Toward Machine Learning-Enhanced High-Throughput Experimentation

Data processing for high-throughput mass spectrometry in drug discovery

INTRODUCTION

High-throughput mass spectrometry that could deliver > 10 times faster sample readout speed than traditional LC-based platforms has emerged as a powerful analytical technique, enabling the rapid analysis of complex biological samples. This increased speed of MS data acquisition has brought a critical demand for automatic data processing capabilities that should match or surpass the speed of data acquisition. Those data processing capabilities should serve the different requirements of drug discovery workflows.

AREAS COVERED

This paper introduced the key steps of the automatic data processing workflows for high-throughput MS technologies. Specific examples and requirements are detailed for different drug discovery applications.

EXPERT OPINION

The demand for automatic data processing in high-throughput mass spectrometry is driven by the need to keep pace with the accelerated speed of data acquisition. The seamless integration of processing capabilities with LIMS, efficient data review mechanisms, and the exploration of future features such as real-time feedback, automatic method optimization, and AI model training is crucial for advancing the drug discovery field. As technology continues to evolve, the synergy between high-throughput mass spectrometry and intelligent data processing will undoubtedly play a pivotal role in shaping the future of high-throughput drug discovery applications.

Design and Validation of a High-Throughput Reductive Catalytic Fractionation Method

Reductive catalytic fractionation (RCF) is a promising method to extract and depolymerize lignin from biomass, and bench-scale studies have enabled considerable progress in the past decade. RCF experiments are typically conducted in pressurized batch reactors with volumes ranging between 50 and 1000 mL, limiting the throughput of these experiments to one to six reactions per day for an individual researcher. Here, we report a high-throughput RCF (HTP-RCF) method in which batch RCF reactions are conducted in 1 mL wells machined directly into Hastelloy reactor plates. The plate reactors can seal high pressures produced by organic solvents by vertically stacking multiple reactor plates, leading to a compact and modular system capable of performing 240 reactions per experiment. Using this setup, we screened solvent mixtures and catalyst loadings for hydrogen-free RCF using 50 mg poplar and 0.5 mL reaction solvent. The system of 1:1 isopropanol/methanol showed optimal monomer yields and selectivity to 4-propyl substituted monomers, and validation reactions using 75 mL batch reactors produced identical monomer yields. To accommodate the low material loadings, we then developed a workup procedure for parallel filtration, washing, and drying of samples and a 1H nuclear magnetic resonance spectroscopy method to measure the RCF oil yield without performing liquid-liquid extraction. As a demonstration of this experimental pipeline, 50 unique switchgrass samples were screened in RCF reactions in the HTP-RCF system, revealing a wide range of monomer yields (21-36%), S/G ratios (0.41-0.93), and oil yields (40-75%). These results were successfully validated by repeating RCF reactions in 75 mL batch reactors for a subset of samples. We anticipate that this approach can be used to rapidly screen substrates, catalysts, and reaction conditions in high-pressure batch reactions with higher throughput than standard batch reactors.

Data-oriented protein kinase drug discovery

The continued growth of data from biological screening and medicinal chemistry provides opportunities for data-driven experimental design and decision making in early-phase drug discovery. Approaches adopted from data science help to integrate internal and public domain data and extract knowledge from historical in-house data. Protein kinase (PK) drug discovery is an exemplary area where large amounts of data are accumulating, providing a valuable knowledge base for discovery projects. Herein, the evolution of PK drug discovery and development of small molecular PK inhibitors (PKIs) is reviewed, highlighting milestone developments in the field and discussing exemplary studies providing a basis for increasing data orientation of PK discovery efforts.

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.

When Yield Prediction Does Not Yield Prediction: An Overview of the Current Challenges

Machine Learning (ML) techniques face significant challenges when predicting advanced chemical properties, such as yield, feasibility of chemical synthesis, and optimal reaction conditions. These challenges stem from the high-dimensional nature of the prediction task and the myriad essential variables involved, ranging from reactants and reagents to catalysts, temperature, and purification processes. Successfully developing a reliable predictive model not only holds the potential for optimizing high-throughput experiments but can also elevate existing retrosynthetic predictive approaches and bolster a plethora of applications within the field. In this review, we systematically evaluate the efficacy of current ML methodologies in chemoinformatics, shedding light on their milestones and inherent limitations. Additionally, a detailed examination of a representative case study provides insights into the prevailing issues related to data availability and transferability in the discipline.

