Skip to content
PaperFren

ML plus transition states predict SNAr barriers

Open paper intelligence

A Gaussian-process model that includes SNAr transition-state features predicts experimental activation energies with 0.77 kcal mol−1 MAE, beating raw DFT.

Source

Machine learning meets mechanistic modelling for accurate prediction of experimental activation energies

Jorner K, Brinck T, Norrby PO, et al. · Chemical science · 2021

doi.org/10.1039/d0sc04896hRead the full paper ↗89 citationscc by

Study at a glance

Design
Computational / modelling — DFT/physical-organic descriptors with GPR for SNAr barriers and selectivity
N
ML on reaction barriers/selectivity; learning curves emphasize 50–150 training samples — no single cohort N
Population
Nucleophilic aromatic substitution reactions (including patent-selectivity cases)
Outcome
Predicted activation free energies and regio-/chemoselectivity accuracy

Structured fields used in claim comparison tables when every cited study has a complete layer.

What they did

Authors combined physical-organic and DFT transition-state descriptors for nucleophilic aromatic substitutions, trained several regressors, and tested regio-/chemoselectivity on patent reactions plus learning curves versus sample size.

What they found

GPR (Matern 3/2) reaches R2 = 0.93 and MAE = 0.77 kcal mol−1 on a held-out set, below 1 kcal mol−1 chemical accuracy. Raw DFT MAE is 2.93. Selectivity top-1 accuracy is 86%. Hybrid models help most with 50–150 samples.

The limits

What it doesn't show

Acid/base catalysis and other reaction classes are not yet in the model; TS geometry still needs reliable computation, which the authors call the bottleneck.

Key terms

SNAr
Nucleophilic aromatic substitution, the prototype reaction modeled here.
Chemical accuracy
About 1 kcal mol−1 error, the target for useful barrier prediction.
GPR
Gaussian process regression; Matern 3/2 kernel was the best model.
Hybrid model
ML that mixes mechanistic TS features with traditional QSRR descriptors.
QSRR
Quantitative structure–reactivity relationship using physical-organic features.

Flashcards

1 / 11

Research intelligence for this paper

See its role on concept claims, tensions it is part of, placement history, and related discoveries.

Open paper intelligence

Quiz yourself

1 / 6

Best test MAE for SNAr Ea is:

Common questions

How accurate is the best model?

MAE 0.77 kcal mol−1 on the external test set.

How bad is uncorrected DFT?

MAE 2.93 kcal mol−1, no better than guessing the mean.

Can it pick the right site?

86% top-1 regio/chemoselectivity on patent data.

When to use hybrid vs DFT-only?

Hybrid shines around 50–150 labeled reactions.

More on Computational chemistry