Machine learning with physicochemical relationships: solubility prediction in organic solvents and water

Boobier, Samuel; Hose, David R. J.; Blacker, A. John; Nguyen, Bao

doi:10.1038/s41467-020-19594-z

Cited by 195 publications

(220 citation statements)

References 44 publications

(50 reference statements)

Supporting

Mentioning

216

Contrasting

Order By: Relevance

“…By inputting ALD in accordance with the atomic sequence in the input le, it is accessible to reconstruct chemical properties of atoms in the ML models. Boobier and co-workers also extracted atomic features in predicting the solubility of solutes in different solvents, but the features of the same elements were summed as descriptors nally 31 . We think this method will lead to the loss of atomic information by virtue of different atomic positions in the molecular structure.…”

Section: Resultsmentioning

confidence: 99%

Highly accurate prediction of reaction performance and enantioselectivity with a uniform machine learning protocol

Sun

Zhu

You

et al. 2021

Preprint

View full text Add to dashboard Cite

Synthetic reactions, especially asymmetric reactions are key components of modern chemistry. Chemists have put enormous experimental effort into recognizing various molecule patterns to enable efficient synthesis and asymmetric catalysis. Recent application of machine learning algorithms and chemoinformatics in this field demonstrated their huge potential in facilitating this process by accurate prediction. However, existing methods are relatively limited to specific designed data set, and only implement single prediction of reaction performance or reaction enantioselectivity, rendering their general use in broader scenarios challenging. Here we provide a uniform machine learning protocol that can predict both reaction performance and enantioselectivity with high accuracy. Reconstruction of molecular chemical space derived from more comprehensive three-dimensional atomic and molecular descriptors allow for training of our neural network-based model over four representative datasets. This uniform machine learning protocol was validated with outperformance of accuracy than other methods over all four cases (C-C, C-N, C-S cross coupling reactions and asymmetric hydrogenation) in the prediction of both reaction performance and enantioselectivity. It was also successfully applied to the out-of-set and sparse set prediction, leveraging its possible wide application in accelerating synthesis improve and molecular architects.

show abstract

Section: Resultsmentioning

confidence: 99%

Highly accurate prediction of reaction performance and enantioselectivity with a uniform machine learning protocol

Sun

Zhu

You

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

“…Sixty five structural group descriptors were used to generate the models. In 2020, Boobier et al proposed to develop models by using MLR, PLS, shallow neural networks, SVM, GP, RF, and ET for solubility prediction in water, ethanol, acetone, and benzene [ 31 ]. The models were constructed on 695 ethanol data, 464 benzene data, and 452 acetone data.…”

Section: Discussionmentioning

confidence: 99%

“…Machine learning algorithms have also been employed to construct models for solubility prediction in various organic solvents [ 7 , 30 , 31 ]. The models were generated on experimental solubility data.…”

Section: Introductionmentioning

confidence: 99%

Prediction of small-molecule compound solubility in organic solvents by machine learning algorithms

Ye¹,

Ouyang²

2021

J Cheminform

View full text Add to dashboard Cite

Rapid solvent selection is of great significance in chemistry. However, solubility prediction remains a crucial challenge. This study aimed to develop machine learning models that can accurately predict compound solubility in organic solvents. A dataset containing 5081 experimental temperature and solubility data of compounds in organic solvents was extracted and standardized. Molecular fingerprints were selected to characterize structural features. lightGBM was compared with deep learning and traditional machine learning (PLS, Ridge regression, kNN, DT, ET, RF, SVM) to develop models for predicting solubility in organic solvents at different temperatures. Compared to other models, lightGBM exhibited significantly better overall generalization (logS ± 0.20). For unseen solutes, our model gave a prediction accuracy (logS ± 0.59) close to the expected noise level of experimental solubility data. lightGBM revealed the physicochemical relationship between solubility and structural features. Our method enables rapid solvent screening in chemistry and may be applied to solubility prediction in other solvents.

show abstract

“…Machine Learning (ML) and Artificial Intelligence (AI) has gained tremendous importance and applications in the chemical industries ( Boobier et al., 2020 ; Dimiduk et al., 2018 ; Ge et al., 2017 ; Liu et al., 2017 ; Shang and You, 2019 ) They are emerging fields that can transform solvent and technology selection. In recent times, ML algorithms have been used to predict the physicochemical properties of organic solvents.…”

Section: Emerging Trends In Designing Solvent Recoverymentioning

confidence: 99%

“…In recent times, ML algorithms have been used to predict the physicochemical properties of organic solvents. An example is the prediction of the solubility of organic solvents in water presented by Boobier et al. (2020) .…”

Section: Emerging Trends In Designing Solvent Recoverymentioning

confidence: 99%

Systems level roadmap for solvent recovery and reuse in industries

2021

View full text Add to dashboard Cite

Summary Recovering waste solvent for reuse presents an excellent alternative to improving the greenness of industrial processes. Implementing solvent recovery practices in the chemical industry is necessary, given the increasing focus on sustainability to promote a circular economy. However, the systematic design of recovery processes is a daunting task due to the complexities associated with waste stream composition, techno-economic analysis, and environmental assessment. Furthermore, the challenges to satisfy the desired product specifications, particularly in pharmaceuticals and specialty chemical industries, may also deter solvent recovery and reuse practices. To this end, this review presents a systems-level approach including various methodologies that can be implemented to design and evaluate efficient solvent recovery pathways.

show abstract

Machine learning with physicochemical relationships: solubility prediction in organic solvents and water

Cited by 195 publications

References 44 publications

Highly accurate prediction of reaction performance and enantioselectivity with a uniform machine learning protocol

Highly accurate prediction of reaction performance and enantioselectivity with a uniform machine learning protocol

Prediction of small-molecule compound solubility in organic solvents by machine learning algorithms

Systems level roadmap for solvent recovery and reuse in industries

Contact Info

Product

Resources

About