Continuous flow synthesis of pyridinium salts accelerated by multi-objective Bayesian optimization with active learning

We report a human-in-the-loop implementation of the multi-objective experimental design via a Bayesian optimization platform (EDBO+) towards the optimization of butylpyridinium bromide synthesis under continuous flow conditions. The algorithm simultaneously optimized reaction yield and production rate (or space-time yield) and generated a well defined Pareto front. The versatility of EDBO+ was demonstrated by expanding the reaction space mid-campaign by increasing the upper temperature limit. Incorporation of continuous flow techniques enabled improved control over reaction parameters compared to common batch chemistry processes, while providing a route towards future automated syntheses and improved scalability. To that end, we applied the open-source Python module, nmrglue, for semi-automated nuclear magnetic resonance (NMR) spectroscopy analysis, and compared the acquired outputs against those obtained through manual processing methods from spectra collected on both low-field (60 MHz) and high-field (400 MHz) NMR spectrometers. The EDBO+ based model was retrained with these four different datasets and the resulting Pareto front predictions provided insight into the effect of data analysis on model predictions. Finally, quaternization of poly(4-vinylpyridine) with bromobutane illustrated the extension of continuous flow chemistry to synthesize functional materials.

Accelerated high-throughput imaging and phenotyping system for small organisms

Studying the complex web of interactions in biological communities requires large multifactorial experiments with sufficient statistical power. Automation tools reduce the time and labor associated with setup, data collection, and analysis in experiments that untangle these webs. We developed tools for high-throughput experimentation (HTE) in duckweeds, small aquatic plants that are amenable to autonomous experimental preparation and image-based phenotyping. We showcase the abilities of our HTE system in a study with 6,000 experimental units grown across 2,000 treatments. These automated tools facilitated the collection and analysis of time-resolved growth data, which revealed finer dynamics of plant-microbe interactions across environmental gradients. Altogether, our HTE system can run experiments with up to 11,520 experimental units and can be adapted for other small organisms.

Late-stage Functionalization for Improving Drug-like Molecular Properties

The development of late-stage functionalization (LSF) methodologies, particularly C-H functionalization, has revolutionized the field of organic synthesis. Over the past decade, medicinal chemists have begun to implement LSF strategies into their drug discovery programs, allowing for the drug discovery process to become more efficient. Most reported applications of late-stage C-H functionalization of drugs and drug-like molecules have been to rapidly diversify screening libraries to explore structure-activity relationships. However, there has been a growing trend toward the use of LSF methodologies as an efficient tool for improving drug-like molecular properties of promising drug candidates. In this review, we have comprehensively reviewed recent progress in this emerging area. Particular emphasis is placed on case studies where multiple LSF techniques were implemented to generate a library of novel analogues with improved drug-like properties. We have critically analyzed the current scope of LSF strategies to improve drug-like properties and commented on how we believe LSF can transform drug discovery in the future. Overall, we aim to provide a comprehensive survey of LSF techniques as tools for efficiently improving drug-like molecular properties, anticipating its continued uptake in drug discovery programs.

A field guide to flow chemistry for synthetic organic chemists

Flow chemistry has unlocked a world of possibilities for the synthetic community, but the idea that it is a mysterious "black box" needs to go. In this review, we show that several of the benefits of microreactor technology can be exploited to push the boundaries in organic synthesis and to unleash unique reactivity and selectivity. By "lifting the veil" on some of the governing principles behind the observed trends, we hope that this review will serve as a useful field guide for those interested in diving into flow chemistry.

