A Machine Learning Approach to Model Interaction Effects: Development and Application to Alcohol Deoxyfluorination.

The application of machine learning (ML) techniques to model high-throughput experimentation (HTE) datasets has seen a recent rise in popularity. Nevertheless, the ability to model the interplay between reaction components, known as interaction effects, with ML remains an outstanding challenge. Using a simulated HTE dataset, we find that the presence of irrelevant features poses a strong obstacle to learning interaction effects with common ML algorithms. To address this problem, we propose a two-part statistical modeling approach for HTE datasets: classical analysis of variance of the experiment to identify systematic effects that impact reaction yield across the experiment followed by regression of individual effects using chemistry-informed features. To illustrate this methodology, we use our previously published alcohol deoxyfluorination dataset comprising 740 reactions to build a compact, interpretable generalized additive model that accounts for each significant effect observed in the dataset. We achieve a sizeable performance boost compared to our previously published random forest model, reducing mean absolute error from 18 to 13% and root-mean-squared error from 22 to 17% on a newly generated validation set. Finally, we demonstrate that this approach can facilitate the generation of new mechanistic hypotheses, which, when probed experimentally, can lead to a deeper understanding of chemical reactivity.

Strategies for Nucleophilic C(sp3)-(Radio)Fluorination.

This Perspective surveys the progress and current limitations of nucleophilic fluorination methodologies. Despite the long and rich history of C(sp3)-F bond construction in chemical research, the inherent challenges associated with this transformation have largely constrained nucleophilic fluorination to a privileged reaction platform. In recent years, the Doyle group─along with many others─has pursued the study and development of this transformation with the intent of generating deeper mechanistic understanding, developing user-friendly fluorination reagents, and contributing to the invention of synthetic methods capable of enabling radiofluorination. Studies from our laboratory are discussed along with recent developments from others in this field. Fluoride reagent development and the mechanistic implications of reagent identity are highlighted. We also outline the chemical space inaccessible by current synthetic technologies and a series of future directions in the field that can potentially fill the existing dark spaces.

Using Data Science To Guide Aryl Bromide Substrate Scope Analysis in a Ni/Photoredox-Catalyzed Cross-Coupling with Acetals as Alcohol-Derived Radical Sources.

Ni/photoredox catalysis has emerged as a powerful platform for C(sp2)-C(sp3) bond formation. While many of these methods typically employ aryl bromides as the C(sp2) coupling partner, a variety of aliphatic radical sources have been investigated. In principle, these reactions enable access to the same product scaffolds, but it can be hard to discern which method to employ because nonstandardized sets of aryl bromides are used in scope evaluation. Herein, we report a Ni/photoredox-catalyzed (deutero)methylation and alkylation of aryl halides where benzaldehyde di(alkyl) acetals serve as alcohol-derived radical sources. Reaction development, mechanistic studies, and late-stage derivatization of a biologically relevant aryl chloride, fenofibrate, are presented. Then, we describe the integration of data science techniques, including DFT featurization, dimensionality reduction, and hierarchical clustering, to delineate a diverse and succinct collection of aryl bromides that is representative of the chemical space of the substrate class. By superimposing scope examples from published Ni/photoredox methods on this same chemical space, we identify areas of sparse coverage and high versus low average yields, enabling comparisons between prior art and this new method. Additionally, we demonstrate that the systematically selected scope of aryl bromides can be used to quantify population-wide reactivity trends and reveal sources of possible functional group incompatibility with supervised machine learning.

Vinyl-Fluorene Molecular Wires for Voltage Imaging with Enhanced Sensitivity and Reduced Phototoxicity.

Fluorescent voltage indicators are an attractive alternative for studying the electrical activity of excitable cells; however, the development of indicators that are both highly sensitive and low in toxicity over long-term experiments remains a challenge. Previously, we reported a fluorene-based voltage-sensitive fluorophore that exhibits much lower phototoxicity than previous voltage indicators in cardiomyocyte monolayers, but suffers from low sensitivity to membrane potential changes. Here, we report that the addition of a single vinyl spacer in the fluorene molecular wire scaffold improves the voltage sensitivity 1.5- to 3.5-fold over fluorene-based voltage probes. Furthermore, we demonstrate the improved ability of the new vinyl-fluorene VoltageFluors to monitor action potential kinetics in both mammalian neurons and human-induced pluripotent stem cell-derived cardiomyocytes. Addition of the vinyl spacer between the aniline donor and fluorene monomer results in indicators that are significantly less phototoxic in cardiomyocyte monolayers. These results demonstrate how structural modification to the voltage sensing domain have a large effect on improving the overall properties of molecular wire-based voltage indicators.

Photoredox Cyanomethylation of Indoles: Catalyst Modification and Mechanism.

The direct cyanomethylation of indoles at the 2- or 3-position was achieved via photoredox catalysis. The versatile nitrile synthon is introduced as a radical generated from bromoacetonitrile, a photocatalyst, and blue LED as a light source. The mechanism of the reaction is explored by determination of the Stern-Volmer quenching constants. By combining photophysical data and mass spectrometry to follow the catalyst decomposition, the catalyst ligands were tuned to enable synthetically useful yields of radical coupling products. A range of indole substrates with alkyl, aryl, halogen, ester, and ether functional groups participate in the reaction, affording products in 16-90% yields. The reaction allows the rapid construction of synthetically useful cyanomethylindoles, products that otherwise require several synthetic steps.

New Molecular Scaffolds for Fluorescent Voltage Indicators.

The ability to non-invasively monitor membrane potential dynamics in excitable cells like neurons and cardiomyocytes promises to revolutionize our understanding of the physiology and pathology of the brain and heart. Here, we report the design, synthesis, and application of a new class of fluorescent voltage indicators that make use of a fluorene-based molecular wire as a voltage-sensing domain to provide fast and sensitive measurements of membrane potential in both mammalian neurons and human-derived cardiomyocytes. We show that the best of the new probes, fluorene VoltageFluor 2 (fVF 2), readily reports on action potentials in mammalian neurons, detects perturbations to the cardiac action potential waveform in human induced pluripotent stem cell-derived cardiomyocytes, shows a substantial decrease in phototoxicity compared to existing molecular wire-based indicators, and can monitor cardiac action potentials for extended periods of time. Together, our results demonstrate the generalizability of a molecular wire approach to voltage sensing and highlight the utility of fVF 2 for interrogating membrane potential dynamics.