Machine-Learning-Guided Discovery of Electrochemical Reactions

Journal of the American Chemical Society, Journal Year: 2022, Volume and Issue: 144(49), P. 22599 - 22610

Published: Dec. 2, 2022

The molecular structures synthesizable by organic chemists dictate the functions they can create. invention and development of chemical reactions are thus critical for to access new desirable functional molecules in all disciplines chemistry. This work seeks expedite exploration emerging areas chemistry devising a machine-learning-guided workflow reaction discovery. Specifically, this study uses machine learning predict competent electrochemical reactions. To end, we first develop representation that enables production general models with limited training data. Next, employ automated experimentation test large number These categorized as or incompetent mixtures, classification model was trained competency. is used screen 38,865 potential silico, predictions identify synthetic mechanistic interest, 80% which found be competent. Additionally, provide 38,865-member set hope accelerating field. We envision adopting such could enable rapid many fields

Language: Английский

---

%V_Bur index and steric maps: from predictive catalysis to machine learning

Chemical Society Reviews, Journal Year: 2023, Volume and Issue: 53(2), P. 853 - 882

Published: Dec. 19, 2023

Steric indices are parameters used in chemistry to describe the spatial arrangement of atoms or groups molecules.

Language: Английский

---

Recent advances in copper-catalyzed decarboxylative reactions of propargylic cyclic carbonates/carbamates

Chemical Communications, Journal Year: 2023, Volume and Issue: 59(49), P. 7483 - 7505

Published: Jan. 1, 2023

Copper-catalyzed decarboxylative reactions are powerful strategies for the construction of widely available skeletons such as allenes, ethynyl-containing heterocycles, and quaternary carbon centers.

Language: Английский

---

A unified ML framework for solubility prediction across organic solvents

Digital Discovery, Journal Year: 2023, Volume and Issue: 2(2), P. 356 - 367

Published: Jan. 1, 2023

A generic framework for enhancing an initial solubility prediction with ML, even simple methods and a modestly sized, sparse dataset. We dissect the setup to show model “locking on” target system as more data are made available.

Language: Английский

---

Catalytic Enantioselective Intramolecular Oxa-Michael Reaction to α,β-Unsaturated Esters and Amides

Journal of the American Chemical Society, Journal Year: 2023, Volume and Issue: 145(23), P. 12771 - 12782

Published: May 30, 2023

A bifunctional iminophosphorane (BIMP)-catalyzed, enantioselective intramolecular oxa-Michael reaction of alcohols to tethered, low electrophilicity Michael acceptors is described. Improved reactivity over previous reports (1 day vs 7 days), excellent yields (up 99%), and enantiomeric ratios 99.5:0.5 er) are demonstrated. The broad scope, enabled by catalyst modularity tunability, includes substituted tetrahydrofurans (THFs) tetrahydropyrans (THPs), oxaspirocycles, sugar natural product derivatives, dihydro-(iso)-benzofurans, iso-chromans. state-of-the-art computational study revealed that the enantioselectivity originates from presence several favorable intermolecular hydrogen bonds between BIMP substrate induce stabilizing electrostatic orbital interactions. newly developed catalytic approach was carried out on multigram scale, multiple adducts were further derivatized an array useful building blocks, providing access enantioenriched biologically active molecules products.

Language: Английский

---

AI for organic and polymer synthesis

Science China Chemistry, Journal Year: 2024, Volume and Issue: 67(8), P. 2461 - 2496

Published: June 26, 2024

Language: Английский

---

Predicting relative efficiency of amide bond formation using multivariate linear regression

Proceedings of the National Academy of Sciences, Journal Year: 2022, Volume and Issue: 119(16)

Published: April 11, 2022

Significance Given the ubiquity of amide coupling reactions, understanding factors which influence success reaction and having means to predict rate would streamline synthetic efforts. This study outlines a data science–based workflow for effective statistical modeling with sparse experimental data. We demonstrated informed substrate selection, collection interpretable molecular descriptors, model development rates. The resulting models illuminate features that impact allow prediction untested

Language: Английский

---

A broadly applicable quantitative relative reactivity model for nucleophilic aromatic substitution (S_NAr) using simple descriptors

Chemical Science, Journal Year: 2022, Volume and Issue: 13(43), P. 12681 - 12695

Published: Jan. 1, 2022

A model for S N Ar reactivity is reported, built from relative rate data obtained by competition studies. Based only on molecular descriptors of the electrophile, predicts and site selectivity many complex substrates.

Language: Английский

---

Prediction of Nucleophilicity and Electrophilicity Based on a Machine‐Learning Approach

ChemPhysChem, Journal Year: 2023, Volume and Issue: 24(14)

Published: May 3, 2023

Nucleophilicity and electrophilicity dictate the reactivity of polar organic reactions. In past decades, Mayr et al. established a quantitative scale for nucleophilicity (N) (E), which proved to be useful tool rationalization chemical reactivity. this study, holistic prediction model was developed through machine-learning approach. rSPOC, an ensemble molecular representation with structural, physicochemical solvent features, purpose. With 1115 nucleophiles, 285 electrophiles, 22 solvents, dataset is currently largest one prediction. The rSPOC trained Extra Trees algorithm showed high accuracy in predicting Mayr's N E parameters R2 0.92 0.93, MAE 1.45 1.45, respectively. Furthermore, practical applications model, instance, NADH, NADPH series enamines potential molecules unknown within seconds. An online platform (http://isyn.luoszgroup.com/) constructed based on current available free scientific community.

Language: Английский

---

Reactivity Prediction of Cu-Catalyzed Halogen Atom Transfer Reactions Using Data-Driven Techniques

Journal of the American Chemical Society, Journal Year: 2023, Volume and Issue: 145(39), P. 21587 - 21599

Published: Sept. 21, 2023

In catalysis, linear free energy relationships (LFERs) are commonly used to identify reaction descriptors that enable the prediction of outcomes and design more effective catalysts. Herein, LFERs established for reductive cleavage C(sp3)-X bond in alkyl halides (RX) by Cu complexes. This represents activation step atom transfer radical polymerization addition/cyclization. The values rate constant, kact, 107 complex/RX couples 5 different solvents spanning over 13 orders magnitude were effectively interpolated equation: log kact = sC(I + C S), where I, C, S are, respectively, initiator, catalyst, solvent parameters, sC is catalyst-specific sensitivity parameter. Furthermore, each these parameters was correlated relevant descriptors, which included dissociation RX its Tolman cone angle θ, electron affinity X, stabilization energy, standard reduction potential complex, polarizability parameter π* solvent, distortion complex transition state. set establishes fundamental properties complexes determine their reactivity need be considered when designing novel systems reactions. Finally, a multivariate regression (MLR) approach adopted develop an objective model surpassed predictive capability LFER equation. Thus, MLR employed predict >2000 pairs.

Language: Английский

---