poses a significant bottleneck to solve the materials challenges of future sustainability in a timely manner. To accelerate this conventional design process, inverse design, which outputs materials with pre-defined target properties, has emerged as a significant materials informatics platform in recent years by leveraging hidden knowledge obtained from materials data. Here, we summarize the latest progress in machine-enabled inverse materials design categorized into three strategies: high-throughput virtual screening, global optimization, and generative models. We analyze challenges for each approach and discuss gaps to be bridged for further accelerated and rational data-driven materials design.

The grand challenge of materials science, discovery of novel materials with target properties, can be greatly accelerated by machine-learned inverse design strategies.     

Attention-based generative models for de novo molecular design

Attention mechanisms have led to many breakthroughs in sequential data modeling but have yet to be incorporated into any generative algorithms for molecular design. Here we explore the impact of adding self-attention layers to generative β-VAE models and show that those with attention are able to learn a complex “molecular grammar” while improving performance on downstream tasks such as accurately sampling from the latent space (“model memory”) or exploring novel chemistries not present in the training data. There is a notable relationship between a model''s architecture, the structure of its latent memory and its performance during inference. We demonstrate that there is an unavoidable tradeoff between model exploration and validity that is a function of the complexity of the latent memory. However, novel sampling schemes may be used that optimize this tradeoff. We anticipate that attention will play an important role in future molecular design algorithms that can make efficient use of the detailed molecular substructures learned by the transformer.

An implementation of attention within the variational autoencoder framework for continuous representation of molecules. The addition of attention significantly increases model performance for complex tasks such as exploration of novel chemistries.     

Evolutionary chemical space exploration for functional materials: computational organic semiconductor discovery

Computational methods, including crystal structure and property prediction, have the potential to accelerate the materials discovery process by enabling structure prediction and screening of possible molecular building blocks prior to their synthesis. However, the discovery of new functional molecular materials is still limited by the need to identify promising molecules from a vast chemical space. We describe an evolutionary method which explores a user specified region of chemical space to identify promising molecules, which are subsequently evaluated using crystal structure prediction. We demonstrate the methods for the exploration of aza-substituted pentacenes with the aim of finding small molecule organic semiconductors with high charge carrier mobilities, where the space of possible substitution patterns is too large to exhaustively search using a high throughput approach. The method efficiently explores this large space, typically requiring calculations on only ∼1% of molecules during a search. The results reveal two promising structural motifs: aza-substituted naphtho[1,2-a]anthracenes with reorganisation energies as low as pentacene and a series of pyridazine-based molecules having both low reorganisation energies and high electron affinities.

Evolutionary optimisation and crystal structure prediction are used to explore chemical space for molecular organic semiconductors.     

Defining and navigating macrocycle chemical space

Macrocyclic compounds (MCs) are of growing interest for inhibition of challenging drug targets. We consider afresh what structural and physicochemical features could be relevant to the bioactivity of this compound class. Using these features, we performed Principal Component Analysis to map oral and non-oral macrocycle drugs and clinical candidates, and also commercially available synthetic MCs, in structure–property space. We find that oral MC drugs occupy defined regions that are distinct from those of the non-oral MC drugs. None of the oral MC regions are effectively sampled by the synthetic MCs. We identify 13 properties that can be used to design synthetic MCs that sample regions overlapping with oral MC drugs. The results advance our understanding of what molecular features are associated with bioactive and orally bioavailable MCs, and illustrate an approach by which synthetic chemists can better evaluate MC designs. We also identify underexplored regions of macrocycle chemical space.

Macrocyclic compounds (MCs) are of high interest for inhibition of challenging drug targets, but existing oral MC drugs occupy regions of chemical space that are not well sampled by many available synthetic MC chemotypes.     

