Organic reactivity from mechanism to machine learning

Jorner, Kjell; Tomberg, Anna; Bauer, Christoph; Sköld, Christian; Norrby, Per‐Ola

doi:10.1038/s41570-021-00260-x

Cited by 110 publications

(106 citation statements)

References 140 publications

Supporting

Mentioning

105

Contrasting

Order By: Relevance

“…[ 1‐3 ] Meanwhile, machine learning (ML) algorithms are gradually involved in the study of organic chemistry. [ 4‐6 ] ML has been widely applied in prediction of molecular and atomic properties, such as aqueous and non‐aqueous p K a , [ 7‐8 ] toxicity of compounds, [ 9 ] power conversion efficiencies of conjugated polymers, [ 10 ] molecular bond energy, [ 11 ] and atom condensed Fukui functions. [ 12 ] In these studies, variables are usually limited to a single molecule in the models.…”

Section: Background and Originality Contentmentioning

confidence: 99%

Prediction of Multicomponent Reaction Yields Using Machine Learning

et al. 2021

View full text Add to dashboard Cite

Main observation and conclusion Prediction of reaction yields using machine learning (ML) can help chemists select high‐yielding reactions and provide prior experience before wet‐lab experimenting to improve efficiency. However, the exploration of a multicomponent organic reaction features many complex variables and limited number of experimental data, which are challenging for the application of ML. Herein, we perform yield prediction for the synthesis of 2‐oxazolidones via Cu‐catalyzed radical‐type oxy‐alkylation of allylamines and herteroarylmethylamines with CO2, which is a three‐component reaction. Using physicochemical descriptors as features to launch ML modelling, we find that XGBoost shows significantly improved performance over linear models and these features are effective for the yield prediction. Moreover, out‐of‐sample prediction indicates the application potential of the model. This study demonstrates great potential of regression‐modelling‐based ML in organic synthesis even with complex factors and a general small size of reaction data, which are generated from the classical research pattern of method for the inquiry of multicomponent reactions.

show abstract

Section: Background and Originality Contentmentioning

confidence: 99%

Prediction of Multicomponent Reaction Yields Using Machine Learning

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Machine learning is currently attracting attention worldwide, and analysing and learning from a population using statistical methods enables immediate prediction of the results from new inputs. In the field of organic chemistry, machine learning is applied for predicting synthetic pathways and reactivity, and optimising reaction conditions [20][21][22][23][24][25][26] . Yu and co-workers reported the prediction model of BDE of carbonyl groups with machine learning in 2020 27 .…”

Section: Machine Learning Enabling Prediction Of the Bond Dissociation Enthalpy Of Hypervalent Iodine From Smilesmentioning

confidence: 99%

Machine learning enabling prediction of the bond dissociation enthalpy of hypervalent iodine from SMILES

Nakajima

Nemoto

2021

Sci Rep

View full text Add to dashboard Cite

Machine learning to create models on the basis of big data enables predictions from new input data. Many tasks formerly performed by humans can now be achieved by machine learning algorithms in various fields, including scientific areas. Hypervalent iodine compounds (HVIs) have long been applied as useful reactive molecules. The bond dissociation enthalpy (BDE) value is an important indicator of reactivity and stability. Experimentally measuring the BDE value of HVIs is difficult, however, and the value has been estimated by quantum calculations, especially density functional theory (DFT) calculations. Although DFT calculations can access the BDE value with high accuracy, the process is highly time-consuming. Thus, we aimed to reduce the time for predicting the BDE by applying machine learning. We calculated the BDE of more than 1000 HVIs using DFT calculations, and performed machine learning. Converting SMILES strings to Avalon fingerprints and learning using a traditional Elastic Net made it possible to predict the BDE value with high accuracy. Furthermore, an applicability domain search revealed that the learning model could accurately predict the BDE even for uncovered inputs that were not completely included in the training data.

show abstract

“…In the last years, the use of machine learning and data-driven approaches proved to be a very effective way to capture patterns from complex chemistry knowledge collections. 13 The extraction of chemical reaction rules from large data sets of traditional organic chemistry reactions 14 is one of the most successful examples 15 of providing transparency and explainability with AI applications in chemistry. While traditional synthetic organic chemistry went through its renaissance period thanks to recent development in machine learning and the availability of public chemical reaction datasets, the impact in biochemistry remained mostly bounded to the context of metabolic pathways prediction.…”

Section: Introductionmentioning

confidence: 99%

Molecular Transformer-aided Biocatalysed Synthesis Planning

Probst¹,

Manica²,

Teukam³

et al. 2021

Preprint

View full text Add to dashboard Cite

Enzyme catalysts are an integral part of green chemistry strategies towards a more sustainable and resource-efficient chemical synthesis. However, the use of enzymes on unreported substrates and their specific stereo- and regioselectivity are domain-specific knowledge factors that require decades of field experience to master. This makes the retrosynthesis of given targets with biocatalysed reactions a significant challenge. Here, we use the molecular transformer architecture to capture the latent knowledge about enzymatic activity from a large data set of publicly available biochemical reactions, extending forward reaction and retrosynthetic pathway prediction to the domain of biocatalysis. We introduce the use of a class token based on the EC classification scheme that allows to capture catalysis patterns among different enzymes belonging to the same hierarchical families. The forward prediction model achieves an accuracy of 49.6% and 62.7%, top-1 and top-5 respectively, while the single-step retrosynthetic model shows a round-trip accuracy of 39.6% and 42.6%, top-1 and top-10 respectively. Trained models and curated data are made publicly available with the hope of promoting enzymatic catalysis and making green chemistry more accessible through the use of digital technologies.

show abstract

Organic reactivity from mechanism to machine learning

Cited by 110 publications

References 140 publications

Prediction of Multicomponent Reaction Yields Using Machine Learning

Prediction of Multicomponent Reaction Yields Using Machine Learning

Machine learning enabling prediction of the bond dissociation enthalpy of hypervalent iodine from SMILES

Molecular Transformer-aided Biocatalysed Synthesis Planning

Contact Info

Product

Resources

About