Artificial Intelligence for Retrosynthetic Planning Needs Both Data and Expert Knowledge

Rapid advancements in artificial intelligence (AI) have enabled breakthroughs across many scientific disciplines. In organic chemistry, the challenge of planning complex multistep chemical syntheses should conceptually be well-suited for AI. Yet, the development of AI synthesis planners trained solely on reaction-example-data has stagnated and is not on par with the performance of "hybrid" algorithms combining AI with expert knowledge. This Perspective examines possible causes of these shortcomings, extending beyond the established reasoning of insufficient quantities of reaction data. Drawing attention to the intricacies and data biases that are specific to the domain of synthetic chemistry, we advocate augmenting the unique capabilities of AI with the knowledge base and the reasoning strategies of domain experts. By actively involving synthetic chemists, who are the end users of any synthesis planning software, into the development process, we envision to bridge the gap between computer algorithms and the intricate nature of chemical synthesis.

DFRscore: Deep Learning-Based Scoring of Synthetic Complexity with Drug-Focused Retrosynthetic Analysis for High-Throughput Virtual Screening

Recently emerging generative AI models enable us to produce a vast number of compounds for potential applications. While they can provide novel molecular structures, the synthetic feasibility of the generated molecules is often questioned. To address this issue, a few recent studies have attempted to use deep learning models to estimate the synthetic accessibility of many molecules rapidly. However, retrosynthetic analysis tools used to train the models rely on reaction templates automatically extracted from a large reaction database that are not domain-specific and may exhibit low chemical correctness. To overcome this limitation, we introduce DFRscore (Drug-Focused Retrosynthetic score), a deep learning-based approach for a more practical assessment of synthetic accessibility in drug discovery. The DFRscore model is trained exclusively on drug-focused reactions, providing a predicted number of minimally required synthetic steps for each compound. This approach enables practitioners to filter out compounds that do not meet their desired level of synthetic accessibility at an early stage of high-throughput virtual screening for accelerated drug discovery. The proposed strategy can be easily adapted to other domains by adjusting the synthesis planning setup of the reaction templates and starting materials.

Accelerated chemical science with AI

In light of the pressing need for practical materials and molecular solutions to renewable energy and health problems, to name just two examples, one wonders how to accelerate research and development in the chemical sciences, so as to address the time it takes to bring materials from initial discovery to commercialization. Artificial intelligence (AI)-based techniques, in particular, are having a transformative and accelerating impact on many if not most, technological domains. To shed light on these questions, the authors and participants gathered in person for the ASLLA Symposium on the theme of 'Accelerated Chemical Science with AI' at Gangneung, Republic of Korea. We present the findings, ideas, comments, and often contentious opinions expressed during four panel discussions related to the respective general topics: 'Data', 'New applications', 'Machine learning algorithms', and 'Education'. All discussions were recorded, transcribed into text using Open AI's Whisper, and summarized using LG AI Research's EXAONE LLM, followed by revision by all authors. For the broader benefit of current researchers, educators in higher education, and academic bodies such as associations, publishers, librarians, and companies, we provide chemistry-specific recommendations and summarize the resulting conclusions.

Computational prediction of complex cationic rearrangement outcomes

Recent years have seen revived interest in computer-assisted organic synthesis1,2. The use of reaction- and neural-network algorithms that can plan multistep synthetic pathways have revolutionized this field1,3-7, including examples leading to advanced natural products6,7. Such methods typically operate on full, literature-derived 'substrate(s)-to-product' reaction rules and cannot be easily extended to the analysis of reaction mechanisms. Here we show that computers equipped with a comprehensive knowledge-base of mechanistic steps augmented by physical-organic chemistry rules, as well as quantum mechanical and kinetic calculations, can use a reaction-network approach to analyse the mechanisms of some of the most complex organic transformations: namely, cationic rearrangements. Such rearrangements are a cornerstone of organic chemistry textbooks and entail notable changes in the molecule's carbon skeleton8-12. The algorithm we describe and deploy at https://HopCat.allchemy.net/ generates, within minutes, networks of possible mechanistic steps, traces plausible step sequences and calculates expected product distributions. We validate this algorithm by three sets of experiments whose analysis would probably prove challenging even to highly trained chemists: (1) predicting the outcomes of tail-to-head terpene (THT) cyclizations in which substantially different outcomes are encoded in modular precursors differing in minute structural details; (2) comparing the outcome of THT cyclizations in solution or in a supramolecular capsule; and (3) analysing complex reaction mixtures. Our results support a vision in which computers no longer just manipulate known reaction types1-7 but will help rationalize and discover new, mechanistically complex transformations.

