Machine learning model for prediction of drug solubility in supercritical solvent: Modeling and experimental validation

An, Feifei; Sayed, Biju Theruvil; Romero-Parra, Rosario Mireya; Hamad, Mohammed Haider; Sivaraman, R.; Foumani, Zahra Zanjani; Rushchitc, Anastasia Andreevna; El-Maghawry, Enas; Alzhrani, Rami M.; Alshehri, Sameer; AboRas, Kareem M.

doi:10.1016/j.molliq.2022.119901

Cited by 23 publications

(5 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The solubility uncertainty can be attributed to many factors, including the force field parameters. Recent advancements in machine learning-based potentials [38][39][40][41][42][43][44][45][46] will alleviate this issue to a great extent.…”

Section: Discussionmentioning

confidence: 99%

Polymorph-Specific Solubility Prediction using Constant Chemical Potential Molecular Dynamics

Neha,

Aggarwal,

Soni

et al. 2024

Preprint

View full text Add to dashboard Cite

Molecular Dynamics (MD) simulations offer a robust approach to understanding material properties within a system. Solubility is defined as the analytical composition of a saturated solution expressed as a proportion of designated solute in a designated solvent, according to IUPAC. It is a critical property of compounds and holds significance across numerous fields. Various computational techniques have been explored for determining solubility, including methods based on chemical potential determination, enhanced sampling simulation, and direct coexistence simulation, and lately, machine learning-based methods have shown promise. In this investigation, we aim to find the solubility values of a compound through Constant Chemical Potential Molecular Dynamics, a method rooted in direct coexistence simulation. The primary purpose of using this method is to overcome the limitation of the direct simulation method by maintaining a constant chemical potential for a sufficiently long time. Urea is chosen as a prototypical system for our study, with a particular focus on three of its polymorphs. Our approach effectively discriminates between the polymorphs of urea based on their respective solubility values; polymorph III is found to have the highest solubility followed by form IV and I.

show abstract

Section: Discussionmentioning

confidence: 99%

Polymorph-Specific Solubility Prediction using Constant Chemical Potential Molecular Dynamics

Neha,

Aggarwal,

Soni

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…At present, one of the most promising approaches for such a task involves the use of machine learning (ML) methods, as evidenced by the recent growth in the literature dedicated to the topic. − However, it is important to note that a majority of the constructed QSPR models are mostly used for the inter- or extrapolation of data, based on a rather small initial set. Notably, there are two noteworthy works , which utilize a larger data set of over a hundred drugs for model training, claiming that their models can successfully predict the properties of new compounds, expanding beyond the original training set.…”

Section: Introductionmentioning

confidence: 99%

Prediction of Drug-like Compounds Solubility in Supercritical Carbon Dioxide: A Comparative Study between Classical Density Functional Theory and Machine Learning Approaches

Makarov,

Kalikin,

Budkov

2024

Ind. Eng. Chem. Res.

View full text Add to dashboard Cite

Supercritical carbon dioxide (scCO 2 ) plays an essential role in various technological procedures, making the solubility of drugs in scCO 2 a crucial aspect of the drug formulation process. This study focuses on utilizing theoretical approaches to predict the solubility of drug-like compounds in scCO2 in order to select the optimum parameters for subsequent experimental procedures. Several machine learning models were developed and compared with a previously established theoretical approach based on classical density functional theory (cDFT). The CatBoost model, utilizing alvaDesc descriptors, demonstrated reasonably accurate predictions for the solubility of 187 drugs (AARD = 1.8%). Meanwhile, the CatBoost model, incorporating CDK descriptors and melting points of drugs as input parameters, exhibited satisfactory accuracy (AARD = 14.3%) in extrapolating predictions for new compounds. Comparing the results between the machine learning approach and the cDFT-based one revealed, on average, a higher accuracy and faster prediction speed for the former. However, cDFT demonstrated a more physical behavior of solubility isotherms compared with the machine learning models. This was particularly evident when the ML models struggled to accurately extrapolate solubility values beyond the experimental range of parameters in the supercritical state. Model CatBoost/CDK is freely accessible at http://chem-predictor.isc-ras.ru/ individual/scco/.

show abstract

“…It is a crucial step to measure and correlate drug solubility to prepare drugs with nanosized and better bioavailability. The process of supercritical for solid-dosage drugs is also considered as green technology because CO 2 gas is usually employed for the drug treatment, and no organic solvent is used for the process 8 – 10 .…”

Section: Introductionmentioning

confidence: 99%

“…Other approaches have been studied for enhancing drug solubility in water, however nanonization is a facile and effective process specifically mechanical approaches which do not use chemical agents for preparation of nanomedicines 10 . The method of supercritical processing can be also developed for continuous processing thereby a hybrid process can be developed using this novel technology.…”

Section: Introductionmentioning

confidence: 99%

Computational intelligence modeling of hyoscine drug solubility and solvent density in supercritical processing: gradient boosting, extra trees, and random forest models

Ghazwani

Begum

2023

Sci Rep

View full text Add to dashboard Cite

This work presents the results of using tree-based models, including Gradient Boosting, Extra Trees, and Random Forest, to model the solubility of hyoscine drug and solvent density based on pressure and temperature as inputs. The models were trained on a dataset of hyoscine drug with known solubility and density values, optimized with WCA algorithm, and their accuracy was evaluated using R2, MSE, MAPE, and Max Error metrics. The results showed that Gradient Boosting and Extra Trees models had high accuracy, with R2 values above 0.96 and low MAPE and Max Error values for both solubility and density output. The Random Forest model was less accurate than the other two models. These findings demonstrate the effectiveness of tree-based models for predicting the solubility and density of chemical compounds and have potential applications in determination of drug solubility prior to process design by correlation of solubility and density to input parameters including pressure and temperature.

show abstract

Machine learning model for prediction of drug solubility in supercritical solvent: Modeling and experimental validation

Cited by 23 publications

References 43 publications

Polymorph-Specific Solubility Prediction using Constant Chemical Potential Molecular Dynamics

Polymorph-Specific Solubility Prediction using Constant Chemical Potential Molecular Dynamics

Prediction of Drug-like Compounds Solubility in Supercritical Carbon Dioxide: A Comparative Study between Classical Density Functional Theory and Machine Learning Approaches

Computational intelligence modeling of hyoscine drug solubility and solvent density in supercritical processing: gradient boosting, extra trees, and random forest models

Contact Info

Product

Resources

About