Towards integrated ADME prediction: past, present and future directions for modelling metabolism by UDP-glucuronosyltransferases

Artificial Intelligence in Drug Metabolism and Excretion Prediction: Recent Advances, Challenges, and Future Perspectives

Drug metabolism and excretion play crucial roles in determining the efficacy and safety of drug candidates, and predicting these processes is an essential part of drug discovery and development. In recent years, artificial intelligence (AI) has emerged as a powerful tool for predicting drug metabolism and excretion, offering the potential to speed up drug development and improve clinical success rates. This review highlights recent advances in AI-based drug metabolism and excretion prediction, including deep learning and machine learning algorithms. We provide a list of public data sources and free prediction tools for the research community. We also discuss the challenges associated with the development of AI models for drug metabolism and excretion prediction and explore future perspectives in the field. We hope this will be a helpful resource for anyone who is researching in silico drug metabolism, excretion, and pharmacokinetic properties.

Discovery of novel hit compounds as potential HDAC1 inhibitors: The case of ligand- and structure-based virtual screening

Histone deacetylases (HDACs) as an important family of epigenetic regulatory enzymes are implicated in the onset and progression of carcinomas. As a result, HDAC inhibition has been proven as a compelling strategy for reversing the aberrant epigenetic changes associated with cancer. However, non-selective profile of most developed HDAC inhibitors (HDACIs) leads to the occurrence of various side effects, limiting their clinical utility. This evidence provides a solid ground for ongoing research aimed at identifying isoform-selective inhibitors. Among the isoforms, HDAC1 have particularly gained increased attention as a preferred target for the design of selective HDACIs. Accordingly, in this paper, we have developed a reliable virtual screening process, combining different ligand- and structure-based methods, to identify novel benzamide-based analogs with potential HDAC1 inhibitory activity. For this purpose, a focused library of 736,160 compounds from PubChem database was first compiled based on 80% structural similarity with four known benzamide-based HDAC1 inhibitors, Mocetinostat, Entinostat, Tacedinaline, and Chidamide. Our inclusive in-house 3D-QSAR model, derived from pharmacophore-based alignment, was then employed as a 3D-query to discriminate hits with the highest predicted HDAC1 inhibitory activity. The selected hits were subjected to subsequent structure-based approaches (induced-fit docking (IFD), MM-GBSA calculations and molecular dynamics (MD) simulation) to retrieve potential compounds with the highest binding affinity for HDAC1 active site. Additionally, in silico ADMET properties and PAINS filtration were also considered for selecting an enriched set of the best drug-like molecules. Finally, six top-ranked hit molecules, CID_38265326, CID_56064109, CID_8136932, CID_55802151, CID_133901641 and CID_18150975 were identified to expose the best stability profiles and binding mode in the HDAC1 active site. The IFD and MD results cooperatively confirmed the interactions of the promising selected hits with critical residues within HDAC1 active site. In summary, the presented computational approach can provide a set of guidelines for the further development of improved benzamide-based derivatives targeting HDAC1 isoform.

A Diarylpentanoid with Potential Activation of the p53 Pathway: Combination of in silico Screening Studies, Synthesis, and Biological Activity Evaluation

In silico studies of a library of diarylpentanoids led us to the identification of potential new MDM2/X ligands. The diarylpentanoids with the best docking scores obeying the druglikeness and ADMET prediction properties were subsequently synthesized and evaluated for their antiproliferative activity on colon cancer HCT116 and fibroblasts HFF-1 cells. The effect on p53-MDM2/X interactions was evaluated through yeast-based assays for compounds showing potent antiproliferative activity in HCT116 cells and low toxicity in normal cells, resulting in the identification of a potential dual inhibitor. Moreover, its antiproliferative effect was significantly reduced in the absence of p53 and in MDA-MB-231 cells expressing a mutant p53 form. The antiproliferative effect of this compound was associated with induction of cell cycle arrest, apoptosis, PARP cleavage and increased p53 and its transcriptional targets, p21 and PUMA, in HCT116 cells. Docking poses and residues involved in the inhibition of p53-MDM2/X interactions were predicted by docking studies.

