Comparison of statistical methods for predicting penetration capacity of drugs into human breast milk using physicochemical, pharmacokinetic and chromatographic descriptors

Wanat, Karolina; Khakimov, Bekzod; Brzezińska, Elżbieta

doi:10.1080/1062936x.2020.1772365

Cited by 5 publications

(16 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To the best of our knowledge, previous studies ,,− on building regression-based models for predicting the M/P ratio of xenobiotic chemicals have employed much smaller data sets in comparison to our data set of 375 chemicals with known M/P ratios, and moreover, in many of these published studies, ,, the authors have not reported the R 2 value on test data. Moreover, none of the previous studies on this topic have published the source code of the developed predictive models.…”

Section: Resultsmentioning

confidence: 98%

“…Experimental measurement of a xenobiotic chemical’s propensity to enter human milk is both a difficult task and ethically impractical. Since the 1980s, there have been attempts to predict the M/P ratio for xenobiotic chemicals, and the initial studies were based on methods incorporating the physicochemical properties of the chemicals while ignoring clinical information or effects of active transport. , Following the initial attempts, several studies employing machine learning algorithms and the quantitative structure–activity relationship (QSAR) principle have been proposed. − Such predictive models have been built on the QSAR principle that the biological or chemical activity of a compound can be quantitatively related to its molecular structure and physicochemical properties. For instance, Yap and Chen developed a regression model based on a general regression neural network and reported an R 2 value of 0.677 and mean squared error (MSE) value of 0.206 on test data .…”

Section: Introductionmentioning

confidence: 99%

“…For their regression model, the authors obtained an R 2 value of 0.7 on train data, and for their classification model, the authors obtained an accuracy of 56.82%, sensitivity of 63.16%, and F -score of 55.81% on test data. Wanat et al developed a random forest (RF)-based regression model and reported an R 2 value of 0.29 on test data …”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Machine Learning Models for Prediction of Xenobiotic Chemicals with High Propensity to Transfer into Human Milk

Vijayaraghavan,

Lakshminarayanan,

Bhargava

et al. 2024

ACS Omega

View full text Add to dashboard Cite

Breast milk serves as a vital source of essential nutrients for infants. However, human milk contamination via the transfer of environmental chemicals from maternal exposome is a significant concern for infant health. The milk to plasma concentration (M/P) ratio is a critical metric that quantifies the extent to which these chemicals transfer from maternal plasma into breast milk, impacting infant exposure. Machine learning-based predictive toxicology models can be valuable in predicting chemicals with a high propensity to transfer into human milk. To this end, we build such classification- and regression-based models by employing multiple machine learning algorithms and leveraging the largest curated data set, to date, of 375 chemicals with known milk-to-plasma concentration (M/P) ratios. Our support vector machine (SVM)-based classifier outperforms other models in terms of different performance metrics, when evaluated on both (internal) test data and an external test data set. Specifically, the SVM-based classifier on (internal) test data achieved a classification accuracy of 77.33%, a specificity of 84%, a sensitivity of 64%, and an F-score of 65.31%. When evaluated on an external test data set, our SVM-based classifier is found to be generalizable with a sensitivity of 77.78%. While we were able to build highly predictive classification models, our best regression models for predicting the M/P ratio of chemicals could achieve only moderate R 2 values on the (internal) test data. As noted in the earlier literature, our study also highlights the challenges in developing accurate regression models for predicting the M/P ratio of xenobiotic chemicals. Overall, this study attests to the immense potential of predictive computational toxicology models in characterizing the myriad of chemicals in the human exposome.

show abstract

Section: Resultsmentioning

confidence: 98%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Machine Learning Models for Prediction of Xenobiotic Chemicals with High Propensity to Transfer into Human Milk