AROPS: A Framework of Automated Reaction Optimization with Parallelized Scheduling

With the development of automated experimental platforms and optimization algorithms, chemists can easily optimize chemical reactions in an automated and high-throughput fashion. However, the modules in existing automated experimental platforms are operated in a linear fashion without orchestrating with the optimization algorithm, thus leaving room for further efficiency improvement. Here, we introduced a framework of automated reaction optimization with parallelized scheduling (AROPS) to realize the integration of the optimization algorithm and module scheduling. AROPS relies on a customized Bayesian optimizer to solve multi-reactor/analyzer reaction optimization problems with three different scheduling modes to arrange tasks for various experimental modules. In addition, a mechanism based on probability of improvement (PI) for discarding unpromising ongoing experiments was developed to facilitate freeing up valuable experimental resources in parallelized optimization. We tested the performance of AROPS using a hardware emulator on three representative benchmark reactions encountered in organic synthesis, illustrating that AROPS can trade off optimization time and cost according to the chemists' preference.

Electrochemical Mechanistic Analysis from Cyclic Voltammograms Based on Deep Learning

For decades, employing cyclic voltammetry for mechanistic investigation has demanded manual inspection of voltammograms. Here, we report a deep-learning-based algorithm that automatically analyzes cyclic voltammograms and designates a probable electrochemical mechanism among five of the most common ones in homogeneous molecular electrochemistry. The reported algorithm will aid researchers' mechanistic analyses, utilize otherwise elusive features in voltammograms, and experimentally observe the gradual mechanism transitions encountered in electrochemistry. An automated voltammogram analysis will aid the analysis of complex electrochemical systems and promise autonomous high-throughput research in electrochemistry with minimal human interference.

Accelerated High-throughput Plant Imaging and Phenotyping System.. [DOI: 10.1101/2022.09.28.509964] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

The complex web of interactions in biological communities is an area of study that requires large multifactorial experiments with sufficient statistical power. The use of automated tools can reduce the time and labor associated with experiment setup, data collection, and analysis in experiments aimed at untangling these webs. Here we demonstrate tools for high-throughput experimentation (HTE) in duckweeds, small aquatic plants that are amenable to autonomous experimental preparation and image-based phenotyping. We showcase the abilities of our HTE system in a study with 6,000 experimental units grown across 1,000 different nutrient environments. The use of our automated tools facilitated the collection and analysis of time-resolved growth data, which revealed finer dynamics of plant-microbe interactions across environmental gradients. Altogether, our HTE system can run experiments of up to 11,520 experimental units and can be adapted to studies with other small organisms.

Machine learning, artificial intelligence, and chemistry: how smart algorithms are reshaping simulation and the laboratory

Machine learning and artificial intelligence are increasingly gaining in prominence through image analysis, language processing, and automation, to name a few applications. Machine learning is also making profound changes in chemistry. From revisiting decades-old analytical techniques for the purpose of creating better calibration curves, to assisting and accelerating traditional in silico simulations, to automating entire scientific workflows, to being used as an approach to deduce underlying physics of unexplained chemical phenomena, machine learning and artificial intelligence are reshaping chemistry, accelerating scientific discovery, and yielding new insights. This review provides an overview of machine learning and artificial intelligence from a chemist’s perspective and focuses on a number of examples of the use of these approaches in computational chemistry and in the laboratory.

Autonomous optimization of non-aqueous Li-ion battery electrolytes via robotic experimentation and machine learning coupling

Developing high-energy and efficient battery technologies is a crucial aspect of advancing the electrification of transportation and aviation. However, battery innovations can take years to deliver. In the case of non-aqueous battery electrolyte solutions, the many design variables in selecting multiple solvents, salts and their relative ratios make electrolyte optimization time-consuming and laborious. To overcome these issues, we propose in this work an experimental design that couples robotics (a custom-built automated experiment named "Clio") to machine-learning (a Bayesian optimization-based experiment planner named "Dragonfly"). An autonomous optimization of the electrolyte conductivity over a single-salt and ternary solvent design space identifies six fast-charging non-aqueous electrolyte solutions in two work-days and forty-two experiments. This result represents a six-fold time acceleration compared to a random search performed by the same automated experiment. To validate the practical use of these electrolytes, we tested them in a 220 mAh graphite∣∣LiNi_0.5Mn_0.3Co_0.2O₂ pouch cell configuration. All the pouch cells containing the robot-developed electrolytes demonstrate improved fast-charging capability against a baseline experiment that uses a non-aqueous electrolyte solution selected a priori from the design space.