PreMetabo: An in silico phase I and II drug metabolism prediction platform

This study aimed to develop a drug metabolism prediction platform using knowledge-based prediction models. Site of Metabolism (SOM) prediction models for four cytochrome P450 (CYP) subtypes were developed along with uridine 5'-diphosphoglucuronosyltransferase (UGT) and sulfotransferase (SULT) substrate classification models. The SOM substrate for a certain CYP was determined using the sum of the activation energy required for the reaction at the reaction site of the substrate and the binding energy of the substrate to the CYP enzyme. Activation energy was calculated using the EaMEAD model and binding energy was calculated by docking simulation. Phase II prediction models were developed to predict whether a molecule is the substrate of a certain phase II conjugate protein, i.e., UGT or SULT. Using SOM prediction models, the predictability of the major metabolite in the top-3 was obtained as 72.5-84.5% for four CYPs, respectively. For internal validation, the accuracy of the UGT and SULT substrate classification model was obtained as 93.94% and 80.68%, respectively. Additionally, for external validation, the accuracy of the UGT substrate classification model was obtained as 81% in the case of 11 FDA-approved drugs. PreMetabo is implemented in a web environment and is available at https://premetabo.bmdrc.kr/.

Novel Series of Methyl 3-(Substituted Benzoyl)-7-Substituted-2-Phenylindolizine-1-Carboxylates as Promising Anti-Inflammatory Agents: Molecular Modeling Studies

The cyclooxygenase-2 (COX-2) enzyme is considered to be an important target for developing novel anti-inflammatory agents. Selective COX-2 inhibitors offer the advantage of lower adverse effects that are commonly associated with non-selective COX inhibitors. In this work, a novel series of methyl 3-(substituted benzoyl)-7-substituted-2-phenylindolizine-1-carboxylates was synthesized and evaluated for COX-2 inhibitory activity. Compound 4e was identified as the most active compound of the series with an IC50 of 6.71 M, which is comparable to the IC50 of indomethacin, a marketed non-steroidal anti-inflammatory drug (NSAID). Molecular modeling and crystallographic studies were conducted to further characterize the compounds and gain better understanding of the binding interactions between the compounds and the residues at the active site of the COX-2 enzyme. The pharmacokinetic properties and potential toxic effects were predicted for all the synthesized compounds, which indicated good drug-like properties. Thus, these synthesized compounds can be considered as potential lead compounds for developing effective anti-inflammatory therapeutic agents.

Prediction of UGT-mediated Metabolism Using the Manually Curated MetaQSAR Database

Even though glucuronidations are the most frequent metabolic reactions of conjugation, both in quantitative and qualitative terms, they have rather seldom been investigated using computational approaches. To fill this gap, we have used the manually collected MetaQSAR metabolic reaction database to generate two models for the prediction of UGT-mediated metabolism, both based on molecular descriptors and implementing the Random Forest algorithm. The first model predicts the occurrence of the reaction and was internally validated with a Matthew correlation coefficient (MCC) of 0.76 and an area under the ROC curve (AUC) of 0.94, and further externally validated using a test set composed of 120 additional xenobiotics (MCC of 0.70 and AUC of 0.90). The second model distinguishes between O- and N-glucuronidations and was optimized by the random undersampling procedure to improve the predictive accuracy during the internal validation, with the recall measure of the minority class increasing from 0.55 to 0.78.

Antihypertensive drugs metabolism: an update to pharmacokinetic profiles and computational approaches

Drug discovery and development is a high-risk enterprise that requires significant investments in capital, time and scientific expertise. The studies of xenobiotic metabolism remain as one of the main topics in the research and development of drugs, cosmetics and nutritional supplements.

Antihypertensive drugs are used for the treatment of high blood pressure, which is one the most frequent symptoms of the patients that undergo cardiovascular diseases such as myocardial infraction and strokes. In current cardiovascular disease pharmacology, four drug clusters - Angiotensin Converting Enzyme Inhibitors, Beta-Blockers, Calcium Channel Blockers and Diuretics - cover the major therapeutic characteristics of the most antihypertensive drugs. The pharmacokinetic and specifically the metabolic profile of the antihypertensive agents are intensively studied because of the broad inter-individual variability on plasma concentrations and the diversity on the efficacy response especially due to the P450 dependent metabolic status they present.

Several computational methods have been developed with the aim to: (i) model and better understand the human drug metabolism; and (ii) enhance the experimental investigation of the metabolism of small xenobiotic molecules. The main predictive tools these methods employ are rule-based approaches, quantitative structure metabolism/activity relationships and docking approaches.

This review paper provides detailed metabolic profiles of the major clusters of antihypertensive agents, including their metabolites and their metabolizing enzymes, and it also provides specific information concerning the computational approaches that have been used to predict the metabolic profile of several antihypertensive drugs.

