Towards global QSAR model building for acute toxicity: Munro database case study

Multi-Endpoint Acute Toxicity Assessment of Organic Compounds Using Large-Scale Machine Learning Modeling

In recent years, alternative animal testing methods such as computational and machine learning approaches have become increasingly crucial for toxicity testing. However, the complexity and scarcity of available biomedical data challenge the development of predictive models. Combining nonlinear machine learning together with multicondition descriptors offers a solution for using data from various assays to create a robust model. This work applies multicondition descriptors (MCDs) to develop a QSTR (Quantitative Structure-Toxicity Relationship) model based on a large toxicity data set comprising more than 80,000 compounds and 59 different end points (122,572 data points). The prediction capabilities of developed single-task multi-end point machine learning models as well as a novel data analysis approach with the use of Convolutional Neural Networks (CNN) are discussed. The results show that using MCDs significantly improves the model and using them with CNN-1D yields the best result (R2train = 0.93, R2ext = 0.70). Several structural features showed a high level of contribution to the toxicity, including van der Waals surface area (VSA), number of nitrogen-containing fragments (nN+), presence of S-P fragments, ionization potential, and presence of C-N fragments. The developed models can be very useful tools to predict the toxicity of various compounds under different conditions, enabling quick toxicity assessment of new compounds.

Assessment of Some Unsymmetrical Porphyrins as Promising Molecules for Photodynamic Therapy of Cutaneous Disorders

In order to select for further development novel photosensitizers for photodynamic therapy in cutaneous disorders, three unsymmetrical porphyrins, namely 5-(4-hydroxy-3-methoxyphenyl)-10,15,20-tris-(4-acetoxy-3-methoxyphenyl) porphyrin (P2.2), 5-(2-hydroxy-5-methoxyphenyl)-10,15,20-tris-(4-carboxymethylphenyl) porphyrin (P3.2), and 5-(2,4-dihydroxyphenyl)-10,15,20-tris-(4-acetoxy-3-methoxyphenyl) porphyrin (P4.2), along with their fully symmetrical counterparts 5,10,15,20-tetrakis-(4-acetoxy-3-methoxyphenyl) porphyrin (P2.1) and 5,10,15,20-tetrakis-(4-carboxymethylphenyl) porphyrin (P3.1) were comparatively evaluated. The absorption and fluorescence properties, as well as atomic force microscopy measurements were performed to evaluate the photophysical characteristics as well as morphological and textural properties of the mentioned porphyrins. The cellular uptake of compounds and the effect of photodynamic therapy on the viability, proliferation, and necrosis of human HaCaT keratinocytes, human Hs27 skin fibroblasts, human skin SCL II squamous cell carcinoma, and B16F10 melanoma cells were assessed in vitro, in correlation with the structural and photophysical properties of the investigated porphyrins, and with the predictions regarding diffusion through cell membranes and ADMET properties. All samples were found to be isotropic and self-similar, with slightly different degrees of aggregability, had a relatively low predicted toxicity (class V), and a predicted long half-life after systemic administration. The in vitro study performed on non-malignant and malignant skin-relevant cells highlighted that the asymmetric P2.2 porphyrin qualified among the five investigated porphyrins to be a promising photosensitizer candidate for PDT in skin disorders. P2.2 was shown to accumulate well within cells, and induced by PDT a massive decrease in the number of metabolically active skin cells, partly due to cell death by necrosis. P2.2 had in this respect a better behavior than the symmetric P.2.1 compound and the related asymmetric compound P4.2. The strong action of P2.2-mediated PDT on normal skin cells might be an important drawback for further development of this compound. Meanwhile, the P3.1 and P3.2 compounds were not able to accumulate well in skin cells, and did not elicit significant PDT in vitro. Taken together, our experiments suggest that P2.2 can be a promising candidate for the development of novel photosensitizers for PDT in skin disorders.

