Quantitative Structure-Pharmacokinetic Relationships for Drug Distribution Properties by Using General Regression Neural Network

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

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 R2 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.

Predicting the aggregation number of cationic surfactants based on ANN-QSAR modeling approaches: understanding the impact of molecular descriptors on aggregation numbers

In this work, a quantitative structure-activity relationship (QSAR) study is performed on some cationic surfactants to evaluate the relationship between the molecular structures of the compounds with their aggregation numbers (AGGNs) in aqueous solution at 25 °C. An artificial neural network (ANN) model is combined with the QSAR study to predict the aggregation number of the surfactants. In the ANN analysis, four out of more than 3000 molecular descriptors were used as input variables, and the complete set of 41 cationic surfactants was randomly divided into a training set of 29, a test set of 6, and a validation set of 6 molecules. After that, a multiple linear regression (MLR) analysis was utilized to build a linear model using the same descriptors and the results were compared statistically with those of the ANN analysis. The square of the correlation coefficient (R ²) and root mean square error (RMSE) of the ANN and MLR models (for the whole data set) were 0.9392, 7.84, and 0.5010, 22.52, respectively. The results of the comparison revealed the efficiency of ANN in detecting a correlation between the molecular structure of surfactants and their AGGN values with a high predictive power due to the non-linearity in the studied data. Based on the ANN algorithm, the relative importance of the selected descriptors was computed and arranged in the following descending order: H-047 > ESpm12x > JGI6> Mor20p. Then, the QSAR data was interpreted and the impact of each descriptor on the AGGNs of the molecules were thoroughly discussed. The results showed there is a correlation between each selected descriptor and the AGGN values of the surfactants.

Optimizing acute stroke outcome prediction models: Comparison of generalized regression neural networks and logistic regressions

Background

Generalized regression neural network (GRNN) and logistic regression (LR) are extensively used in the medical field; however, the better model for predicting stroke outcome has not been established. The primary goal of this study was to compare the accuracies of GRNN and LR models to identify the most optimal model for the prediction of acute stroke outcome, as well as explore useful biomarkers for predicting the prognosis of acute stroke patients.

Method

In a single-center study, 216 (80% for the training set and 20% for the test set) acute stroke patients admitted to the Shenzhen Second People’s Hospital between December 2019 to June 2021 were retrospectively recruited. The functional outcomes of the patients were measured using Barthel Index (BI) on discharge. A training set was used to optimize the GRNN and LR models. The test set was utilized to validate and compare the performances of GRNN and LR in predicting acute stroke outcome based on the area under the receiver operating characteristic curve (AUROC), accuracy, sensitivity, and the Kappa value.

Result

The LR analysis showed that age, the National Institute Health Stroke Scale score, BI index, hemoglobin, and albumin were independently associated with stroke outcome. After validating in test set using these variables, we found that the GRNN model showed a better performance based on AUROC (0.931 vs 0.702), sensitivity (0.933 vs 0.700), specificity (0.889 vs 0.722), accuracy (0.896 vs 0.729), and the Kappa value (0.775 vs 0.416) than the LR model.

Conclusion

Overall, the GRNN model demonstrated superior performance to the LR model in predicting the prognosis of acute stroke patients. In addition to its advantage in not affected by implicit interactions and complex relationship in the data. Thus, we suggested that GRNN could be served as the optimal statistical model for acute stroke outcome prediction. Simultaneously, prospective validation based on more variables of the GRNN model for the prediction is required in future studies.