Computer-designed repurposing of chemical wastes into drugs

As the chemical industry continues to produce considerable quantities of waste chemicals^1,2, it is essential to devise 'circular chemistry'^3-8 schemes to productively back-convert at least a portion of these unwanted materials into useful products. Despite substantial progress in the degradation of some classes of harmful chemicals⁹, work on 'closing the circle'-transforming waste substrates into valuable products-remains fragmented and focused on well known areas^10-15. Comprehensive analyses of which valuable products are synthesizable from diverse chemical wastes are difficult because even small sets of waste substrates can, within few steps, generate millions of putative products, each synthesizable by multiple routes forming densely connected networks. Tracing all such syntheses and selecting those that also meet criteria of process and 'green' chemistries is, arguably, beyond the cognition of human chemists. Here we show how computers equipped with broad synthetic knowledge can help address this challenge. Using the forward-synthesis Allchemy platform¹⁶, we generate giant synthetic networks emanating from approximately 200 waste chemicals recycled on commercial scales, retrieve from these networks tens of thousands of routes leading to approximately 300 important drugs and agrochemicals, and algorithmically rank these syntheses according to the accepted metrics of sustainable chemistry^17-19. Several of these routes we validate by experiment, including an industrially realistic demonstration on a 'pharmacy on demand' flow-chemistry platform²⁰. Wide adoption of computerized waste-to-valuable algorithms can accelerate productive reuse of chemicals that would otherwise incur storage or disposal costs, or even pose environmental hazards.

Is Organic Chemistry Really Growing Exponentially?

In terms of molecules and specific reaction examples, organic chemistry features an impressive, exponential growth. However, new reaction classes/types that fuel this growth are being discovered at a much slower and only linear (or even sublinear) rate. The proportion of newly discovered reaction types to all reactions being performed keeps decreasing, suggesting that synthetic chemistry becomes more reliant on reusing the well-known methods. The newly discovered chemistries are more complex than decades ago and allow for the rapid construction of complex scaffolds in fewer numbers of steps. We study these and other trends in the function of time, reaction-type popularity and complexity based on the algorithm that extracts generalized reaction class templates. These analyses are useful in the context of computer-assisted synthesis, machine learning (to estimate the numbers of models with sufficient reaction statistics), and identifying erroneous entries in reaction databases.

Molecule Edit Graph Attention Network: Modeling Chemical Reactions as Sequences of Graph Edits

The central challenge in automated synthesis planning is to be able to generate and predict outcomes of a diverse set of chemical reactions. In particular, in many cases, the most likely synthesis pathway cannot be applied due to additional constraints, which requires proposing alternative chemical reactions. With this in mind, we present Molecule Edit Graph Attention Network (MEGAN), an end-to-end encoder-decoder neural model. MEGAN is inspired by models that express a chemical reaction as a sequence of graph edits, akin to the arrow pushing formalism. We extend this model to retrosynthesis prediction (predicting substrates given the product of a chemical reaction) and scale it up to large data sets. We argue that representing the reaction as a sequence of edits enables MEGAN to efficiently explore the space of plausible chemical reactions, maintaining the flexibility of modeling the reaction in an end-to-end fashion and achieving state-of-the-art accuracy in standard benchmarks. Code and trained models are made available online at https://github.com/molecule-one/megan.

Chemist Ex Machina: Advanced Synthesis Planning by Computers

Teaching computers to plan multistep syntheses of arbitrary target molecules-including natural products-has been one of the oldest challenges in chemistry, dating back to the 1960s. This Account recapitulates two decades of our group's work on the software platform called Chematica, which very recently achieved this long-sought objective and has been shown capable of planning synthetic routes to complex natural products, several of which were validated in the laboratory.For the machine to plan syntheses at an expert level, it must know the rules describing chemical reactions and use these rules to expand and search the networks of synthetic options. The rules must be of high quality: They must delineate accurately the scope of admissible substituents, capture all relevant stereochemical information, detect potential reactivity conflicts, and protection requirements. They should yield only those synthons that are chemically stable and energetically allowed (e.g., not too strained) and should be able to extrapolate beyond examples already published in the literature. In parallel, the network-search algorithms must be able to assign meaningful scores to the sets of synthons they encounter, make judicious choices which of the network's branches to expand, and when to withdraw from unpromising ones. They must be able to strategize over multiple steps to resolve intermittent reactivity conflicts, exchange functional groups, or overcome local maxima of molecular complexity.Meeting all these requirements makes the problem of computer-driven retrosynthesis very multifaceted, combining expert and AI approaches further supplemented by quantum-mechanical and molecular-mechanics calculations. Development of Chematica has been a very long and gradual process because all these components are needed. Any shortcuts-for example, reliance on only expert or only data-based approaches-yield chemically naïve and often erroneous syntheses, especially for complex targets. On the bright side, once all the requisite algorithms are implemented-as they now are-they not only streamline conventional synthetic planning but also enable completely new modalities that would challenge any human chemist, for example, synthesis with multiple constraints imposed simultaneously or library-wide syntheses in which the machine constructs "global plans" leading to multiple targets and benefiting from the use of common intermediates. These types of analyses will have profound impact on the practice of chemical industry, designing more economical, more green, and less hazardous pathways.