Open Source Bayesian Models. 1. Application to ADME/Tox and Drug Discovery Datasets

On the order of hundreds of absorption, distribution, metabolism, excretion, and toxicity (ADME/Tox) models have been described in the literature in the past decade which are more often than not inaccessible to anyone but their authors. Public accessibility is also an issue with computational models for bioactivity, and the ability to share such models still remains a major challenge limiting drug discovery. We describe the creation of a reference implementation of a Bayesian model-building software module, which we have released as an open source component that is now included in the Chemistry Development Kit (CDK) project, as well as implemented in the CDD Vault and in several mobile apps. We use this implementation to build an array of Bayesian models for ADME/Tox, in vitro and in vivo bioactivity, and other physicochemical properties. We show that these models possess cross-validation receiver operator curve values comparable to those generated previously in prior publications using alternative tools. We have now described how the implementation of Bayesian models with FCFP6 descriptors generated in the CDD Vault enables the rapid production of robust machine learning models from public data or the user’s own datasets. The current study sets the stage for generating models in proprietary software (such as CDD) and exporting these models in a format that could be run in open source software using CDK components. This work also demonstrates that we can enable biocomputation across distributed private or public datasets to enhance drug discovery.

Binding Mode Investigation of Polyphenols fromScrophulariaTargeting Human Aldose Reductase Using Molecular Docking and Molecular Dynamics Simulations

Aldose reductase (ALR2), a vital enzyme involved in polyol pathway, has befitted as a novel drug target in antidiabetes drug discovery process. In the present study, the binding mode and pharmacokinetic properties of potential polyphenolic compounds with reported aldose reductase inhibitory activity from the genusScrophulariahave been investigated. The human ALR2 enzyme (PDB ID: 2FZD) acted as the receptor in the current study. Among the compounds investigated, acacetin, a methoxy flavonoid, displayed the stable binding to the active site of ALR2 with least binding energy value. Molecular interaction analysis revealed that acacetin interrupts the proton donation mechanism, necessary for the catalytic activity of ALR2, by forming H-bond with Tyr48 (proton donor). In addition, acacetin also possessed favorable ADME properties and complies with Lipinski’s rule of 5 representing the possible drug-like nature compared to other polyphenols. Interestingly, the biological activity predictions also ranked acacetin with higher probability score for aldose reductase inhibition activity. Moreover, the molecular dynamics simulation of ALR2-acacetin complex was validated for the stability of ligand binding and the refined complex was used for generation of receptor-ligand pharmacophore model. Thus, the molecular insights of receptor-ligand interactions gained from the present study can be utilized for the development of novel aldose reductase inhibitors fromScrophularia.

Substrate selectivity of human intestinal UDP-glucuronosyltransferases (UGTs): in silico and in vitro insights

The current drug development process aims to produce safe, effective drugs within a reasonable time and at a reasonable cost. Phase II metabolism (glucuronidation) can affect drug action and pharmacokinetics to a considerable extent and so its studies and prediction at initial stages of drug development are very imperative. Extensive glucuronidation is an obstacle to oral bioavailability because the first-pass glucuronidation [or premature clearance by UDP-glucuronosyltransferases (UGTs)] of orally administered agents frequently results in poor oral bioavailability and lack of efficacy. Modeling of new chemical entities/drugs for UGTs and their kinetic data can be useful in understanding the binding patterns to be used in the design of better molecules. This review concentrates on first-pass glucuronidation by intestinal UGTs, including their topology, expression profile, and pharmacogenomics. In addition, recent advances are discussed with respect to substrate selectivity at the binding pocket, structural requirements, and mechanism of enzyme actions.

Accurate prediction of glucuronidation of structurally diverse phenolics by human UGT1A9 using combined experimental and in silico approaches

PURPOSE

Catalytic selectivity of human UGT1A9, an important membrane-bound enzyme catalyzing glucuronidation of xenobiotics, was determined experimentally using 145 phenolics and analyzed by 3D-QSAR methods.

METHODS

Catalytic efficiency of UGT1A9 was determined by kinetic profiling. Quantitative structure activity relationships were analyzed using CoMFA and CoMSIA techniques. Molecular alignment of substrate structures was made by superimposing the glucuronidation site and its adjacent aromatic ring to achieve maximal steric overlap. For a substrate with multiple active glucuronidation sites, each site was considered a separate substrate.