Mechanisms, Challenges, and Opportunities of Dual Ni/Photoredox-Catalyzed C(sp2)-C(sp3) Cross-Couplings

The merging of photoredox and nickel catalysis has revolutionized the field of C-C cross-coupling. However, in comparison to the development of synthetic methods, detailed mechanistic investigations of these catalytic systems are lagging. To improve the mechanistic understanding, computational tools have emerged as powerful tools to elucidate the factors controlling reactivity and selectivity in these complex catalytic transformations. Based on the reported computational studies, it appears that the mechanistic picture of catalytic systems is not generally applicable, but is rather dependent on the specific choice of substrate, ligands, photocatalysts, etc. Given the complexity of these systems, the need for more accurate computational methods, readily available and user-friendly dynamics simulation tools, and data-driven approaches is clear in order to understand at the molecular level the mechanisms of these transformations. In particular, we anticipate that such improvement of theoretical methods will become crucial to advance the understanding of excited-state properties and dynamics of key species, as well as to enable faster and unbiased exploration of reaction pathways. Further, with greater collaboration between computational, experimental, and spectroscopic communities, the mechanistic investigation of photoredox/Ni dual-catalytic reactions is expected to thrive quickly, facilitating the design of novel catalytic systems and promoting our understanding of the reaction selectivity.

Working at the interfaces of data science and synthetic electrochemistry

Electrochemistry is quickly entering the mainstream of synthetic organic chemistry. The diversity of new transformations enabled by electrochemistry is to a large extent a consequence of the unique features and reaction parameters in electrochemical systems including redox mediators, applied potential, electrode material, and cell construction. While offering chemists new means to control reactivity and selectivity, these additional features also increase the dimensionalities of a reaction system and complicate its optimization. This challenge, however, has spawned increasing adoption of data science tools to aid reaction discovery as well as development of high-throughput screening platforms that facilitate the generation of high quality datasets. In this Perspective, we provide an overview of recent advances in data-science driven electrochemistry with an emphasis on the opportunities and challenges facing this growing subdiscipline.

Machine Learning-Assisted Design of Material Properties

Designing functional materials requires a deep search through multidimensional spaces for system parameters that yield desirable material properties. For cases where conventional parameter sweeps or trial-and-error sampling are impractical, inverse methods that frame design as a constrained optimization problem present an attractive alternative. However, even efficient algorithms require time- and resource-intensive characterization of material properties many times during optimization, imposing a design bottleneck. Approaches that incorporate machine learning can help address this limitation and accelerate the discovery of materials with targeted properties. In this article, we review how to leverage machine learning to reduce dimensionality in order to effectively explore design space, accelerate property evaluation, and generate unconventional material structures with optimal properties. We also discuss promising future directions, including integration of machine learning into multiple stages of a design algorithm and interpretation of machine learning models to understand how design parameters relate to material properties. Expected final online publication date for the Annual Review of Chemical and Biomolecular Engineering, Volume 13 is October 2022. Please see http://www.annualreviews.org/page/journal/pubdates for revised estimates.

Technological Innovations in Photochemistry for Organic Synthesis: Flow Chemistry, High-Throughput Experimentation, Scale-up, and Photoelectrochemistry

Photoinduced chemical transformations have received in recent years a tremendous amount of attention, providing a plethora of opportunities to synthetic organic chemists. However, performing a photochemical transformation can be quite a challenge because of various issues related to the delivery of photons. These challenges have barred the widespread adoption of photochemical steps in the chemical industry. However, in the past decade, several technological innovations have led to more reproducible, selective, and scalable photoinduced reactions. Herein, we provide a comprehensive overview of these exciting technological advances, including flow chemistry, high-throughput experimentation, reactor design and scale-up, and the combination of photo- and electro-chemistry.

High-throughput parallelized testing of membrane electrode assemblies for CO2 reduction

High-throughput characterization of electrochemical reactions can accelerate discovery and optimization cycles, and provide the data required for further acceleration via machine-learning guided experiment planning. There are a range of high...

