Response to Comment on “Predicting reaction performance in C–N cross-coupling using machine learning”

Estrada, Jesús G.; Ahneman, Derek T.; Sheridan, Robert P.; Dreher, Spencer D.; Doyle, Abigail G.

doi:10.1126/science.aat8763

Cited by 61 publications

(78 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“… 3 – 4 This overarching issue in reaction optimization is often exasperated by subtle connections across several reaction variables, wherein modest structural changes to any or a few of these can have a profound effect on the experimental outcome. 5 – 6 , 7 These factors combined with the number of dimensions under study in most reactions, are the underlying reasons for why optimization is decidedly empirical. 8 – 9 This situation is particularly common in the area of asymmetric catalysis, wherein seemingly minor structural variations to any reaction component can have acute and non-intuitive influences on the observed enantioselectivity.…”

mentioning

confidence: 99%

Holistic prediction of enantioselectivity in asymmetric catalysis

Reid

Sigman

2019

Nature

252

269

View full text Add to dashboard Cite

Summary: When faced with unfamiliar reaction space, synthetic chemists typically apply reported conditions (reagents, catalyst, solvent, additives) from closely-related reactions to new substrate types. Unfortunately, this approach often fails due to subtle, albeit important, differences in reaction requirements. Consequently, a significant goal in synthetic chemistry is the ability to transfer chemical observations from one reaction to another, quantitatively. Here, we present such a platform by developing a holistic, data-driven workflow for deriving statistical models for one set of reactions that can be applied to predict out-of-sample examples. As a validating case study, published enantioselectivity data sets that employ BINOL-derived chiral phosphoric acids for a range of nucleophilic addition reactions to imines were combined and statistical models developed. These models reveal the general interactions imparting asymmetric induction and allow the quantitative transfer of this information to new reaction components. The disclosed techniques create opportunities for translating comprehensive reaction analysis to diverse chemical space, streamlining both catalyst and reaction development.

show abstract

mentioning

confidence: 99%

Holistic prediction of enantioselectivity in asymmetric catalysis

Reid

Sigman

2019

Nature

252

269

View full text Add to dashboard Cite

show abstract

“…Thus, the authors split the isoxazole additives into a variety of representative training and test sets and could prove good performance of the chemical feature model in these cases. 57 The same division for out-of-sample prediction using multiple fingerprints features (MFF) as input, showed comparable correlation in three of four test sets. 58 It should be emphasized that the onehot encoded models performed significantly worse ( Figure 5b).…”

Section: Resultsmentioning

confidence: 94%

A Structure-Based Platform for Predicting Chemical Reactivity

Sandfort¹,

Strieth‐Kalthoff²,

Kühnemund³

et al. 2019

Preprint

View full text Add to dashboard Cite

Despite their enormous potential, machine learning methods have only found limited application in predicting reaction outcomes, as current models are often highly complex and, most importantly, are not transferrable to different problem sets. Herein, we present the direct utilization of Lewis structures in a machine learning platform for diverse applications in organic chemistry. Therefore, an input based on multiple fingerprint features (MFF) as a universal molecular representation was developed and used for problem sets of increasing complexity: First, molecular properties across a diverse array of molecules could be predicted accurately. Next, reaction outcomes such as stereoselectivities and yields were predicted for experimental data sets that were previously evaluated using (complex) problem-oriented descriptor models. As a final application, a systematic high-throughput data set showed good correlation when using the MFF model, which suggests that this approach is general and ready for immediate adoption by chemists.

show abstract

“…The original assumptions regarding the significance and validity of the random-forest (chemical-feature) model to describe important and general chemical features were also confirmed (Estrada et al, 2018).…”

Section: Signs Of Controversymentioning

confidence: 81%

“…Despite the increasing number of works on the topic, the models proposed and practices carried out by chemists are entailing serious concerns (Chuang and Keiser, 2018a). Several technical challenges, pitfalls, and potentials of ML, and also the reliability of the results, have been discussed by some authors (Ahneman et al, 2018;Chuang and Keiser, 2018a,b;Estrada et al, 2018) corroborating some critical remarks on the fragility of purely data-based approaches (Microsoft, 2018). "If data can speak for themselves, they can also lie for themselves."…”

Section: Introductionmentioning

confidence: 98%

Deep Learning for Deep Chemistry: Optimizing the Prediction of Chemical Patterns

2019

View full text Add to dashboard Cite

Computational Chemistry is currently a synergistic assembly between ab initio calculations, simulation, machine learning (ML) and optimization strategies for describing, solving and predicting chemical data and related phenomena. These include accelerated literature searches, analysis and prediction of physical and quantum chemical properties, transition states, chemical structures, chemical reactions, and also new catalysts and drug candidates. The generalization of scalability to larger chemical problems, rather than specialization, is now the main principle for transforming chemical tasks in multiple fronts, for which systematic and cost-effective solutions have benefited from ML approaches, including those based on deep learning (e.g. quantum chemistry, molecular screening, synthetic route design, catalysis, drug discovery). The latter class of ML algorithms is capable of combining raw input into layers of intermediate features, enabling bench-to-bytes designs with the potential to transform several chemical domains. In this review, the most exciting developments concerning the use of ML in a range of different chemical scenarios are described. A range of different chemical problems and respective rationalization, that have hitherto been inaccessible due to the lack of suitable analysis tools, is thus detailed, evidencing the breadth of potential applications of these emerging multidimensional approaches. Focus is given to the models, algorithms and methods proposed to facilitate research on compound design and synthesis, materials design, prediction of binding, molecular activity, and soft matter behavior. The information produced by pairing Chemistry and ML, through data-driven analyses, neural network predictions and monitoring of chemical systems, allows (i) prompting the ability to understand the complexity of chemical data, (ii) streamlining and designing experiments, (ii) discovering new molecular targets and materials, and also (iv) planning or rethinking forthcoming chemical challenges. In fact, optimization engulfs all these tasks directly.

show abstract

Response to Comment on “Predicting reaction performance in C–N cross-coupling using machine learning”

Cited by 61 publications

References 6 publications

Holistic prediction of enantioselectivity in asymmetric catalysis

Holistic prediction of enantioselectivity in asymmetric catalysis

A Structure-Based Platform for Predicting Chemical Reactivity

Deep Learning for Deep Chemistry: Optimizing the Prediction of Chemical Patterns

Contact Info

Product

Resources

About