RetroBioCat as a computer-aided synthesis planning tool for biocatalytic reactions and cascades

As the enzyme toolbox for biocatalysis has expanded, so has the potential for the construction of powerful enzymatic cascades for efficient and selective synthesis of target molecules. Additionally, recent advances in computer-aided synthesis planning are revolutionising synthesis design in both synthetic biology and organic chemistry. However, the potential for biocatalysis is not well captured by tools currently available in either field. Here we present RetroBioCat, an intuitive and accessible tool for computer-aided design of biocatalytic cascades, freely available at retrobiocat.com. Our approach uses a set of expertly encoded reaction rules encompassing the enzyme toolbox for biocatalysis, and a system for identifying literature precedent for enzymes with the correct substrate specificity where this is available. Applying these rules for automated biocatalytic retrosynthesis, we show our tool to be capable of identifying promising biocatalytic pathways to target molecules, validated using a test-set of recent cascades described in the literature.

Artificial intelligence and automation in computer aided synthesis planning

In this perspective we deal with questions pertaining to the development of synthesis planning technologies over the course of recent years.

Computational planning of the synthesis of complex natural products

Training algorithms to computationally plan multistep organic syntheses has been a challenge for more than 50 years^1-7. However, the field has progressed greatly since the development of early programs such as LHASA^1,7, for which reaction choices at each step were made by human operators. Multiple software platforms^6,8-14 are now capable of completely autonomous planning. But these programs 'think' only one step at a time and have so far been limited to relatively simple targets, the syntheses of which could arguably be designed by human chemists within minutes, without the help of a computer. Furthermore, no algorithm has yet been able to design plausible routes to complex natural products, for which much more far-sighted, multistep planning is necessary^15,16 and closely related literature precedents cannot be relied on. Here we demonstrate that such computational synthesis planning is possible, provided that the program's knowledge of organic chemistry and data-based artificial intelligence routines are augmented with causal relationships^17,18, allowing it to 'strategize' over multiple synthetic steps. Using a Turing-like test administered to synthesis experts, we show that the routes designed by such a program are largely indistinguishable from those designed by humans. We also successfully validated three computer-designed syntheses of natural products in the laboratory. Taken together, these results indicate that expert-level automated synthetic planning is feasible, pending continued improvements to the reaction knowledge base and further code optimization.

Synthetic connectivity, emergence, and self-regeneration in the network of prebiotic chemistry

The challenge of prebiotic chemistry is to trace the syntheses of life’s key building blocks from a handful of primordial substrates. Here we report a forward-synthesis algorithm that generates a full network of prebiotic chemical reactions accessible from these substrates under generally accepted conditions. This network contains both reported and previously unidentified routes to biotic targets, as well as plausible syntheses of abiotic molecules. It also exhibits three forms of nontrivial chemical emergence, as the molecules within the network can act as catalysts of downstream reaction types; form functional chemical systems, including self-regenerating cycles; and produce surfactants relevant to primitive forms of biological compartmentalization. To support these claims, computer-predicted, prebiotic syntheses of several biotic molecules as well as a multistep, self-regenerative cycle of iminodiacetic acid were validated by experiment.

Computer-generated "synthetic contingency" plans at times of logistics and supply problems: scenarios for hydroxychloroquine and remdesivir

A computer program for retrosynthetic planning helps develop multiple "synthetic contingency" plans for hydroxychloroquine and also routes leading to remdesivir, both promising but yet unproven medications against COVID-19. These plans are designed to navigate, as much as possible, around known and patented routes and to commence from inexpensive and diverse starting materials, so as to ensure supply in case of anticipated market shortages of commonly used substrates. Looking beyond the current COVID-19 pandemic, development of similar contingency syntheses is advocated for other already-approved medications, in case such medications become urgently needed in mass quantities to face other public-health emergencies.

Data Augmentation and Pretraining for Template-Based Retrosynthetic Prediction in Computer-Aided Synthesis Planning

This work presents efforts to augment the performance of data-driven machine learning algorithms for reaction template recommendation used in computer-aided synthesis planning software. Often, machine learning models designed to perform the task of prioritizing reaction templates or molecular transformations are focused on reporting high-accuracy metrics for the one-to-one mapping of product molecules in reaction databases to the template extracted from the recorded reaction. The available templates that get selected for inclusion in these machine learning models have been previously limited to those that appear frequently in the reaction databases and exclude potentially useful transformations. By augmenting open-access data sets of organic reactions with explicitly calculated template applicability and pretraining a template-relevance neural network on this augmented applicability data set, we report an increase in the template applicability recall and an increase in the diversity of predicted precursors. The augmentation and pretraining effectively teaches the neural network an increased set of templates that could theoretically lead to successful reactions for a given target. Even on a small data set of well-curated reactions, the data augmentation and pretraining methods resulted in an increase in top-1 accuracy, especially for rare templates, indicating that these strategies can be very useful for small data sets.

A tutored discourse on microcontrollers, single board computers and their applications to monitor and control chemical reactions

This Tutored Discourse constitutes a preliminary exposure on how synthesis chemists can engage positively with inexpensive, low-power microcontrollers to aid control, monitoring and optimisation of chemical reactions.

Multitask prediction of site selectivity in aromatic C–H functionalization reactions

Aromatic C–H functionalization reactions are an important part of the synthetic chemistry toolbox.