RESULTS

3D-QSAR analyses produced statistically reliable models with good predictive power (CoMFA: q2 = 0.548, r2 = 0.949, r pred 2 = 0.775; CoMSIA: q2 = 0.579, r2 = 0.876, r pred 2 = 0.700). Contour coefficient maps were applied to elucidate structural features among substrates that are responsible for selectivity differences. Contour coefficient maps were overlaid in the catalytic pocket of a homology model of UGT1A9, enabling identification of the UGT1A9 catalytic pocket with a high degree of confidence.

CONCLUSION

CoMFA/CoMSIA models can predict substrate selectivity and in vitro clearance of UGT1A9. Our findings also provide a possible molecular basis for understanding UGT1A9 functions and substrate selectivity.

Understanding substrate selectivity of human UDP-glucuronosyltransferases through QSAR modeling and analysis of homologous enzymes

The UDP-glucuronosyltransferase (UGT) enzyme catalyzes the glucuronidation reaction which is a major metabolic and detoxification pathway in humans. Understanding the mechanisms for substrate recognition by UGT assumes great importance in an attempt to predict its contribution to xenobiotic/drug disposition in vivo. Spurred on by this interest, 2D/3D-quantitative structure activity relationships and pharmacophore models have been established in the absence of a complete mammalian UGT crystal structure. This review discusses the recent progress in modeling human UGT substrates including those with multiple sites of glucuronidation. A better understanding of UGT active site contributing to substrate selectivity (and regioselectivity) from the homologous enzymes (i.e. plant and bacterial UGTs, all belong to family 1 of glycosyltransferase (GT1)) is also highlighted, as these enzymes share a common catalytic mechanism and/or overlapping substrate selectivity.

Toward in silico structure-based ADMET prediction in drug discovery

Quantitative structure-activity relationship (QSAR) methods and related approaches have been used to investigate the molecular features that influence the absorption, distribution, metabolism, excretion and toxicity (ADMET) of drugs. As the three-dimensional structures of several major ADMET proteins become available, structure-based (docking-scoring) computations can be carried out to complement or to go beyond QSAR studies. Applying docking-scoring methods to ADMET proteins is a challenging process because they usually have a large and flexible binding cavity; however, promising results relating to metabolizing enzymes have been reported. After reviewing current trends in the field we applied structure-based methods in the context of receptor flexibility in a case study involving the phase II metabolizing sulfotransferases. Overall, the explored concepts and results suggested that structure-based ADMET profiling will probably join the mainstream during the coming years.

Structure–metabolism relationships for the glucuronidation of flavonoids by UGT1A3 and UGT1A9

Objectives

This study tries to find structure–metabolism relationships between flavonoids and human UGT1A3 and UGT1A9.

Methods

The glucuronidation of flavonoids was studied with recombinant UGT1A3 and UGT1A9, and the glucuronidation activity was determined by HPLC.

Key findings

Of the flavonoids studied, it was shown for the first time that baicalein, quercetin-3-OCH2OCH3, quercetin-4′-CH3, quercetin-3′-OCH3 and quercetin-3′-Br are substrates of UGT1A3. Wogonin, baicalein, quercetin-4′-Cl, quercetin-3-OCH2OCH3, quercetin-3-O-arabinoside, quercetin-4′-CH3, quercetin-3′-OCH3 and quercetin-3′-Br are the newly reported substrates of UGT1A9. The preferred substrates for UGT1A3 and UGT1A9 contain the hydroxyl group at the C7-position. The glycon and the position of the B ring have conspicuous influences on the glucuronidation activity, and other chemical structures of flavonoids have minor effects.

Conclusions

From the quantitative study, UGT1A9 in general has higher glucuronidation efficiency than UGT1A3.

Three-dimensional quantitative structure-activity relationship studies on UGT1A9-mediated 3-O-glucuronidation of natural flavonols using a pharmacophore-based comparative molecular field analysis model