Machine learning in pharmacometrics: Opportunities and challenges

The explosive growth in medical devices, imaging and diagnostics, computing, and communication and information technologies in drug development and healthcare has created an ever-expanding data landscape that the pharmacometrics (PMX) research community must now traverse. The tools of machine learning (ML) have emerged as a powerful computational approach in other data-rich disciplines but its effective utilization in the pharmaceutical sciences and PMX modelling is in its infancy. ML-based methods can complement PMX modelling by enabling the information in diverse sources of big data, e.g. population-based public databases and disease-specific clinical registries, to be harnessed because they are capable of efficiently identifying salient variables associated with outcomes and delineating their interdependencies. ML algorithms are computationally efficient, have strong predictive capabilities and can enable learning in the big data setting. ML algorithms can be viewed as providing a computational bridge from big data to complement PMX modelling. This review provides an overview of the strengths and weaknesses of ML approaches vis-à-vis population methods, assesses current research into ML applications in the pharmaceutical sciences and provides perspective for potential opportunities and strategies for the successful integration and utilization of ML in PMX.

Towards cheminformatics-based estimation of drug therapeutic index: Predicting the protective index of anticonvulsants using a new quantitative structure-index relationship approach

The overall efficacy and safety profile of a new drug is partially evaluated by the therapeutic index in clinical studies and by the protective index (PI) in preclinical studies. In-silico predictive methods may facilitate the assessment of these indicators. Although QSAR and QSTR models can be used for predicting PI, their predictive capability has not been evaluated. To test this capability, we developed QSAR and QSTR models for predicting the activity and toxicity of anticonvulsants at accuracy levels above the literature-reported threshold (LT) of good QSAR models as tested by both the internal 5-fold cross validation and external validation method. These models showed significantly compromised PI predictive capability due to the cumulative errors of the QSAR and QSTR models. Therefore, in this investigation a new quantitative structure-index relationship (QSIR) model was devised and it showed improved PI predictive capability that superseded the LT of good QSAR models. The QSAR, QSTR and QSIR models were developed using support vector regression (SVR) method with the parameters optimized by using the greedy search method. The molecular descriptors relevant to the prediction of anticonvulsant activities, toxicities and PIs were analyzed by a recursive feature elimination method. The selected molecular descriptors are primarily associated with the drug-like, pharmacological and toxicological features and those used in the published anticonvulsant QSAR and QSTR models. This study suggested that QSIR is useful for estimating the therapeutic index of drug candidates.

High-Performance Personalized Heartbeat Classification Model for Long-Term ECG Signal

Long-term electrocardiogram (ECG) has become one of the important diagnostic assist methods in clinical cardiovascular domain. Long-term ECG is primarily used for the detection of various cardiovascular diseases that are caused by various cardiac arrhythmia such as myocardial infarction, cardiomyopathy, and myocarditis. In the past few years, the development of an automatic heartbeat classification method has been a challenge. With the accumulation of medical data, personalized heartbeat classification of a patient has become possible. For the long-term data accumulation method, such as the holter, it is difficult to obtain the analysis results in a short time using the original method of serial design. The pressure to develop a personalized automatic classification model is high. To solve these challenges, this paper implemented a parallel general regression neural network (GRNN) to classify the heartbeat, and achieved a 95% accuracy according to the Association for the Advancement of Medical Instrumentation. We designed an online learning program to form a personalized classification model for patients. The achieved accuracy of the model is 88% compared to the specific ECG data of the patients. The efficiency of the parallel GRNN with GTX780Ti can improve by 450 times.

Development and experimental test of support vector machines virtual screening method for searching Src inhibitors from large compound libraries

BACKGROUND

Src plays various roles in tumour progression, invasion, metastasis, angiogenesis and survival. It is one of the multiple targets of multi-target kinase inhibitors in clinical uses and trials for the treatment of leukemia and other cancers. These successes and appearances of drug resistance in some patients have raised significant interest and efforts in discovering new Src inhibitors. Various in-silico methods have been used in some of these efforts. It is desirable to explore additional in-silico methods, particularly those capable of searching large compound libraries at high yields and reduced false-hit rates.

