Novel Computational Approach by Combining Machine Learning with Molecular Thermodynamics for Predicting Drug Solubility in Solvents

Ge, Kai; Ji, Yuanhui

doi:10.1021/acs.iecr.1c00998

Cited by 31 publications

(36 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It should be pointed out that NIST provides, through their ‘Thermo Data Engine’ (TDE), a novel method for extending the UNIFAC parameter table. , Other descriptors can be alternatively used. Within the SAFT framework, several innovative attempts exist with the use of ab initio descriptors, through neural network-type approaches . The COSMO models use as descriptors the so-called sigma profiles that also originate from ab initio calculations.…”

Section: Models and Simulationmentioning

confidence: 99%

“…Within the SAFT framework, several innovative attempts exist with the use of ab initio descriptors, through neural network-type approaches. 50 The COSMO models 51 use as descriptors the so-called sigma profiles that also originate from ab initio calculations. Alternative approaches based on artificial intelligence are seriously investigated.…”

Section: Macroscopic Thermodynamic Models For Equilibrium Properties ...mentioning

confidence: 99%

See 1 more Smart Citation

A View on the Future of Applied Thermodynamics

Hemptinne

Kontogeorgis

Dohrn

et al. 2022

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

Thermodynamics is the science of the interactions between energy and matter. It was formalized in the late 19th century and remains an essential piece in solving many technological challenges that society faces today. Yet, it is often considered complex and challenging, perhaps because it is often taught within a rigid mathematical framework, without highlighting the extensive range of applications and the tools that it offers for understanding and elaborating a sustainable future. The authors of this paper have performed an industrial survey (Kontogeorgis et al., Ind. Eng. Chem. Res., 2021, 60, 13, 4987-5013), which pointed out that thermodynamics is indeed a cornerstone of many processes in a large range of industries, but that as of today, many questions and needs remain unanswered. Some missing answers are caused by a lack of knowledge of the existing tools (educational issue), some by the unavailability of models, parameters or by the lack of transferability of the concepts from one system to another. In other cases, simply, no generally accepted approach exists, and fundamental research is required for understanding the phenomena. In all cases, data are needed, either to understand, develop, or validate the models. Specific recent examples of applied thermodynamics research relevant to industrial practice are discussed. This manuscript aims not only at promoting research but also at encouraging highly trained professionals to engage in education, laboratory work, fundamental developments, and/or model validation. Such professionals should find positions both in academia and in industry, as well as with software vendors. Collaboration between academia, industry, and software vendors is essential in order to foster new developments and serve the goals of sustainable development and circular economy.

show abstract

Section: Models and Simulationmentioning

confidence: 99%

Section: Macroscopic Thermodynamic Models For Equilibrium Properties ...mentioning

confidence: 99%

A View on the Future of Applied Thermodynamics

Hemptinne

Kontogeorgis

Dohrn

et al. 2022

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

show abstract

“…A combined molecular thermodynamic and ML‐based model was developed to predict the solubility of drugs using 16 molecular descriptors. The predictive model was trained using 5 different ML algorithms MLR, ANN, RF, extremely randomized trees (ET), and SVM (Ge & Ji, 2021). Finally, a single‐hidden‐layer neural network was found as a good predictive model for drug solubility which can be used for purpose of the drug development and drug solvent screening.…”

Section: Approaches In Drug Discoverymentioning

confidence: 99%

Machine learning approaches and their applications in drug discovery and design

et al. 2022

View full text Add to dashboard Cite

This review is focused on several machine learning approaches used in chemoinformatics. Machine learning approaches provide tools and algorithms to improve drug discovery. Many physicochemical properties of drugs like toxicity, absorption, drug‐drug interaction, carcinogenesis, and distribution have been effectively modeled by QSAR techniques. Machine learning is a subset of artificial intelligence, and this technique has shown tremendous potential in the field of drug discovery. Techniques discussed in this review are capable of modeling non‐linear datasets, as well as big data of increasing depth and complexity. Various machine learning‐based approaches are being used for drug target prediction, modeling the structure of drug target, binding site prediction, ligand‐based similarity searching, de novo designing of ligands with desired properties, developing scoring functions for molecular docking, building QSAR model for biological activity prediction, and prediction of pharmacokinetic and pharmacodynamic properties of ligands. In recent years, these predictive tools and models have achieved good accuracy. By the use of more related input data, relevant parameters, and appropriate algorithms, the accuracy of these predictions can be further improved.

show abstract

“…Unfortunately, its qualitative screening performance without kijs was not examined. Recently, Ge and Ji 43 proposed an empirical machine learning-based methodology combined with PC-SAFT to estimate the solubility of APIs using a number of molecular descriptors including known PC-SAFT parameters for pure APIs. Finally, the API parameters can also be estimated using a group contribution (GC) approach (e.g., ref 44 ) but this approach can generally be limited for the same reasons as UNIFAC.…”

Section: Introductionmentioning

confidence: 99%

Purely Predicting the Pharmaceutical Solubility: What to Expect from PC-SAFT and COSMO-RS?

Klajmon

2022

Mol. Pharmaceutics

View full text Add to dashboard Cite

A pair of popular thermodynamic models for pharmaceutical applications, namely, the perturbed-chain statistical associating fluid theory (PC-SAFT) equation of state and the conductor-like screening model for real solvents (COSMO-RS), are thoroughly benchmarked for their performance in predicting the solubility of active pharmaceutical ingredients (APIs) in pure solvents. The ultimate goal is to provide an illustration of what to expect from these progressive frameworks when applied to the thermodynamic solubility of APIs based on activity coefficients in a purely predictive regime without specific experimental solubility data (the fusion properties of pure APIs were taken from experiments). While this kind of prediction represents the typical modus operandi of the first-principles-aided COSMO-RS, PC-SAFT is a relatively highly parametrized model that relies on experimental data, against which its pure-substance and binary interaction parameters (kij) are fitted. Therefore, to make this benchmark as fair as possible, we omitted any binary parameters of PC-SAFT (that is, kij = 0 in all cases) and preferred pure-substance parameter sets for APIs not trained to experimental solubility data. This computational approach, together with a detailed assessment of the obtained solubility predictions against a large experimental data set, revealed that COSMO-RS convincingly outperformed PC-SAFT both qualitatively (that is, COSMO-RS was better in solvent ranking) and quantitatively, even though the former is independent of both substance-and mixture-specific experimental data. Regarding quantitative comparison, COSMO-RS outperformed PC-SAFT for nine of the ten APIs and for 63 % of the API-solvent systems, with root-mean-square deviations of the predicted data from the entire experimental data set being 0.82 and 1.44 log units, respectively.The results were further analyzed to expand the picture of the performance of both models with respect to the individual APIs and solvents. Interestingly, in many cases, both models were found to qualitatively incorrectly predict the direction of deviations from ideality. Furthermore, we examined how the solubility predictions from both models are sensitive to different API parametrizations.

show abstract

Novel Computational Approach by Combining Machine Learning with Molecular Thermodynamics for Predicting Drug Solubility in Solvents

Cited by 31 publications

References 41 publications

A View on the Future of Applied Thermodynamics

A View on the Future of Applied Thermodynamics

Machine learning approaches and their applications in drug discovery and design

Purely Predicting the Pharmaceutical Solubility: What to Expect from PC-SAFT and COSMO-RS?

Contact Info

Product

Resources

About