Glucuronidation is often recognized as one of the rate-determining factors that limit the bioavailability of flavonols. Hence, design and synthesis of more bioavailable flavonols would benefit from the establishment of predictive models of glucuronidation using kinetic parameters [e.g., K(m), V(max), intrinsic clearance (CL(int)) = V(max)/K(m)] derived for flavonols. This article aims to construct position (3-OH)-specific comparative molecular field analysis (CoMFA) models to describe UDP-glucuronosyltransferase (UGT) 1A9-mediated glucuronidation of flavonols, which can be used to design poor UGT1A9 substrates. The kinetics of recombinant UGT1A9-mediated 3-O-glucuronidation of 30 flavonols was characterized, and kinetic parameters (K(m), V(max), CL(int)) were obtained. The observed K(m), V(max), and CL(int) values of 3-O-glucuronidation ranged from 0.04 to 0.68 μM, 0.04 to 12.95 nmol/mg/min, and 0.06 to 109.60 ml/mg/min, respectively. To model UGT1A9-mediated glucuronidation, 30 flavonols were split into the training (23 compounds) and test (7 compounds) sets. These flavonols were then aligned by mapping the flavonols to specific common feature pharmacophores, which were used to construct CoMFA models of V(max) and CL(int), respectively. The derived CoMFA models possessed good internal and external consistency and showed statistical significance and substantive predictive abilities (V(max) model: q(2) = 0.738, r(2) = 0.976, r(pred)(2) = 0.735; CL(int) model: q(2) = 0.561, r(2) = 0.938, r(pred)(2) = 0.630). The contour maps derived from CoMFA modeling clearly indicate structural characteristics associated with rapid or slow 3-O-glucuronidation. In conclusion, the approach of coupling CoMFA analysis with a pharmacophore-based structural alignment is viable for constructing a predictive model for regiospecific glucuronidation rates of flavonols by UGT1A9.

In silico pharmacology for drug discovery: applications to targets and beyond

Computational (in silico) methods have been developed and widely applied to pharmacology hypothesis development and testing. These in silico methods include databases, quantitative structure-activity relationships, similarity searching, pharmacophores, homology models and other molecular modeling, machine learning, data mining, network analysis tools and data analysis tools that use a computer. Such methods have seen frequent use in the discovery and optimization of novel molecules with affinity to a target, the clarification of absorption, distribution, metabolism, excretion and toxicity properties as well as physicochemical characterization. The first part of this review discussed the methods that have been used for virtual ligand and target-based screening and profiling to predict biological activity. The aim of this second part of the review is to illustrate some of the varied applications of in silico methods for pharmacology in terms of the targets addressed. We will also discuss some of the advantages and disadvantages of in silico methods with respect to in vitro and in vivo methods for pharmacology research. Our conclusion is that the in silico pharmacology paradigm is ongoing and presents a rich array of opportunities that will assist in expediating the discovery of new targets, and ultimately lead to compounds with predicted biological activity for these novel targets.

Comparative modelling of the human UDP-glucuronosyltransferases: insights into structure and mechanism

UDP-glucuronosyltranferases (UGTs) affect the disposition of drugs and other xenobiotics by catalysing the conjugation of glucuronic acid to available oxygen, nitrogen, and sulfur atoms. Several related mammalian isoforms of UGT are expressed that have different binding affinities and turnover rates for the substrates they encounter in the liver and other tissues. Because no high-resolution structural information is available to dissect the enzyme-substrate interactions that give rise to different specificities, a search was conducted to find the best available templates to use for comparative protein modelling. Sequence identity analysis was used to identify some recently crystallized plant UGTs as homologues of microsomal UGTs. Because UGTs contain a Rossman fold motif predicted to bind the UDP-containing sugar donor substrate, this consensus sequence was used to aid sequence alignment, as were other conserved residues thought to be involved in catalysis or substrate binding, and the predicted secondary structure. Docking of UDP-glucuronic acid to a model of UGT1A1 resulted in a root mean square deviation of only 0.37 angstroms vs. UDP co-crystallized with the plant UGT71G1 template. The significance of a comparative model generated for UGT1A1 with respect to both the sugar donor and aglycone binding sites, and mechanism, is discussed.

The importance of local chemical structure for chemical metabolism by human uridine 5'-diphosphate-glucuronosyltransferase

The uridine 5'-diphosphate- (UDP-)glucuronosyltransferase (UGT) family of enzymes catalyzes the conjugation of chemicals containing a suitable nucleophilic atom with glucuronic acid. Despite the importance of glucuronidation as an elimination and detoxification mechanism for drugs, environmental chemicals, and endogenous compounds, the structural features of substrates that confer isoform selectivity are poorly understood. The relationship between the local molecular structure of nucleophilic atoms of chemicals and the ability of UGT isoforms to glucuronidate the nucleophilic atoms was investigated here. The proximity of an aromatic ring to the nucleophilic atom was highly associated with a greater likelihood of glucuronidation by most UGT isoforms. Similarly, most UGT isoforms were found to have a statistically significant preference for oxygen over nitrogen as the nucleophilic atom. The converse was established only for UGT1A4. Naïve Bayes models were trained to predict the site of glucuronidation for eight UGT isoforms on the basis of the partial charge and Fukui function of the nucleophilic atom and whether an aromatic ring was attached to the nucleophilic atom. On average, the cross-validated sensitivity and specificity of the models were approximately 75-80%. For all but UGT2B7, the area under the receiver operating characteristics curve of the model was greater than 0.8, indicating strong predictive ability. A chemical diversity analysis of the currently available data indicates bias toward chemicals containing phenolic groups, and it is likely that the availability of chemical data sets with greater diversity will facilitate further insights into the structural features of substrates that confer enzyme selectivity.