RESULTS

We evaluated support vector machines (SVM) as virtual screening tools for searching Src inhibitors from large compound libraries. SVM trained and tested by 1,703 inhibitors and 63,318 putative non-inhibitors correctly identified 93.53%~ 95.01% inhibitors and 99.81%~ 99.90% non-inhibitors in 5-fold cross validation studies. SVM trained by 1,703 inhibitors reported before 2011 and 63,318 putative non-inhibitors correctly identified 70.45% of the 44 inhibitors reported since 2011, and predicted as inhibitors 44,843 (0.33%) of 13.56M PubChem, 1,496 (0.89%) of 168 K MDDR, and 719 (7.73%) of 9,305 MDDR compounds similar to the known inhibitors.

CONCLUSIONS

SVM showed comparable yield and reduced false hit rates in searching large compound libraries compared to the similarity-based and other machine-learning VS methods developed from the same set of training compounds and molecular descriptors. We tested three virtual hits of the same novel scaffold from in-house chemical libraries not reported as Src inhibitor, one of which showed moderate activity. SVM may be potentially explored for searching Src inhibitors from large compound libraries at low false-hit rates.

Identification of small molecule aggregators from large compound libraries by support vector machines

Small molecule aggregators non-specifically inhibit multiple unrelated proteins, rendering them therapeutically useless. They frequently appear as false hits and thus need to be eliminated in high-throughput screening campaigns. Computational methods have been explored for identifying aggregators, which have not been tested in screening large compound libraries. We used 1319 aggregators and 128,325 non-aggregators to develop a support vector machines (SVM) aggregator identification model, which was tested by four methods. The first is five fold cross-validation, which showed comparable aggregator and significantly improved non-aggregator identification rates against earlier studies. The second is the independent test of 17 aggregators discovered independently from the training aggregators, 71% of which were correctly identified. The third is retrospective screening of 13M PUBCHEM and 168K MDDR compounds, which predicted 97.9% and 98.7% of the PUBCHEM and MDDR compounds as non-aggregators. The fourth is retrospective screening of 5527 MDDR compounds similar to the known aggregators, 1.14% of which were predicted as aggregators. SVM showed slightly better overall performance against two other machine learning methods based on five fold cross-validation studies of the same settings. Molecular features of aggregation, extracted by a feature selection method, are consistent with published profiles. SVM showed substantial capability in identifying aggregators from large libraries at low false-hit rates.

Virtual screening of Abl inhibitors from large compound libraries by support vector machines

Abl promotes cancers by regulating cell morphogenesis, motility, growth, and survival. Successes of several marketed and clinical trial Abl inhibitors against leukemia and other cancers and appearances of reduced efficacies and drug resistances have led to significant interest in and efforts for developing new Abl inhibitors. In silico methods of pharmacophore, fragment, and molecular docking have been used in some of these efforts. It is desirable to explore other in silico methods capable of searching large compound libraries at high yields and reduced false-hit rates. We evaluated support vector machines (SVM) as a virtual screening tool for searching Abl inhibitors from large compound libraries. SVM trained and tested by 708 inhibitors and 65,494 putative noninhibitors correctly identified 84.4 to 92.3% inhibitors and 99.96 to 99.99% noninhibitors in 5-fold cross validation studies. SVM trained by 708 pre-2008 inhibitors and 65 494 putative noninhibitors correctly identified 50.5% of the 91 inhibitors reported since 2008 and predicted as inhibitors 29,072 (0.21%) of 13.56M PubChem, 659 (0.39%) of 168K MDDR, and 330 (5.0%) of 6638 MDDR compounds similar to the known inhibitors. SVM showed comparable yields and substantially reduced false-hit rates against two similarity based and another machine learning VS methods based on the same training and testing data sets and molecular descriptors. These suggest that SVM is capable of searching Abl inhibitors from large compound libraries at low false-hit rates.

