An architecture of deep learning in QSPR modeling for the prediction of critical properties using molecular signatures

Yang, Su; Wang, Zihao; Jin, Saimeng; Shen, Weifeng; Ren, Jingzheng; Eden, Mario R.

doi:10.1002/aic.16678

Cited by 76 publications

(53 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Chemical compounds are vectorized to become a suitable input for the machine-learning algorithms to process. Various molecular representation methods have been implemented, simpler methods such as one-hot encoding [ 26 , 27 ] or vectorization of bond strings [ 28 ]. Molecular descriptors have also been shown to be useful for the representation of compounds [ 29 , 30 ].…”

Section: Related Workmentioning

confidence: 99%

Prediction of Chromatography Conditions for Purification in Organic Synthesis Using Deep Learning

2021

View full text Add to dashboard Cite

In this research, a process for developing normal-phase liquid chromatography solvent systems has been proposed. In contrast to the development of conditions via thin-layer chromatography (TLC), this process is based on the architecture of two hierarchically connected neural network-based components. Using a large database of reaction procedures allows those two components to perform an essential role in the machine-learning-based prediction of chromatographic purification conditions, i.e., solvents and the ratio between solvents. In our paper, we build two datasets and test various molecular vectorization approaches, such as extended-connectivity fingerprints, learned embedding, and auto-encoders along with different types of deep neural networks to demonstrate a novel method for modeling chromatographic solvent systems employing two neural networks in sequence. Afterward, we present our findings and provide insights on the most effective methods for solving prediction tasks. Our approach results in a system of two neural networks with long short-term memory (LSTM)-based auto-encoders, where the first predicts solvent labels (by reaching the classification accuracy of 0.950 ± 0.001) and in the case of two solvents, the second one predicts the ratio between two solvents (R2 metric equal to 0.982 ± 0.001). Our approach can be used as a guidance instrument in laboratories to accelerate scouting for suitable chromatography conditions.

show abstract

Section: Related Workmentioning

confidence: 99%

Prediction of Chromatography Conditions for Purification in Organic Synthesis Using Deep Learning

2021

View full text Add to dashboard Cite

show abstract

“…As proven by many studies [20][21][22][23][24][25], machine learning (ML) is an efficient and promising approach for building quantitative structure-property relationship (QSPR) models to predict various properties for chemical compounds. Until now, ML techniques have obtained wide applications and great successes in predicting IL properties, including melting point [26], CO 2 solubility [27], viscosity [28], etc.…”

Section: Introductionmentioning

confidence: 99%

Machine Learning for Ionic Liquid Toxicity Prediction

2020

Self Cite

View full text Add to dashboard Cite

In addition to proper physicochemical properties, low toxicity is also desirable when seeking suitable ionic liquids (ILs) for specific applications. In this context, machine learning (ML) models were developed to predict the IL toxicity in leukemia rat cell line (IPC-81) based on an extended experimental dataset. Following a systematic procedure including framework construction, hyper-parameter optimization, model training, and evaluation, the feedforward neural network (FNN) and support vector machine (SVM) algorithms were adopted to predict the toxicity of ILs directly from their molecular structures. Based on the ML structures optimized by the five-fold cross validation, two ML models were established and evaluated using IL structural descriptors as inputs. It was observed that both models exhibited high predictive accuracy, with the SVM model observed to be slightly better than the FNN model. For the SVM model, the determination coefficients were 0.9289 and 0.9202 for the training and test sets, respectively. The satisfactory predictive performance and generalization ability make our models useful for the computer-aided molecular design (CAMD) of environmentally friendly ILs.

show abstract

“…and critical properties [39]. Often SMILES strings are used as the input for the DNN model [36][37][38][39]. Such an approach proves useful in the screening and development of green solvents with respect to unconventional and novel compounds.…”

Section: Introductionmentioning

confidence: 99%

“…Among others, AI supported QSPRs have been used for the prediction of octanol–water partition coefficients [ 36 ], solvation free energies [ 37 ], gas chromatographic retention indices [ 38 ]. and critical properties [ 39 ]. Often SMILES strings are used as the input for the DNN model [ 36 , 37 , 38 , 39 ].…”

Section: Introductionmentioning

confidence: 99%

SUSSOL—Using Artificial Intelligence for Greener Solvent Selection and Substitution

2020

View full text Add to dashboard Cite

Solvents come in many shapes and types. Looking for solvents for a specific application can be hard, and looking for green alternatives for currently used nonbenign solvents can be even harder. We describe a new methodology for solvent selection and substitution, by applying Artificial Intelligence (AI) software to cluster a database of solvents based on their physical properties. The solvents are processed by a neural network, the Self-organizing Map of Kohonen, which results in a 2D map of clusters. The resulting clusters are validated both chemically and statistically and are presented in user-friendly visualizations by the SUSSOL (Sustainable Solvents Selection and Substitution Software) software. The software helps the user in exploring the solvent space and in generating and evaluating a list of possible alternatives for a specific solvent. The alternatives are ranked based on their safety, health, and environment scores. Cases are discussed to demonstrate the possibilities of our approach and to show that it can help in the search for more sustainable and greener solvents. The SUSSOL software makes intuitive sense and in most case studies, the software confirms the findings in literature, thus providing a sound platform for selecting the most sustainable solvent candidate.

show abstract

An architecture of deep learning in QSPR modeling for the prediction of critical properties using molecular signatures

Cited by 76 publications

References 43 publications

Prediction of Chromatography Conditions for Purification in Organic Synthesis Using Deep Learning

Prediction of Chromatography Conditions for Purification in Organic Synthesis Using Deep Learning

Machine Learning for Ionic Liquid Toxicity Prediction

SUSSOL—Using Artificial Intelligence for Greener Solvent Selection and Substitution

Contact Info

Product

Resources

About