Reaction Templates: Bridging Synthesis Knowledge and Artificial Intelligence.

ConspectusThe field of chemical research boasts a long history of developing software to automate synthesis planning and reaction prediction. Early software relied heavily on expert systems, requiring significant effort to encode vast amounts of synthesis knowledge into a computer-readable format. However, recent advancements in deep learning have shifted the focus toward AI models, offering improved prediction capabilities. Despite these advancements, current AI models often lack the integration of known synthesis rules and intuitions, creating a gap that hinders interpretability and future development of the models. To bridge them, our research group has been actively working on incorporating reaction templates into deep learning models, achieving promising results across various applications.In this Account, we present our latest works to incorporate the known synthesis knowledge into the deep learning models through the utilization of reaction templates. We begin by highlighting the limitations of early computer programs heavily reliant on hand-coded rules. These programs, while providing a foundation for the field, presented limitations in scalability and adaptability. We then introduce SMARTS (SMILES arbitrary target specification), a popular Python-readable format for representing chemical reactions. This format of reaction encoding facilitates the quick integration of synthesis knowledge into AI models built using the Python language. With the SMARTS-based reaction templates, we introduce our recent efforts of developing an AI model for reaction-based molecule optimization. Subsequently, we discuss the recent efforts to automate the extraction of reaction templates from vast chemical reaction databases. This approach eliminates the previously required manual effort of encoding knowledge, a process that could be time-consuming and prone to error when dealing with large data sets. By customizing the automated extraction algorithm, we have developed powerful AI models for specific tasks such as retrosynthesis (LocalRetro), reaction outcome prediction (LocalTransform), and atom-to-atom mapping (LocalMapper). These models, aligned with the intuition of chemists, demonstrate the effectiveness of incorporating reaction templates into deep learning frameworks.Looking toward the future, we believe that utilizing reaction templates to connect known chemical knowledge and AI models holds immense potential for various applications. Not only can this approach significantly benefit future AI models focused on challenging tasks like reaction mechanism labeling and prediction, but we anticipate it can also extend its reach to the realm of inorganic synthesis. By integrating synthesis knowledge, we can not only achieve improved performance but also enhance the interpretability of AI models, paving the way for further advancements in AI-powered chemical synthesis.

PredPS: Attention-based graph neural network for predicting stability of compounds in human plasma.

Stability of compounds in the human plasma is crucial for maintaining sufficient systemic drug exposure and considered an essential factor in the early stages of drug discovery and development. The rapid degradation of compounds in the plasma can result in poor in vivo efficacy. Currently, there are no open-source software programs for predicting human plasma stability. In this study, we developed an attention-based graph neural network, PredPS to predict the plasma stability of compounds in human plasma using in-house and open-source datasets. The PredPS outperformed the two machine learning and two deep learning algorithms that were used for comparison indicating its stability-predicting efficiency. PredPS achieved an area under the receiver operating characteristic curve of 90.1%, accuracy of 83.5%, sensitivity of 82.3%, and specificity of 84.6% when evaluated using 5-fold cross-validation. In the early stages of drug discovery, PredPS could be a helpful method for predicting the human plasma stability of compounds. Saving time and money can be accomplished by adopting an in silico-based plasma stability prediction model at the high-throughput screening stage. The source code for PredPS is available at https://bitbucket.org/krict-ai/predps and the PredPS web server is available at https://predps.netlify.app.

PredAOT: a computational framework for prediction of acute oral toxicity based on multiple random forest models.

BACKGROUND: Acute oral toxicity of drug candidates can lead to drug development failure; thus, predicting the acute oral toxicity of small compounds is important for successful drug development. However, evaluation of the acute oral toxicity of small compounds considered in the early stages of drug discovery is limited because of cost and time. Here, we developed a computational framework, PredAOT, that predicts the acute oral toxicity of small compounds in mice and rats. METHODS: PredAOT is based on multiple random forest models for the accurate prediction of acute oral toxicity. A total of 6226 and 6238 compounds evaluated in mice and rats, respectively, were used to train the models. RESULTS: PredAOT has the advantage of predicting acute oral toxicity in mice and rats simultaneously, and its prediction performance is similar to or better than that of existing tools. CONCLUSION: PredAOT will be a useful tool for the quick and accurate prediction of the acute oral toxicity of small compounds in mice and rats during drug development.

Structure-Based Synthesizability Prediction of Crystals Using Partially Supervised Learning.

Predicting the synthesizability of inorganic materials is one of the major challenges in accelerated material discovery. A widely employed approximate approach is to consider the thermodynamic decomposition stability due to its simplicity of computing, but it is notorious for either producing too many candidates or missing important metastable materials. These results, however, are not unexcepted since the synthesizability is a complex phenomenon, and the thermodynamic stability is just one contributor. Here, we suggest a machine-learning model to quantify the probability of synthesis based on the partially supervised learning of materials database. We adapted the positive and unlabeled machine learning (PU learning) by implementing the graph convolutional neural network as a classifier in which the model outputs crystal-likeness scores (CLscore). The model shows 87.4% true positive (CLscore > 0.5) prediction accuracy for the test set of experimentally reported cases (9356 materials) in the Materials Project. We further validated the model by predicting the synthesizability of newly reported experimental materials in the last 5 years (2015-2019) with an 86.2% true positive rate using the model trained with the database as of the end of year 2014. Our analysis shows that our model captures the structural motif for synthesizability beyond what is possible by Ehull. We find that 71 materials among the top 100 high-scoring virtual materials have indeed been previously synthesized in the literature. With the proposed data-driven metric of the crystal-likeness score, high-throughput virtual screenings and generative models can benefit significantly by effectively reducing the chemical space that needs to be explored experimentally in the future toward more rational materials design.

One dimensional building blocks for molecular separation: laminated graphitic nanoribbons.

Herein, a new carbon-based graphitic membrane composed of laminated graphitic nanoribbons with a nanometer-scale width and micrometer-scale length, the graphitic nanoribbon membrane, is reported. Compared to the existing graphitic membranes, such as those composed of graphene oxide and carbon nanotubes, the developed membrane exhibits several unique characteristics in pressure-driven systems. First, the short diffusion length through its interlayer and the free volume of its stacked nanoribbons result in high solvent flux regardless of solvent polarity (water: 25-250 L m-2 h-1 bar-1; toluene: ∼975 L m-2 h-1 bar-1; hexane: ∼240 L m-2 h-1 bar-1). The flux value for water is one order of magnitude higher, while that for nonpolar organic solvents is two to three orders of magnitude greater than the corresponding flux values obtained through commercially available nanofiltration membranes. Second, the membrane exhibits good separation performance, particularly with organic dye molecules (∼100%) and trivalent ions (∼60%), maintaining high solvent flux during extended filtration. Finally, the membrane exhibits high stability in various fluids, e.g., 1 M HCl solution, 1 M NaOH solution, toluene, ethanol, and water, as well as under hydraulic pressures of up to 50 bar. Electron microscopy observation and simulation results suggest that such distinctive features of the membrane are related to the entangled thin multilayers of the graphitic nanoribbons, which possibly originate from the high aspect ratio and narrow width of the nanoribbons.