Structure-based methods for the prediction of drug metabolism

There is a tantalising possibility that we may be able to predict the metabolism of a drug directly from its structure, thus obviating the requirement for animal tests in this area. There are a number of techniques that can be used to estimate a range of events associated with metabolism, and may allow us to achieve this aim. This paper considers the role of (quantitative) structure-activity relationships, and pharmacophore and homology modelling in the prediction of metabolism. Examples are also presented where such approaches have been formalised into expert systems. Clearly, many advances have been made in this area in recent years. Discussed herein is the importance of fully integrating the diverse systems and approaches available to fulfil the aspiration to predict metabolism directly from structure.

In vitro–in vivo correlation for drugs and other compounds eliminated by glucuronidation in humans: Pitfalls and promises

Enzymes of the UDP-glucuronosyltransferase (UGT) superfamily are responsible for the metabolism of many drugs, environmental chemicals and endogenous compounds. Identification of the UGT(s) involved in the metabolism of a given compound ('reaction phenotyping') currently relies on multiple confirmatory approaches, which may be confounded by the dependence of UGT activity on enzyme source, incubation conditions, and the occurrence of atypical glucuronidation kinetics. However, the increasing availability of substrate and inhibitor 'probes' for the individual UGTs provides the prospect for reliable phenotyping of glucuronidation reactions using human liver microsomes or hepatocytes, thereby providing data directly relevant to drug metabolism in humans. While the feasibility of computational prediction of UGT substrate selectivity has been demonstrated, the development of easily interpretable and generalisable models requires further improvement in the datasets available for analysis. Quantitative prediction of the hepatic clearance of glucuronidated drugs and the magnitude of inhibitory interactions based on in vitro kinetic data is more problematic. Intrinsic clearance (CL(int)) values generated using human liver microsomes under-predict in vivo hepatic clearance, typically by an order of magnitude. In vivo clearances of glucuronidated drugs are also generally under-predicted by CL(int) values from human hepatocytes, but to a lesser extent than observed with the microsomal model. While it is anticipated that systematic analysis of the potential causes of under-prediction may provide more reliable in vitro-in vivo scaling strategies, mechanistic interpretation of in vitro-in vivo correlation more broadly awaits further advances in our understanding of the structural and cellular determinants of UGT activity.

WHAT ARE ARTIFICIAL NEURAL NETWORKS AND WHAT THEY CAN DO?

The artificial neural networks (ANN) are very often applied in many areas of toxicology for the solving of complex problems, such as the prediction of chemical compound properties and quantitative structure-activity relationship. The aim of this contribution is to give the basic knowledge about conception of ANN, theirs division and finally, the typical application of ANN will be discussed. Due to the diversity of architectures and adaptation algorithms, the ANNs are used in the broad spectrum of applications from the environmental processes modeling, through the optimization to quantitative structure-activity relationship (QSAR) methods. In addition, especially ANNs with Kohonen learning are very effective classification tool. The ANNs are mostly applied in cases, where the commonly used methods does not work.

Rapid Prediction of Chemical Metabolism by Human UDP-glucuronosyltransferase Isoforms Using Quantum Chemical Descriptors Derived with the Electronegativity Equalization Method

This study aimed to evaluate in silico models based on quantum chemical (QC) descriptors derived using the electronegativity equalization method (EEM) and to assess the use of QC properties to predict chemical metabolism by human UDP-glucuronosyltransferase (UGT) isoforms. Various EEM-derived QC molecular descriptors were calculated for known UGT substrates and nonsubstrates. Classification models were developed using support vector machine and partial least squares discriminant analysis. In general, the most predictive models were generated with the support vector machine. Combining QC and 2D descriptors (from previous work) using a consensus approach resulted in a statistically significant improvement in predictivity (to 84%) over both the QC and 2D models and the other methods of combining the descriptors. EEM-derived QC descriptors were shown to be both highly predictive and computationally efficient. It is likely that EEM-derived QC properties will be generally useful for predicting ADMET and physicochemical properties during drug discovery.