A new perspective on predicting the reaction rate constants of hydrated electrons for organic contaminants: Exploring molecular structure characterization methods and ambient conditions

Hydrated electrons (eaq-) exhibit rapid degradation of diverse persistent organic contaminants (OCs) and hold great promise as a formidable reducing agent in water treatment. However, the diverse structures of compounds exert different influences on the second-order rate constant of hydrated electron reactions (keaq-), while the same OCs demonstrate notable discrepancies in keaq- values across different pH levels. This study aims to develop machine learning (ML) models that can effectively simulate the intricate reaction kinetics between eaq- and OCs. Furthermore, the introduction of the pH variable enables a comprehensive investigation into the impact of ambient conditions on this process, thereby improving the practicality of the model. A dataset encompassing 701 keaq- values derived from 351 peer-reviewed publications was compiled. To comprehensively investigate compound properties, this study introduced molecular descriptor (MD), molecular fingerprint (MF), and the integration of both (MD + MF) as model variables. Furthermore, 60 sets of predictive models were established utilizing two variable screening methodologies (MLR and RF) and ten prominent algorithms. Through statistical parameter analysis, it was determined that descriptors combined with MD and MF, the RF screening method, and the symbolism algorithm exhibited the best predictive efficacy. Importantly, the combination of descriptor models exhibited significantly superior performance compared to individual MF and MD models. Notably, the optimal model, denoted as RF - (MF + MD) - LGB, exhibited highly satisfactory predictive results (R2tra = 0.967, Q2tra = 0.840, R2ext = 0.761). The mechanistic explanation study based on Shapley Additive Explanations (SHAP) values further elucidated the crucial influences of polarity, pH, molecular weight, electronegativity, carbon-carbon double bonds, and molecular topology on the degradation of OCs by eaq-. The proposed modeling approach, particularly the integration of MF and MD, alongside the introduction of pH, may furnish innovative ideas for advanced reduction or oxidation processes (ARPs/AOPs) and machine learning applications in other domains.

Towards Anticancer and Antibacterial Agents: Design and Synthesis of 1,2,3-Triazol-quinobenzothiazine Derivatives

In this paper, we describe a new method for synthesizing hybrid combinations of 1,2,3-triazoles with a tetracyclic quinobenzothiazinium system. The developed approach allowed for the production of a series of new azaphenothiazine derivatives with the 1,2,3-triazole system in different positions of the benzene ring. In practice, the methodology consists of the reaction of triazole aniline derivatives with thioquinanthrenediinium bis-chloride. The structure of the products was determined by 1H-NMR, 13C-NMR spectroscopy, and HR-MS spectrometry, respectively. Moreover, the spatial structure of the molecule and the arrangement of molecules in the crystal (unit cell) were determined by X-ray crystallography. The anticancer activity profiles of the synthesized compounds were tested in vitro against human cancer cells of the A549, SNB-19, and T47D lines and the normal NHDF cell line. Additional tests of antibacterial activity against methicillin-sensitive and methicillin-resistant staphylococci, vancomycin-sensitive and vancomycin-resistant enterococci, and two mycobacterial strains were also performed. In fact, the dependence of anticancer and antibacterial activity on the substituent type and its position in the quinobenzothiazinium system was observed. Furthermore, the distance-guided property evaluation was performed using principal component analysis (PCA) and hierarchical clustering analysis (HCA) on the pool of the calculated descriptors. Finally, the theoretically approximated partition coefficients (clogP) were (inter-)correlated with each other and cross-compared with the empirically specified logPTLC parameters.