A prediction model for blood-brain barrier permeation and analysis on its parameter biologically

The objective of this paper is to build a reliable model based on the artificial neural network (ANN) for predicting the blood-brain barrier (BBB) permeability and reveal the effects of the molecular descriptor on the BBB permeability. Eight descriptors including high-affinity P-gp substrate probability and plasma protein binding ratio are selected to develop the model. The three layers feedforward neural network (8-5-1) is employed for the prediction of logBB. By analyzing the experimental results, polar surface area (PSA) seems to be the most important factor for BBB permeability. Different from traditional view, the Abraham's hydrogen-bond basicity (HBB) can make a positive contribution to logBB in rational range. The experimental results show that the ANN based model with eight selected descriptors as inputs can achieve good performance for logBB prediction, and the results of sensitivity analysis can be confirmed by the present biological and chemical research.

Evaluation of virtual screening performance of support vector machines trained by sparsely distributed active compounds

Virtual screening performance of support vector machines (SVM) depends on the diversity of training active and inactive compounds. While diverse inactive compounds can be routinely generated, the number and diversity of known actives are typically low. We evaluated the performance of SVM trained by sparsely distributed actives in six MDDR biological target classes composed of a high number of known actives (983-1645) of high, intermediate, and low structural diversity (muscarinic M1 receptor agonists, NMDA receptor antagonists, thrombin inhibitors, HIV protease inhibitors, cephalosporins, and renin inhibitors). SVM trained by regularly sparse data sets of 100 actives show improved yields at substantially reduced false-hit rates compared to those of published studies and those of Tanimoto-based similarity searching method based on the same data sets and molecular descriptors. SVM trained by very sparse data sets of 40 actives (2.4%-4.1% of the known actives) predicted 17.5-39.5%, 23.0-48.1%, and 70.2-92.4% of the remaining 943-1605 actives in the high, intermediate, and low diversity classes, respectively, 13.8-68.7% of which are outside the training compound families. SVM predicted 99.97% and 97.1% of the 9.997 M PUBCHEM and 167K remaining MDDR compounds as inactive and 2.6%-8.3% of the 19,495-38,483 MDDR compounds similar to the known actives as active. These suggest that SVM has substantial capability in identifying novel active compounds from sparse active data sets at low false-hit rates.

A support vector machines approach for virtual screening of active compounds of single and multiple mechanisms from large libraries at an improved hit-rate and enrichment factor

Support vector machines (SVM) and other machine-learning (ML) methods have been explored as ligand-based virtual screening (VS) tools for facilitating lead discovery. While exhibiting good hit selection performance, in screening large compound libraries, these methods tend to produce lower hit-rate than those of the best performing VS tools, partly because their training-sets contain limited spectrum of inactive compounds. We tested whether the performance of SVM can be improved by using training-sets of diverse inactive compounds. In retrospective database screening of active compounds of single mechanism (HIV protease inhibitors, DHFR inhibitors, dopamine antagonists) and multiple mechanisms (CNS active agents) from large libraries of 2.986 million compounds, the yields, hit-rates, and enrichment factors of our SVM models are 52.4-78.0%, 4.7-73.8%, and 214-10,543, respectively, compared to those of 62-95%, 0.65-35%, and 20-1200 by structure-based VS and 55-81%, 0.2-0.7%, and 110-795 by other ligand-based VS tools in screening libraries of >or=1 million compounds. The hit-rates are comparable and the enrichment factors are substantially better than the best results of other VS tools. 24.3-87.6% of the predicted hits are outside the known hit families. SVM appears to be potentially useful for facilitating lead discovery in VS of large compound libraries.

Machine learning approaches for predicting compounds that interact with therapeutic and ADMET related proteins

Computational methods for predicting compounds of specific pharmacodynamic and ADMET (absorption, distribution, metabolism, excretion and toxicity) property are useful for facilitating drug discovery and evaluation. Recently, machine learning methods such as neural networks and support vector machines have been explored for predicting inhibitors, antagonists, blockers, agonists, activators and substrates of proteins related to specific therapeutic and ADMET property. These methods are particularly useful for compounds of diverse structures to complement QSAR methods, and for cases of unavailable receptor 3D structure to complement structure-based methods. A number of studies have demonstrated the potential of these methods for predicting such compounds as substrates of P-glycoprotein and cytochrome P450 CYP isoenzymes, inhibitors of protein kinases and CYP isoenzymes, and agonists of serotonin receptor and estrogen receptor. This article is intended to review the strategies, current progresses and underlying difficulties in using machine learning methods for predicting these protein binders and as potential virtual screening tools. Algorithms for proper representation of the structural and physicochemical properties of compounds are also evaluated.