Scaffold-based molecular design with a graph generative model

Searching for new molecules in areas like drug discovery often starts from the core structures of known molecules. Such a method has called for a strategy of designing derivative compounds retaining a particular scaffold as a substructure. On this account, our present work proposes a graph generative model that targets its use in scaffold-based molecular design. Our model accepts a molecular scaffold as input and extends it by sequentially adding atoms and bonds. The generated molecules are then guaranteed to contain the scaffold with certainty, and their properties can be controlled by conditioning the generation process on desired properties. The learned rule of extending molecules can well generalize to arbitrary kinds of scaffolds, including those unseen during learning. In the conditional generation of molecules, our model can simultaneously control multiple chemical properties despite the search space constrained by fixing the substructure. As a demonstration, we applied our model to designing inhibitors of the epidermal growth factor receptor and show that our model can employ a simple semi-supervised extension to broaden its applicability to situations where only a small amount of data is available.

We propose a scaffold-based graph generative model for designing novel drug candidates that include the desired scaffold as a substructure.     

Linchpins empower promiscuous electrophiles to enable site-selective modification of histidine and aspartic acid in proteins

The conservation of chemoselectivity becomes invalid for multiple electrophilic warheads during protein bioconjugation. Consequently, it leads to unpredictable heterogeneous labeling of proteins. Here, we report that a linchpin can create a unique chemical space to enable site-selectivity for histidine and aspartic acid modifications overcoming the pre-requisite of chemoselectivity.

Linchpin-enabled promiscuous electrophile uncovers an unchartered reactivity landscape for the precision engineering of proteins.     

Structure-based de novo drug design using 3D deep generative models

Deep generative models are attracting much attention in the field of de novo molecule design. Compared to traditional methods, deep generative models can be trained in a fully data-driven way with little requirement for expert knowledge. Although many models have been developed to generate 1D and 2D molecular structures, 3D molecule generation is less explored, and the direct design of drug-like molecules inside target binding sites remains challenging. In this work, we introduce DeepLigBuilder, a novel deep learning-based method for de novo drug design that generates 3D molecular structures in the binding sites of target proteins. We first developed Ligand Neural Network (L-Net), a novel graph generative model for the end-to-end design of chemically and conformationally valid 3D molecules with high drug-likeness. Then, we combined L-Net with Monte Carlo tree search to perform structure-based de novo drug design tasks. In the case study of inhibitor design for the main protease of SARS-CoV-2, DeepLigBuilder suggested a list of drug-like compounds with novel chemical structures, high predicted affinity, and similar binding features to those of known inhibitors. The current version of L-Net was trained on drug-like compounds from ChEMBL, which could be easily extended to other molecular datasets with desired properties based on users'' demands and applied in functional molecule generation. Merging deep generative models with atomic-level interaction evaluation, DeepLigBuilder provides a state-of-the-art model for structure-based de novo drug design and lead optimization.

DeepLigBuilder, a novel deep generative model for structure-based de novo drug design, directly generates 3D structures of drug-like compounds in the target binding site.     

Is it time for biocatalysis in fragment-based drug discovery?

The use of biocatalysts for fragment-based drug discovery has yet to be fully investigated, despite the promise enzymes hold for the synthesis of poly-functional, non-protected small molecules. Here we analyze products of the biocatalysis literature to demonstrate the potential for not only fragment generation, but also the enzyme-mediated elaboration of these fragments. Our analysis demonstrates that biocatalytic products can readily populate 3D chemical space, offering diverse catalytic approaches to help generate new, bioactive molecules.

This perspective discusses how biocatalysis could play an important role in the future fragment-based drug discovery.     

Retrosynthetic accessibility score (RAscore) – rapid machine learned synthesizability classification from AI driven retrosynthetic planning

Computer aided synthesis planning (CASP) is part of a suite of artificial intelligence (AI) based tools that are able to propose synthesis routes to a wide range of compounds. However, at present they are too slow to be used to screen the synthetic feasibility of millions of generated or enumerated compounds before identification of potential bioactivity by virtual screening (VS) workflows. Herein we report a machine learning (ML) based method capable of classifying whether a synthetic route can be identified for a particular compound or not by the CASP tool AiZynthFinder. The resulting ML models return a retrosynthetic accessibility score (RAscore) of any molecule of interest, and computes at least 4500 times faster than retrosynthetic analysis performed by the underlying CASP tool. The RAscore should be useful for pre-screening millions of virtual molecules from enumerated databases or generative models for synthetic accessibility and produce higher quality databases for virtual screening of biological activity.