Cytotoxicity Analysis and In Silico Studies of Three Plant Extracts with Potential Application in Treatment of Endothelial Dysfunction

Endothelial dysfunction is the basis of the physiopathological mechanisms of vascular diseases. In addition to the therapeutic activity of plant extracts, cytotoxicity is significant. This research evaluates the cytotoxicity of three vegetal extracts (Calendulae flos extract-CE, Ginkgo bilobae folium extract-GE, and Sophorae flos extract-SE). In vitro evaluation was performed using an endothelial cell line model (Human Pulmonary Artery Endothelial Cells-HPAEC) when a dose-dependent cytotoxic activity was observed after 72 h. The IC50 values were calculated for all extracts: Calendulae flos extract (IC50 = 91.36 μg/mL), Sophorae flos extract (IC50 = 68.61 μg/mL), and Ginkgo bilobae folium extract (IC50 = 13.08 μg/mL). Therefore, at the level of HPAEC cells, the cytotoxicity of the extracts follows the order GE > SE > CE. The apoptotic mechanism implied in cell death was predicted for several phytocompounds using the PASS algorithm and molecular docking simulations, highlighting potential interactions with caspases-3 and -8. In vivo analysis was performed through brine shrimp lethality assay (BSLA) when lethal, behavioral, and cytological effects were evaluated on Artemia salina larvae. The viability examined after 24 h (assessment of lethal effects) follows the same sequence: CE > SE > GE. In addition, the predicted cell permeability was observed mainly for GE constituents through in silico studies. However, the extracts can be considered nontoxic according to Clarckson's criteria because no BSL% was registered at 1200 µg/mL. The obtained data reveal that all three extracts are safe for human use and suitable for incorporation in further pharmaceutical formulations.

Novel quantitative structure activity relationship models for predicting hexadecane/air partition coefficients of organic compounds

Predicting the logarithm of hexadecane/air partition coefficient (L) for organic compounds is crucial for understanding the environmental behavior and fate of organic compounds and developing prediction models with polyparameter linear free energy relationships. Herein, two quantitative structure activity relationship (QSAR) models were developed with 1272 L values for the organic compounds by using multiple linear regression (MLR) and support vector machine (SVM) algorithms. On the basis of the OECD principles, the goodness of fit, robustness and predictive ability for the developed models were evaluated. The SVM model was first developed, and the predictive capability for the SVM model is slightly better than that for the MLR model. The applicability domain (AD) of these two models has been extended to include more kinds of emerging pollutants, i.e., oraganosilicon compounds. The developed QSAR models can be used for predicting L values of various organic compounds. The van der Waals interactions between the organic compound and the hexadecane have a significant effect on the L value of the compound. These in silico models developed in current study can provide an alternative to experimental method for high-throughput obtaining L values of organic compounds.

Molecular Toxicity Virtual Screening Applying a Quantized Computational SNN-Based Framework

Spiking neural networks are biologically inspired machine learning algorithms attracting researchers' attention for their applicability to alternative energy-efficient hardware other than traditional computers. In the current work, spiking neural networks have been tested in a quantitative structure-activity analysis targeting the toxicity of molecules. Multiple public-domain databases of compounds have been evaluated with spiking neural networks, achieving accuracies compatible with high-quality frameworks presented in the previous literature. The numerical experiments also included an analysis of hyperparameters and tested the spiking neural networks on molecular fingerprints of different lengths. Proposing alternatives to traditional software and hardware for time- and resource-consuming tasks, such as those found in chemoinformatics, may open the door to new research and improvements in the field.

Versatile in silico modelling of microplastics adsorption capacity in aqueous environment based on molecular descriptor and machine learning