Prediction of factor Xa inhibitors by machine learning methods

Factor Xa (FXa) inhibitors have been explored as anticoagulants for treatment and prevention of thrombotic diseases. Molecular docking, pharmacophore, quantitative structure-activity relationships, and support vector machines (SVM) have been used for computer prediction of FXa inhibitors. These methods achieve promising prediction accuracies of 69-80% for FXa inhibitors and 85-99% for non-inhibitors. Prediction performance, particularly for inhibitors, may be further improved by exploring methods applicable to more diverse range of compounds and by using more appropriate set of molecular descriptors. We tested the capability of several machine learning methods (C4.5 decision tree, k-nearest neighbor, probabilistic neural network, and support vector machine) by using a much more diverse set of 1098 compounds (360 inhibitors and 738 non-inhibitors) than those in other studies. A feature selection method was used for selecting molecular descriptors appropriate for distinguishing FXa inhibitors and non-inhibitors. The prediction accuracies of these methods are 89.1-97.5% for FXa inhibitors and 92.3-98.1% for non-inhibitors. In particular, compared to other studies, support vector machine gives a substantially improved accuracy of 94.6% for FXa non-inhibitors and maintains a comparable accuracy of 98.1% for inhibitors, based-on a more rigorous test with more diverse range of compounds. Our study suggests that machine learning methods such as SVM are useful for facilitating the prediction of FXa inhibitors.

MODEL—molecular descriptor lab: A web-based server for computing structural and physicochemical features of compounds

Molecular descriptors represent structural and physicochemical features of compounds. They have been extensively used for developing statistical models, such as quantitative structure activity relationship (QSAR) and artificial neural networks (NN), for computer prediction of the pharmacodynamic, pharmacokinetic, or toxicological properties of compounds from their structure. While computer programs have been developed for computing molecular descriptors, there is a lack of a freely accessible one. We have developed a web-based server, MODEL (Molecular Descriptor Lab), for computing a comprehensive set of 3,778 molecular descriptors, which is significantly more than the approximately 1,600 molecular descriptors computed by other software. Our computational algorithms have been extensively tested and the computed molecular descriptors have been used in a number of published works of statistical models for predicting variety of pharmacodynamic, pharmacokinetic, and toxicological properties of compounds. Several testing studies on the computed molecular descriptors are discussed. MODEL is accessible at http://jing.cz3.nus.edu.sg/cgi-bin/model/model.cgi free of charge for academic use.

Quantitative structure-pharmacokinetic/pharmacodynamic relationships

Quantitative structure-activity relationships have long been considered a vital component of drug discovery and development, providing insight into the role of molecular properties in the biological activity of similar and unrelated compounds. Recognition that in vitro bioassay and/or pre-clinical activity are insufficient for anticipating which compounds are suitable leads for further development has shifted the focus toward integrated pharmacokinetic (PK) and pharmacodynamic (PD) processes. Over the last decade, considerable progress has been made in constructing empirical and mechanistic quantitative structure-PK relationships (QSPKR), as well as diverse mechanism-based pharmacodynamic models of drug effects. In this review, traditional and contemporary approaches to developing QSPKR models are discussed, along with selected examples of attempts to couple QSPKR and pharmacodynamic models to anticipate the intensity and time-course of the pharmacological effects of new or related compounds, or quantitative structure-pharmacodynamic relationships modeling. Such models are in accordance with the goals of systems biology and the ideal of designing drugs and delivery systems from first principles.