The retrosynthetic accessibility score (RAscore) is based on AI driven retrosynthetic planning, and is useful for rapid scoring of synthetic feasability and pre-screening of large datasets of virtual/generated molecules.     

Controlled synthesis of organic two-dimensional nanostructures via reaction-driven,cooperative supramolecular polymerization

The bottom-up approach of supramolecular polymerization is an effective synthetic method for functional organic nanostructures. However, the uncontrolled growth and polydisperse structural outcome often lead to low functional efficiency. Thus, precise control over the structural characteristics of supramolecular polymers is the current scientific hurdle. Research so far has tended to focus on systems with inherent kinetic control by the presence of metastable state monomers either through conformational molecular design or by exploring pathway complexity. The need of the hour is to create generic strategies for dormant states of monomers that can be extended to different molecules and various structural organizations and dimensions. Here we venture to demonstrate chemical reaction-driven cooperative supramolecular polymerization as an alternative strategy for the controlled synthesis of organic two-dimensional nanostructures. In our approach, the dynamic imine bond is exploited to convert a non-assembling dormant monomer to an activated amphiphilic structure in a kinetically controlled manner. The chemical reaction governed retarded nucleation–elongation growth provides control over dispersity and size.

We report the kinetically controlled supramolecular polymerization of organic two-dimensional charge-transfer nanostructures via a chemical reaction (imine)-driven approach.     

A review on machine learning algorithms for the ionic liquid chemical space

There are thousands of papers published every year investigating the properties and possible applications of ionic liquids. Industrial use of these exceptional fluids requires adequate understanding of their physical properties, in order to create the ionic liquid that will optimally suit the application. Computational property prediction arose from the urgent need to minimise the time and cost that would be required to experimentally test different combinations of ions. This review discusses the use of machine learning algorithms as property prediction tools for ionic liquids (either as standalone methods or in conjunction with molecular dynamics simulations), presents common problems of training datasets and proposes ways that could lead to more accurate and efficient models.

In this review article, the authors discuss the use of machine learning algorithms as tools for the prediction of physical and chemical properties of ionic liquids.     

Development and application of ubiquitin-based chemical probes

Protein ubiquitination regulates almost every process in eukaryotic cells. The study of the many enzymes involved in the ubiquitination system and the development of ubiquitination-associated therapeutics are important areas of current research. Synthetic tools such as ubiquitin-based chemical probes have been making an increasing contribution to deciphering various biochemical components involved in ubiquitin conjugation, recruitment, signaling, and deconjugation. In the present minireview, we summarize the progress of ubiquitin-based chemical probes with an emphasis on their various structures and chemical synthesis. We discuss the utility of the ubiquitin-based chemical probes for discovering and profiling ubiquitin-dependent signaling systems, as well as the monitoring and visualization of ubiquitin-related enzymatic machinery. We also show how the probes can serve to elucidate the molecular mechanism of recognition and catalysis. Collectively, the development and application of ubiquitin-based chemical probes emphasizes the importance and utility of chemical protein synthesis in modern chemical biology.

This article reviews the design, synthesis, and application of different classes of Ub-based chemical probes.     

Relating atomic energy,radius and electronegativity through compression

Trends in atomic properties are well-established tools for guiding the analysis and discovery of materials. Here, we show how compression can reveal a long sought-after connection between two central chemical concepts – van-der-Waals (vdW) radii and electronegativity – and how these relate to the driving forces behind chemical and physical transformations.

Compression is used to derive a long sought-after connection between two central chemical concepts – van-der-Waals (vdW) radii and electronegativity – and how these relate to the driving forces behind chemical and physical transformations.     