To comprehensively evaluate the hazards of microplastics and their coexisting organic pollutants, the sorption capacity of microplastics is a major issue that is quantified through the microplastic-aqueous sorption coefficient (K_d). Almost all quantitative structure-property relationship (QSPR) models that describe K_d apply only to narrow, relatively homogeneous groups of reactants. Herein, non-hybrid QSPR-based models were developed to predict PE-water (K_PE-w), PE-seawater (K_PE-sw), PVC-water (K_PVC-w) and PP-seawater (K_PP-sw) sorption coefficients at different temperatures, with eight machine learning algorithms. Moreover, novel hybrid intelligent models for predicting K_d more accurately were innovatively developed by applying GA, PSO and AdaBoost algorithms to optimize MLP and ELM models. The results indicated that all three optimization algorithms could improve the robustness and predictability of the standalone MLP and ELM models. In all models trained with K_PE-w, K_PE-sw, K_PVC-w and K_PP-sw data sets, GBDT-1 and XGBoost-1 models, MLP-GA-2 and MLP-PSO-2 models, MLR-3 and MLR-4 models performed better in terms of goodness of fit (R_adj²: 0.907-0.999), robustness (Q_BOOT²: 0.900-0.937) and predictability (R_ext²: 0.889-0.970), respectively. Analyzing the descriptors revealed that temperature, lipophilicity, ionization potential and molecular size were correlated closely with the adsorption capacity of microplastics to organic pollutants. The proposed QSPR models may assist in initial environmental exposure assessments without imposing heavy costs in the early experimental phase.

Principles and Procedures for Assessment of Acute Toxicity Incorporating In Silico Methods

Acute toxicity in silico models are being used to support an increasing number of application areas including (1) product research and development, (2) product approval and registration as well as (3) the transport, storage and handling of chemicals. The adoption of such models is being hindered, in part, because of a lack of guidance describing how to perform and document an in silico analysis. To address this issue, a framework for an acute toxicity hazard assessment is proposed. This framework combines results from different sources including in silico methods and in vitro or in vivo experiments. In silico methods that can assist the prediction of in vivo outcomes (i.e., LD50) are analyzed concluding that predictions obtained using in silico approaches are now well-suited for reliably supporting assessment of LD50-based acute toxicity for the purpose of GHS classification. A general overview is provided of the endpoints from in vitro studies commonly evaluated for predicting acute toxicity (e.g., cytotoxicity/cytolethality as well as assays targeting specific mechanisms). The increased understanding of pathways and key triggering mechanisms underlying toxicity and the increased availability of in vitro data allow for a shift away from assessments solely based on endpoints such as LD50, to mechanism-based endpoints that can be accurately assessed in vitro or by using in silico prediction models. This paper also highlights the importance of an expert review of all available information using weight-of-evidence considerations and illustrates, using a series of diverse practical use cases, how in silico approaches support the assessment of acute toxicity.

In Silico Models for Predicting Acute Systemic Toxicity

In this chapter, we give a brief overview of the regulatory requirements for acute systemic toxicity information in the European Union, and we review structure-based computational models that are available and potentially useful in the assessment of acute systemic toxicity. Emphasis is placed on quantitative structure-activity relationship (QSAR) models implemented by means of a range of software tools. The most recently published literature models for acute systemic toxicity are also discussed, and perspectives for future developments in this field are offered.

The Role of Natural Products as Sources of Therapeutic Agents for Innovative Drug Discovery

Emerging threats to human health require a concerted effort in search of both preventive and treatment strategies, placing natural products at the center of efforts to obtain new therapies and reduce disease spread and associated mortality. The therapeutic value of compounds found in plants has been known for ages, resulting in their utilization in homes and in clinics for the treatment of many ailments ranging from common headache to serious conditions such as wounds. Despite the advancement observed in the world, plant based medicines are still being used to treat many pathological conditions or are used as alternatives to modern medicines. In most cases, these natural products or plant-based medicines are used in an un-purified state as extracts. A lot of research is underway to identify and purify the active compounds responsible for the healing process. Some of the current drugs used in clinics have their origins as natural products or came from plant extracts. In addition, several synthetic analogues are natural product-based or plant-based. With the emergence of novel infectious agents such as the SARS-CoV-2 in addition to already burdensome diseases such as diabetes, cancer, tuberculosis and HIV/AIDS, there is need to come up with new drugs that can cure these conditions. Natural products offer an opportunity to discover new compounds that can be converted into drugs given their chemical structure diversity. Advances in analytical processes make drug discovery a multi-dimensional process involving computational designing and testing and eventual laboratory screening of potential drug candidates. Lead compounds will then be evaluated for safety, pharmacokinetics and efficacy. New technologies including Artificial Intelligence, better organ and tissue models such as organoids allow virtual screening, automation and high-throughput screening to be part of drug discovery. The use of bioinformatics and computation means that drug discovery can be a fast and efficient process and enable the use of natural products structures to obtain novel drugs. The removal of potential bottlenecks resulting in minimal false positive leads in drug development has enabled an efficient system of drug discovery. This review describes the biosynthesis and screening of natural products during drug discovery as well as methods used in studying natural products.

