Selectivity in organocatalysis—From qualitative to quantitative predictive models

Sterling, Alistair J.; Zavitsanou, Stamatia; Ford, Joseph; Duarte, Fernanda

doi:10.1002/wcms.1518

Cited by 28 publications

(37 citation statements)

References 70 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After conformation analysis of the peptide catalyst based on Miller’s pioneering studies on β -turn-containing tetrapeptides 62 , 63 , it was found that the approach of the radical adduct in the transition state (TS) is dictated by the C=O ··· H–N interaction in the backbone (highlighted with blue dash lines). Moreover, the Si and Re faces of the carbon radical interact differently with the peptidic thiol in the transition states due to non-covalent dispersion interactions 64 , 65 , with TS- Si displaying strong C–H ··· π interactions between the proline and the phenyl ring of the radical adduct. In contrast, only weak C−H ··· π interactions are identified in TS- Re .…”

Section: Resultsmentioning

confidence: 99%

Visible-light mediated catalytic asymmetric radical deuteration at non-benzylic positions

et al. 2022

View full text Add to dashboard Cite

Site- and enantioselective incorporation of deuterium into organic compounds is of broad interest in organic synthesis, especially within the pharmaceutical industry. While catalytic approaches relying on two-electron reaction manifolds have allowed for stereoselective delivery of a formal deuteride (D–) or deuteron (D+) at benzylic positions, complementary strategies that make use of one-electron deuterium atom transfer and target non-benzylic positions remain elusive. Here we report a photochemical approach for asymmetric radical deuteration by utilizing readily available peptide- or sugar-derived thiols as the catalyst and inexpensive deuterium oxide as the deuterium source. This metal-free platform enables four types of deuterofunctionalization reactions of exocyclic olefins and allows deuteration at non-benzylic positions with high levels of enantioselectivity and deuterium incorporation. Computational studies reveal that attractive non-covalent interactions are responsible for stereocontrol. We anticipate that our findings will open up new avenues for asymmetric deuteration.

show abstract

Section: Resultsmentioning

confidence: 99%

Visible-light mediated catalytic asymmetric radical deuteration at non-benzylic positions

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Although useful for the rationalization and prediction of absolute stereochemistry, the simultaneous analysis of multiple stereochemical models may limit the application. Further, simple models that can be sketched by hand are qualitative and do not provide a quantitative prediction of the enantioselectivity output . Accordingly, we deployed statistical modeling tools, which could greatly enhance the application and predictive power of our reaction models. − To connect these nine disparate reaction types, a comprehensive multivariate linear regression (cMLR) model that relates the features of all of the reaction components to the experimentally obtained enantioselectivity outcomes conveyed as ΔΔ G ‡ would be required (see SI for full details on workflow, statistical models considered, and additional validation tests) .…”

Section: Resultsmentioning

confidence: 99%

Comprehensive Stereochemical Models for Selectivity Prediction in Diverse Chiral Phosphate-Catalyzed Reaction Space

2021

View full text Add to dashboard Cite

The reactivity landscape of chiral phosphate catalysis is rapidly expanding and currently ranges from the hydrogenation of enals to the electrophilic activation of allenamides. Despite the importance of such transformations for the stereocontrolled synthesis of a diverse array of organic compounds, the preferred pathway and reasons for stereocontrol have not been firmly established, making it difficult to develop new reactions more generally. Here, we address this challenge by integrating traditional transition state calculations with statistical tools to rapidly connect and analyze several types of chiral phosphate-catalyzed enantioselective transformations. Detailed density functional theory (DFT) calculations of carefully selected case studies reveal that this set of superficially unrelated reactions operate, in many cases, through a single mechanism involving two hydrogen-bonding interactions from the iminium intermediate and nucleophile to the catalyst. From the transition state structures, we rationalize the different factors on which the enantioselectivity depends, focusing on the orientation of the reactants with respect to the catalyst. These theoretical analyses led to the construction of stereochemical models that correlate the magnitude and explain the sense of enantioselectivity for over 200 chemical reactions. We demonstrate how the resulting models can be used to assist reaction application to include additional substrates and develop related transformations. Ultimately, our findings represent a framework for formulating mechanistically relevant correlations driven by high-level transition state analysis, and this strategy should be broadly applicable to other catalytic systems widely applied in asymmetric synthesis.

show abstract

“…As opposed to physics, chemistry rarely has known theoretical models, and chemists usually work using their own intuition and empirical evidence. Based on this, chemists have developed the field of "chemoinformatics" where machine learning has been used extensively [18], [19], and recently generative architectures similar to the ones developed in this work have also been used [20]- [22]. The main difference in this work is that we did not aim to predict synthesizability, but we focused instead on tracking real-time experimental changes based on user-input.…”

Section: Related Workmentioning

confidence: 99%

Predicting Real-time Scientific Experiments Using Transformer models and Reinforcement Learning

Gutiérrez

2021

2021 20th IEEE International Conference on Machine Learning and Applications (ICMLA)

View full text Add to dashboard Cite

Life and physical sciences have always been quick to adopt the latest advances in machine learning to accelerate scientific discovery. Examples of this are cell segmentation or cancer detection. Nevertheless, these exceptional results are based on mining previously created datasets to discover patterns or trends. Recent advances in AI have been demonstrated in realtime scenarios like self-driving cars or playing video games. However, these new techniques have not seen widespread adoption in life or physical sciences because experimentation can be slow. To tackle this limitation, this work aims to adapt generative learning algorithms to model scientific experiments and accelerate their discovery using in-silico simulations. We particularly focused on real-time experiments, aiming to model how they react to user inputs. To achieve this, here we present an encoder-decoder architecture based on the Transformer model to simulate realtime scientific experimentation, predict its future behaviour and manipulate it on a step-by-step basis. As a proof of concept, this architecture was trained to map a set of mechanical inputs to the oscillations generated by a chemical reaction. The model was paired with a Reinforcement Learning controller to show how the simulated chemistry can be manipulated in real-time towards user-defined behaviours. Our results demonstrate how generative learning can model real-time scientific experimentation to track how it changes through time as the user manipulates it, and how the trained models can be paired with optimisation algorithms to discover new phenomena beyond the physical limitations of lab experimentation. This work paves the way towards building surrogate systems where physical experimentation interacts with machine learning on a step-by-step basis.

show abstract

Selectivity in organocatalysis—From qualitative to quantitative predictive models

Cited by 28 publications

References 70 publications

Visible-light mediated catalytic asymmetric radical deuteration at non-benzylic positions

Visible-light mediated catalytic asymmetric radical deuteration at non-benzylic positions

Comprehensive Stereochemical Models for Selectivity Prediction in Diverse Chiral Phosphate-Catalyzed Reaction Space

Predicting Real-time Scientific Experiments Using Transformer models and Reinforcement Learning

Contact Info

Product

Resources

About