On the crystal chemistry of inorganic nitrides: crystal-chemical parameters,bonding behavior,and opportunities in the exploration of their compositional space

The scarcity of nitrogen in Earth''s crust, combined with challenging synthesis, have made inorganic nitrides a relatively unexplored class of compounds compared to their naturally abundant oxide counterparts. To facilitate exploration of their compositional space via a priori modeling, and to help a posteriori structure verification not limited to inferring the oxidation state of redox-active cations, we derive a suite of bond-valence parameters and Lewis acid strength values for 76 cations observed bonding to N³⁻, and further outline a baseline statistical knowledge of bond lengths for these compounds. Examination of structural and electronic effects responsible for the functional properties and anomalous bonding behavior of inorganic nitrides shows that many mechanisms of bond-length variation ubiquitous to oxide and oxysalt compounds (e.g., lone-pair stereoactivity, the Jahn–Teller and pseudo Jahn–Teller effects) are similarly pervasive in inorganic nitrides, and are occasionally observed to result in greater distortion magnitude than their oxide counterparts. We identify promising functional units for exploring uncharted chemical spaces of inorganic nitrides, e.g. multiple-bond metal centers with promise regarding the development of a post-Haber–Bosch process proceeding at milder reaction conditions, and promote an atomistic understanding of chemical bonding in nitrides relevant to such pursuits as the development of a model of ion substitution in solids, a problem of great relevance to semiconductor doping whose solution would fast-track the development of compound solar cells, battery materials, electronics, and more.

Navigating high-return chemical spaces in inorganic nitrides via identification of coordination units bearing functional properties.     

Unraveling the effect of defects,domain size,and chemical doping on photophysics and charge transport in covalent organic frameworks

Understanding the underlying physical mechanisms that govern charge transport in two-dimensional (2D) covalent organic frameworks (COFs) will facilitate the development of novel COF-based devices for optoelectronic and thermoelectric applications. In this context, the low-energy mid-infrared absorption contains valuable information about the structure–property relationships and the extent of intra- and inter-framework “hole” polaron delocalization in doped and undoped polymeric materials. In this study, we provide a quantitative characterization of the intricate interplay between electronic defects, domain sizes, pore volumes, chemical dopants, and three dimensional anisotropic charge migration in 2D COFs. We compare our simulations with recent experiments on doped COF films and establish the correlations between polaron coherence, conductivity, and transport signatures. By obtaining the first quantitative agreement with the measured absorption spectra of iodine doped (aza)triangulene-based COF, we highlight the fundamental differences between the underlying microstructure, spectral signatures, and transport physics of polymers and COFs. Our findings provide conclusive evidence of why iodine doped COFs exhibit lower conductivity compared to doped polythiophenes. Finally, we propose new research directions to address existing limitations and improve charge transport in COFs for applications in functional molecular electronic devices.

This study highlights the importance of mid-infrared spectral signatures and discusses the fundamental mechanisms driving charge transport in COFs. Our analysis can hopefully guide the rational design of new COFs yielding higher conductivities.     

Chemical control of peptide material phase transitions

Progressive solute-rich polymer phase transitions provide pathways for achieving ordered supramolecular assemblies. Intrinsically disordered protein domains specifically regulate information in biological networks via conformational ordering. Here we consider a molecular tagging strategy to control ordering transitions in polymeric materials and provide a proof-of-principle minimal peptide phase network captured with a dynamic chemical network.

Substrate initiated assembly of a dynamic chemical network.     

Physically inspired deep learning of molecular excitations and photoemission spectra

Modern functional materials consist of large molecular building blocks with significant chemical complexity which limits spectroscopic property prediction with accurate first-principles methods. Consequently, a targeted design of materials with tailored optoelectronic properties by high-throughput screening is bound to fail without efficient methods to predict molecular excited-state properties across chemical space. In this work, we present a deep neural network that predicts charged quasiparticle excitations for large and complex organic molecules with a rich elemental diversity and a size well out of reach of accurate many body perturbation theory calculations. The model exploits the fundamental underlying physics of molecular resonances as eigenvalues of a latent Hamiltonian matrix and is thus able to accurately describe multiple resonances simultaneously. The performance of this model is demonstrated for a range of organic molecules across chemical composition space and configuration space. We further showcase the model capabilities by predicting photoemission spectra at the level of the GW approximation for previously unseen conjugated molecules.