Artificial intelligence and machine learning assisted drug delivery for effective treatment of infectious diseases

In the era of antimicrobial resistance, the prevalence of multidrug-resistant microorganisms that resist conventional antibiotic treatment has steadily increased. Thus, it is now unquestionable that infectious diseases are significant global burdens that urgently require innovative treatment strategies. Emerging studies have demonstrated that artificial intelligence (AI) can transform drug delivery to promote effective treatment of infectious diseases. In this review, we propose to evaluate the significance, essential principles, and popular tools of AI in drug delivery for infectious disease treatment. Specifically, we will focus on the achievements and key findings of current research, as well as the applications of AI on drug delivery throughout the whole antimicrobial treatment process, with an emphasis on drug development, treatment regimen optimization, drug delivery system and administration route design, and drug delivery outcome prediction. To that end, the challenges of AI in drug delivery for infectious disease treatments and their current solutions and future perspective will be presented and discussed.

Are In Silico Approaches Applicable As a First Step for the Prediction of e-Liquid Toxicity in e-Cigarettes?

Recent studies have raised concerns about e-cigarette liquid inhalation toxicity by reporting the presence of chemicals with European Union CLP toxicity classification. In this scenario, the regulatory context is still developing and is not yet up to date with vaping current reality. Due to the paucity of toxicological studies, robust data regarding which components in e-liquids exhibit potential toxicities, are still inconsistent. In this study we applied computational methods for estimating the toxicity of poorly studied chemicals as a useful tool for predicting the acute toxicity of chemicals contained in e-liquids. The purpose of this study was 3-fold: (a) to provide a lower tier assessment of the potential health concerns associated with e-liquid ingredients, (b) to prioritize e-liquid ingredients by calculating the e-tox index, and (c) to estimate acute toxicity of e-liquid mixtures. QSAR models were generated using QSARINS software to fill the acute toxicity data gap of 264 e-liquid ingredients. As a second step, the potential acute toxicity of e-liquids mixtures was evaluated. Our preliminary data suggest that a computational approach may serve as a roadmap to enable regulatory bodies to better regulate e-liquid composition and to contribute to consumer health protection.

Application of cytochrome P450 reactivity on the characterization of chemical compounds and its association with repeated-dose toxicity

Repeated-dose toxicity (RDT) studies are one of the critical studies to assess chemical safety. There have been some studies attempting to predict RDT endpoints based on chemical substructures, but it remains very difficult to establish such a method, and a more detailed characterization of chemical compounds seems necessary. Cytochrome P450s (P450s) comprise multiple forms with different substrate specificities and play important roles in both the detoxification and metabolic activation of xenobiotics. In this study, we investigated possible use of P450 reactivity of chemical compounds to classify the compounds. A total of 148 compounds with available rat RDT test data were used as test compounds and subjected to inhibition assays against 18 human and rat P450s. Among the tested compounds, 82 compounds inhibited at least one P450 form. Hierarchical clustering analyses using the P450 inhibitory profiles divided the 82 compounds into nine groups, some of which showed characteristic chemical and biological properties. Principal component analyses of the P450 inhibition data in combination with the calculated chemical descriptors demonstrated that P450 inhibition data were plotted differently than most chemical descriptors in the loading plots. Finally, association analyses between P450 inhibition and RDT endpoints showed that some endpoints related to the liver, kidney and hematology were significantly associated with the inhibition of some P450s. Our present results suggest that the P450 reactivity profiles can be used as novel descriptors for characterizing chemical compounds for the investigation of the toxicity mechanism and/or the establishment of a toxicity prediction model.