Vijayaraghavan,

Lakshminarayanan,

Bhargava

et al. 2024

ACS Omega

View full text Add to dashboard Cite

show abstract

“…The QSAR/QSPR approach aims to find correlations between structural features or physicochemical constants of a drug and its biological activity, and can be applied to predict physical and chemical properties by means of descriptors that explain changes in the physical or chemical properties of that drug group. A number of linear regression models have been reported for QSAR models predicting M/P ratios [17][18][19][20][21][22][23], but because M/P ratio data are collected from individual reports, uncertainties in subjects, measurement methods, and variations in the number of cases may affect the models. A classification model was also constructed based on the idea that prediction by linear regression is not realistic [24,25].…”

Section: Introductionmentioning

confidence: 99%

Prediction model for milk transfer of drugs by primarily evaluating the area under the curve using QSAR/QSPR

et al. 2023

View full text Add to dashboard Cite

Purpose Information on milk transferability of drugs is important for patients who wish to breastfeed. The purpose of this study is to develop a prediction model for milk-to-plasma drug concentration ratio based on area under the curve (M/PAUC). The quantitative structure–activity/property relationship (QSAR/QSPR) approach was used to predict compounds involved in active transport during milk transfer. Methods We collected M/P ratio data from literature, which were curated and divided into M/PAUC ≥ 1 and M/PAUC < 1. Using the ADMET Predictor® and ADMET Modeler™, we constructed two types of binary classification models: an artificial neural network (ANN) and a support vector machine (SVM). Results M/P ratios of 403 compounds were collected, M/PAUC data were obtained for 173 compounds, while 230 compounds only had M/Pnon-AUC values reported. The models were constructed using 129 of the 173 compounds, excluding colostrum data. The sensitivity of the ANN model was 0.969 for the training set and 0.833 for the test set, while the sensitivity of the SVM model was 0.971 for the training set and 0.667 for the test set. The contribution of the charge-based descriptor was high in both models. Conclusions We built a M/PAUC prediction model using QSAR/QSPR. These predictive models can play an auxiliary role in evaluating the milk transferability of drugs.

show abstract

“…In the previous articles [ 1 , 2 ], we presented a comparison of statistical methods in the study of drug excretion into breast milk with the use of the M/P descriptor. It was shown that the multiple linear regression (MLR) and random forest (RF) analyses were most effective in describing this pharmacokinetic phenomenon, with the use of chromatographic data and physicochemical properties of the tested compounds.…”

Section: Introductionmentioning

confidence: 99%

Statistical Methods in the Study of Protein Binding and Its Relationship to Drug Bioavailability in Breast Milk

Wanat

Brzezińska

2022

Molecules

Self Cite

View full text Add to dashboard Cite

Protein binding (PB) is indicated as the factor most severely limiting distribution in the organism, reducing the bioavailability of the drug, but also minimizing the penetration of xenobiotics into the fetus or the body of a breastfed child. Therefore, PB is an important aspect to be analyzed and monitored in the design of new drug substances. In this paper, several statistical analyses have been introduced to find the relationship between protein binding and the amount of drug in breast milk and to select molecular descriptors responsible for both pharmacokinetic phenomena. Along with descriptors related to the physicochemical properties of drugs, chromatographic descriptors from TLC and HPLC experiments were also used. Both methods used modification of the stationary phase, using bovine serum albumin (BSA) in TLC and human serum albumin (HSA) in HPLC. The use of the chromatographic data in the protein binding study was found to be positive —the most effective application of normal-phase TLC and HPLCHSA data was found. Statistical analyses also confirmed the prognostic value of affinity chromatography data and protein binding itself as the most important parameters in predicting drug excretion into breast milk.

show abstract

Comparison of statistical methods for predicting penetration capacity of drugs into human breast milk using physicochemical, pharmacokinetic and chromatographic descriptors

Cited by 5 publications

References 30 publications

Machine Learning Models for Prediction of Xenobiotic Chemicals with High Propensity to Transfer into Human Milk

Machine Learning Models for Prediction of Xenobiotic Chemicals with High Propensity to Transfer into Human Milk

Prediction model for milk transfer of drugs by primarily evaluating the area under the curve using QSAR/QSPR

Statistical Methods in the Study of Protein Binding and Its Relationship to Drug Bioavailability in Breast Milk

Contact Info

Product

Resources

About