A physically-inspired machine learning model for orbital energies is developed that can be augmented with delta learning to obtain photoemission spectra, ionization potentials, and electron affinities with experimental accuracy.     

Illuminating elite patches of chemical space

In the past few years, there has been considerable activity in both academic and industrial research to develop innovative machine learning approaches to locate novel, high-performing molecules in chemical space. Here we describe a new and fundamentally different type of approach that provides a holistic overview of how high-performing molecules are distributed throughout a search space. Based on an open-source, graph-based implementation [J. H. Jensen, Chem. Sci., 2019, 10, 3567–3572] of a traditional genetic algorithm for molecular optimisation, and influenced by state-of-the-art concepts from soft robot design [J. B. Mouret and J. Clune, Proceedings of the Artificial Life Conference, 2012, pp. 593–594], we provide an algorithm that (i) produces a large diversity of high-performing, yet qualitatively different molecules, (ii) illuminates the distribution of optimal solutions, and (iii) improves search efficiency compared to both machine learning and traditional genetic algorithm approaches.

We report a novel algorithm that produces a large diversity of high-performing molecules, illuminates the distribution of optimal solutions, and improves search efficiency compared to both machine learning and genetic algorithm approaches.     

Deep generative design with 3D pharmacophoric constraints

Generative models have increasingly been proposed as a solution to the molecular design problem. However, it has proved challenging to control the design process or incorporate prior knowledge, limiting their practical use in drug discovery. In particular, generative methods have made limited use of three-dimensional (3D) structural information even though this is critical to binding. This work describes a method to incorporate such information and demonstrates the benefit of doing so. We combine an existing graph-based deep generative model, DeLinker, with a convolutional neural network to utilise physically-meaningful 3D representations of molecules and target pharmacophores. We apply our model, DEVELOP, to both linker and R-group design, demonstrating its suitability for both hit-to-lead and lead optimisation. The 3D pharmacophoric information results in improved generation and allows greater control of the design process. In multiple large-scale evaluations, we show that including 3D pharmacophoric constraints results in substantial improvements in the quality of generated molecules. On a challenging test set derived from PDBbind, our model improves the proportion of generated molecules with high 3D similarity to the original molecule by over 300%. In addition, DEVELOP recovers 10× more of the original molecules compared to the baseline DeLinker method. Our approach is general-purpose, readily modifiable to alternate 3D representations, and can be incorporated into other generative frameworks. Code is available at https://github.com/oxpig/DEVELOP.

A novel deep generative model combines convolution and graph neural networks to allow 3D-aware molecular design. We show how 3D pharmacophoric information can be incorporated into generative models and apply our model to both linker and R-group design.     

Evaluation of fully-functionalized diazirine tags for chemical proteomic applications

The use of photo-affinity reagents for the mapping of noncovalent small molecule–protein interactions has become widespread. Recently, several ‘fully-functionalized’ (FF) chemical tags have been developed wherein a photoactivatable capture group, an enrichment handle, and a functional group for synthetic conjugation to a molecule of interest are integrated into a single modular tag. Diazirine-based FF tags in particular are increasingly employed in chemical proteomic investigations; however, despite routine usage, their relative utility has not been established. Here, we systematically evaluate several diazirine-containing FF tags, including a terminal diazirine analog developed herein, for chemical proteomic investigations. Specifically, we compared the general reactivity of five diazirine tags and assessed their impact on the profiles of various small molecules, including fragments and known inhibitors revealing that such tags can have profound effects on the proteomic profiles of chemical probes. Our findings should be informative for chemical probe design, photo-affinity reagent development, and chemical proteomic investigations.

The chemical proteomic properties of five diazirine-based, fully-functionalized photoaffinity tags, including a newly developed, minimal tag, were compared. This study provides guidance for the development of new photoaffinity probes.