Development of QSAR machine learning-based models to forecast the effect of substances on malignant melanoma cells

SK-MEL-5 is a human melanoma cell line that has been used in various studies to explore new therapies against melanoma in different in vitro experiments. Based on this study we report on the development of quantitative structure-activity relationship (QSAR) models able to predict the cytotoxic effect of diverse chemical compounds on this cancer cell line. The dataset of cytotoxic and inactive compounds were downloaded from the PubChem database. It contains the data for all chemical compounds for which cytotoxicity results expressed by GI₅₀ was recorded. In total 13 blocks of molecular descriptors were computed and used, after appropriate pre-processing in building QSAR models with four machine learning classifiers: Random forest (RF), gradient boosting, support vector machine and random k-nearest neighbors. Among the 186 models reported none had a positive predictive value (PPV) higher than 0.90 in both nested cross-validation and on an external dataset testing, but 7 models had a PPV higher than 0.85 in both evaluations, all seven using the RFs algorithm as a classifier, and topological descriptors, information indices, 2D-autocorrelation descriptors, P-VSA-like descriptors, and edge-adjacency descriptors as sets of features used for classification. The y-scrambling test was associated with considerably worse performance (confirming the non-random character of the models) and the applicability domain was assessed through three different methods.

Natural Products for Drug Discovery in the 21st Century: Innovations for Novel Drug Discovery

The therapeutic properties of plants have been recognised since time immemorial. Many pathological conditions have been treated using plant-derived medicines. These medicines are used as concoctions or concentrated plant extracts without isolation of active compounds. Modern medicine however, requires the isolation and purification of one or two active compounds. There are however a lot of global health challenges with diseases such as cancer, degenerative diseases, HIV/AIDS and diabetes, of which modern medicine is struggling to provide cures. Many times the isolation of "active compound" has made the compound ineffective. Drug discovery is a multidimensional problem requiring several parameters of both natural and synthetic compounds such as safety, pharmacokinetics and efficacy to be evaluated during drug candidate selection. The advent of latest technologies that enhance drug design hypotheses such as Artificial Intelligence, the use of 'organ-on chip' and microfluidics technologies, means that automation has become part of drug discovery. This has resulted in increased speed in drug discovery and evaluation of the safety, pharmacokinetics and efficacy of candidate compounds whilst allowing novel ways of drug design and synthesis based on natural compounds. Recent advances in analytical and computational techniques have opened new avenues to process complex natural products and to use their structures to derive new and innovative drugs. Indeed, we are in the era of computational molecular design, as applied to natural products. Predictive computational softwares have contributed to the discovery of molecular targets of natural products and their derivatives. In future the use of quantum computing, computational softwares and databases in modelling molecular interactions and predicting features and parameters needed for drug development, such as pharmacokinetic and pharmacodynamics, will result in few false positive leads in drug development. This review discusses plant-based natural product drug discovery and how innovative technologies play a role in next-generation drug discovery.

Machine learning in chemoinformatics and drug discovery

Chemoinformatics is an established discipline focusing on extracting, processing and extrapolating meaningful data from chemical structures. With the rapid explosion of chemical 'big' data from HTS and combinatorial synthesis, machine learning has become an indispensable tool for drug designers to mine chemical information from large compound databases to design drugs with important biological properties. To process the chemical data, we first reviewed multiple processing layers in the chemoinformatics pipeline followed by the introduction of commonly used machine learning models in drug discovery and QSAR analysis. Here, we present basic principles and recent case studies to demonstrate the utility of machine learning techniques in chemoinformatics analyses; and we discuss limitations and future directions to guide further development in this evolving field.