Prediction of estrogen receptor agonists and characterization of associated molecular descriptors by statistical learning methods

Specific estrogen receptor (ER) agonists have been used for hormone replacement therapy, contraception, osteoporosis prevention, and prostate cancer treatment. Some ER agonists and partial-agonists induce cancer and endocrine function disruption. Methods for predicting ER agonists are useful for facilitating drug discovery and chemical safety evaluation. Structure-activity relationships and rule-based decision forest models have been derived for predicting ER binders at impressive accuracies of 87.1-97.6% for ER binders and 80.2-96.0% for ER non-binders. However, these are not designed for identifying ER agonists and they were developed from a subset of known ER binders. This work explored several statistical learning methods (support vector machines, k-nearest neighbor, probabilistic neural network and C4.5 decision tree) for predicting ER agonists from comprehensive set of known ER agonists and other compounds. The corresponding prediction systems were developed and tested by using 243 ER agonists and 463 ER non-agonists, respectively, which are significantly larger in number and structural diversity than those in previous studies. A feature selection method was used for selecting molecular descriptors responsible for distinguishing ER agonists from non-agonists, some of which are consistent with those used in other studies and the findings from X-ray crystallography data. The prediction accuracies of these methods are comparable to those of earlier studies despite the use of significantly more diverse range of compounds. SVM gives the best accuracy of 88.9% for ER agonists and 98.1% for non-agonists. Our study suggests that statistical learning methods such as SVM are potentially useful for facilitating the prediction of ER agonists and for characterizing the molecular descriptors associated with ER agonists.

A data base for partition of volatile organic compounds and drugs from blood/plasma/serum to brain, and an LFER analysis of the data

Literature values of the in vivo distribution (BB) of drugs from blood, plasma, or serum to rat brain have been assembled for 207 compounds (233 data points). We find that data on in vivo distribution from blood, plasma, and serum to rat brain can all be combined. Application of our general linear free energy relationship (LFER) to the 207 compounds yields an equation in log BB, with R2=0.75 and a standard deviation, SD, of 0.33 log units. An equation for a training set predicts the test set of data with a standard deviation of 0.31 log units. We further find that the in vivo data cannot simply be combined with in vitro data on volatile organic and inorganic compounds, because there is a systematic difference between the two sets of data. Use of an indicator variable allows the two sets to be combined, leading to a LFER equation for 302 compounds (328 data points) with R2=0.75 and SD=0.30 log units. A training equation was then used to predict a test set with SD=0.25 log units.

Quantitative structure-pharmacokinetic relationships for drug clearance by using statistical learning methods

Quantitative structure-pharmacokinetic relationships (QSPkR) have increasingly been used for the prediction of the pharmacokinetic properties of drug leads. Several QSPkR models have been developed to predict the total clearance (CL(tot)) of a compound. These models give good prediction accuracy but they are primarily based on a limited number of related compounds which are significantly lesser in number and diversity than the 503 compounds with known CL(tot) described in the literature. It is desirable to examine whether these and other statistical learning methods can be used for predicting the CL(tot) of a more diverse set of compounds. In this work, three statistical learning methods, general regression neural network (GRNN), support vector regression (SVR) and k-nearest neighbour (KNN) were explored for modeling the CL(tot) of all of the 503 known compounds. Six different sets of molecular descriptors, DS-MIXED, DS-3DMoRSE, DS-ATS, DS-GETAWAY, DS-RDF and DS-WHIM, were evaluated for their usefulness in the prediction of CL(tot). GRNN-, SVR- and KNN-developed models have average-fold errors in the range of 1.63 to 1.96, 1.66-1.95 and 1.90-2.23, respectively. For the best GRNN-, SVR- and KNN-developed models, the percentage of compounds with predicted CL(tot) within two-fold error of actual values are in the range of 61.9-74.3% and are comparable or slightly better than those of earlier studies. QSPkR models developed by using DS-MIXED, which is a collection of constitutional, geometrical, topological and electrotopological descriptors, generally give better prediction accuracies than those developed by using other descriptor sets. These results suggest that GRNN, SVR, and their consensus model are potentially useful for predicting QSPkR properties of drug leads.