Prospects and challenges for autonomous catalyst discovery viewed from an experimental perspective

Autonomous catalysis research requires elaborate integration of operando experiments into automated workflows. Suitable experimental data for analysis by artificial intelligence can be measured more readily according to standard operating procedures.

Accelerate Synthesis of Metal-Organic Frameworks by a Robotic Platform and Bayesian Optimization

Synthesis of materials with desired structures, e.g., metal-organic frameworks (MOFs), involves optimization of highly complex chemical and reaction spaces due to multiple choices of chemical elements and reaction parameters/routes. Traditionally, realizing such an aim requires rapid screening of these nonlinear spaces by experimental conduction with human intuition, which is quite inefficient and may cause errors or bias. In this work, we report a platform that integrates a synthesis robot with the Bayesian optimization (BO) algorithm to accelerate the synthesis of MOFs. This robotic platform consists of a direct laser writing apparatus, precursor injecting and Joule-heating components. It can automate the MOFs synthesis upon fed reaction parameters that are recommended by the BO algorithm. Without any prior knowledge, this integrated platform continuously improves the crystallinity of ZIF-67, a demo MOF employed in this study, as the number of operation iterations increases. This work represents a methodology enabled by a data-driven synthesis robot, which achieves the goal of material synthesis with targeted structures, thus greatly shortening the reaction time and reducing energy consumption. It can be easily generalized to other material systems, thus paving a new route to the autonomous discovery of a variety of materials in a cost-effective way in the future.

Toward autonomous design and synthesis of novel inorganic materials

Autonomous experimentation driven by artificial intelligence (AI) provides an exciting opportunity to revolutionize inorganic materials discovery and development. Herein, we review recent progress in the design of self-driving laboratories, including robotics to automate materials synthesis and characterization, in conjunction with AI to interpret experimental outcomes and propose new experimental procedures. We focus on efforts to automate inorganic synthesis through solution-based routes, solid-state reactions, and thin film deposition. In each case, connections are made to relevant work in organic chemistry, where automation is more common. Characterization techniques are primarily discussed in the context of phase identification, as this task is critical to understand what products have formed during synthesis. The application of deep learning to analyze multivariate characterization data and perform phase identification is examined. To achieve "closed-loop" materials synthesis and design, we further provide a detailed overview of optimization algorithms that use active learning to rationally guide experimental iterations. Finally, we highlight several key opportunities and challenges for the future development of self-driving inorganic materials synthesis platforms.

Late-stage C–H functionalization offers new opportunities in drug discovery

Over the past decade, the landscape of molecular synthesis has gained major impetus by the introduction of late-stage functionalization (LSF) methodologies. C-H functionalization approaches, particularly, set the stage for new retrosynthetic disconnections, while leading to improvements in resource economy. A variety of innovative techniques have been successfully applied to the C-H diversification of pharmaceuticals, and these key developments have enabled medicinal chemists to integrate LSF strategies in their drug discovery programmes. This Review highlights the significant advances achieved in the late-stage C-H functionalization of drugs and drug-like compounds, and showcases how the implementation of these modern strategies allows increased efficiency in the drug discovery process. Representative examples are examined and classified by mechanistic patterns involving directed or innate C-H functionalization, as well as emerging reaction manifolds, such as electrosynthesis and biocatalysis, among others. Structurally complex bioactive entities beyond small molecules are also covered, including diversification in the new modalities sphere. The challenges and limitations of current LSF methods are critically assessed, and avenues for future improvements of this rapidly expanding field are discussed. We, hereby, aim to provide a toolbox for chemists in academia as well as industrial practitioners, and introduce guiding principles for the application of LSF strategies to access new molecules of interest.

Toward machine learning-enhanced high-throughput experimentation for chemistry

High-throughput experimentation in chemistry allows for quick and automated exploration of chemical space to, for example, discover new drugs. Combining machine learning techniques with high-throughput experimentation has the potential to speed up and improve chemical space exploration and optimization.

Continuous biphasic chemical processes in a four-phase segmented flow reactor

A four-phase segmented flow regime for continuous biphasic reaction processes is introduced, characterized over 1500 automatically conducted experiments, and used for biphasic ligand exchange of CdSe quantum dots.