Evaluating Molecular Properties Involved in Transport of Small Molecules in Stratum Corneum: A Quantitative Structure-Activity Relationship for Skin Permeability

The skin permeability (Kp) defines the rate of a chemical penetrating across the stratum corneum. This value is widely used to quantitatively describe the transport of molecules in the outermost layer of epidermal skin and indicate the significance of skin absorption. This study defined a Kp quantitative structure-activity relationship (QSAR) based on 106 chemical substances of Kp measured using human skin and interpreted the molecular interactions underlying transport behavior of small molecules in the stratum corneum. The Kp QSAR developed in this study identified four molecular descriptors that described the molecular cyclicity in the molecule reflecting local geometrical environments, topological distances between pairs of oxygen and chlorine atoms, lipophilicity, and similarity to antineoplastics in molecular properties. This Kp QSAR considered the octanol-water partition coefficient to be a direct influence on transdermal movement of molecules. Moreover, the Kp QSAR identified a sub-domain of molecular properties initially defined to describe the antineoplastic resemblance of a compound as a significant factor in affecting transdermal permeation of solutes. This finding suggests that the influence of molecular size on the chemical’s skin-permeating capability should be interpreted with other relevant physicochemical properties rather than being represented by molecular weight alone.

Nontest Methods to Predict Acute Toxicity: State of the Art for Applications of In Silico Methods

The assessment of acute toxicity of chemicals by in silico methods is actually done by two methodologies, read-across and QSAR. The two approaches are strongly based on the similarity between the chemical for which a risk assessment is required and the reference chemical(s) for which the experimental data are known. Here, we describe the two methodologies with some main publications as illustrations and the in silico data associated with acute toxicity endpoints (ECHA, REACH) accessible via eChemPortal.

QSAR modeling in ecotoxicological risk assessment: application to the prediction of acute contact toxicity of pesticides on bees (Apis mellifera L.)

Despite their indisputable importance around the world, the pesticides can be dangerous for a range of species of ecological importance such as honeybees (Apis mellifera L.). Thus, a particular attention should be paid to their protection, not only for their ecological importance by contributing to the maintenance of wild plant diversity, but also for their economic value as honey producers and crop-pollinating agents. For all these reasons, the environmental protection requires the resort of risk assessment of pesticides. The goal of this work was therefore to develop a validated QSAR model to predict contact acute toxicity (LD₅₀) of 111 pesticides to bees because the QSAR models devoted to this species are very scarce. The analysis of the statistical parameters of this model and those published in the literature shows that our model is more efficient. The QSAR model was assessed according to the OECD principles for the validation of QSAR models. The calculated values for the internal and external validation statistic parameters (Q ² and [Formula: see text] are greater than 0.85. In addition to this validation, a mathematical equation derived from the ANN model was used to predict the LD₅₀ of 20 other pesticides. A good correlation between predicted and experimental values was found (R ² = 0.97 and RMSE = 0.14). As a result, this equation could be a means of predicting the toxicity of new pesticides.

Prediction of biodegradability of aromatics in water using QSAR modeling

The study was aimed at developing models for predicting the biodegradability of aromatic water pollutants. For that purpose, 36 single-benzene ring compounds, with different type, number and position of substituents, were used. The biodegradability was estimated according to the ratio of the biochemical (BOD₅) and chemical (COD) oxygen demand values determined for parent compounds ((BOD₅/COD)₀), as well as for their reaction mixtures in half-life achieved by UV-C/H₂O₂ process ((BOD₅/COD)_t1/2). The models correlating biodegradability and molecular structure characteristics of studied pollutants were derived using quantitative structure-activity relationship (QSAR) principles and tools. Upon derivation of the models and calibration on the training and subsequent testing on the test set, 3- and 5-variable models were selected as the most predictive for (BOD₅/COD)₀ and (BOD₅/COD)_t1/2, respectively, according to the values of statistical parameters R² and Q². Hence, 3-variable model predicting (BOD₅/COD)₀ possessed R²=0.863 and Q²=0.799 for training set, and R²=0.710 for test set, while 5-variable model predicting (BOD₅/COD)_1/2 possessed R²=0.886 and Q²=0.788 for training set, and R²=0.564 for test set. The selected models are interpretable and transparent, reflecting key structural features that influence targeted biodegradability and can be correlated with the degradation mechanisms of studied compounds by UV-C/H₂O₂.

Construction of Metabolism Prediction Models for CYP450 3A4, 2D6, and 2C9 Based on Microsomal Metabolic Reaction System

During the past decades, there have been continuous attempts in the prediction of metabolism mediated by cytochrome P450s (CYP450s) 3A4, 2D6, and 2C9. However, it has indeed remained a huge challenge to accurately predict the metabolism of xenobiotics mediated by these enzymes. To address this issue, microsomal metabolic reaction system (MMRS)—a novel concept, which integrates information about site of metabolism (SOM) and enzyme—was introduced. By incorporating the use of multiple feature selection (FS) techniques (ChiSquared (CHI), InfoGain (IG), GainRatio (GR), Relief) and hybrid classification procedures (Kstar, Bayes (BN), K-nearest neighbours (IBK), C4.5 decision tree (J48), RandomForest (RF), Support vector machines (SVM), AdaBoostM1, Bagging), metabolism prediction models were established based on metabolism data released by Sheridan et al. Four major biotransformations, including aliphatic C-hydroxylation, aromatic C-hydroxylation, N-dealkylation and O-dealkylation, were involved. For validation, the overall accuracies of all four biotransformations exceeded 0.95. For receiver operating characteristic (ROC) analysis, each of these models gave a significant area under curve (AUC) value >0.98. In addition, an external test was performed based on dataset published previously. As a result, 87.7% of the potential SOMs were correctly identified by our four models. In summary, four MMRS-based models were established, which can be used to predict the metabolism mediated by CYP3A4, 2D6, and 2C9 with high accuracy.

In Silico Models for Acute Systemic Toxicity

In this chapter, we give an overview of the regulatory requirements for acute systemic toxicity information in the European Union, and we review the availability of structure-based computational models that are available and potentially useful in the assessment of acute systemic toxicity. The most recently published literature models for acute systemic toxicity are also discussed, and perspectives for future developments in this field are offered.

A k-nearest neighbor classification of hERG K(+) channel blockers

A series of 172 molecular structures that block the hERG K⁺ channel were used to develop a classification model where, initially, eight types of PaDEL fingerprints were used for k-nearest neighbor model development. A consensus model constructed using Extended-CDK, PubChem and Substructure count fingerprint-based models was found to be a robust predictor of hERG activity. This consensus model demonstrated sensitivity and specificity values of 0.78 and 0.61 for the internal dataset compounds and 0.63 and 0.54 for the external (PubChem) dataset compounds, respectively. This model has identified the highest number of true positives (i.e. 140) from the PubChem dataset so far, as compared to other published models, and can potentially serve as a basis for the prediction of hERG active compounds. Validating this model against FDA-withdrawn substances indicated that it may even be useful for differentiating between mechanisms underlying QT prolongation.

Acute Toxicity-Supported Chronic Toxicity Prediction: A k-Nearest Neighbor Coupled Read-Across Strategy

A k-nearest neighbor (k-NN) classification model was constructed for 118 RDT NEDO (Repeated Dose Toxicity New Energy and industrial technology Development Organization; currently known as the Hazard Evaluation Support System (HESS)) database chemicals, employing two acute toxicity (LD₅₀)-based classes as a response and using a series of eight PaDEL software-derived fingerprints as predictor variables. A model developed using Estate type fingerprints correctly predicted the LD₅₀ classes for 70 of 94 training set chemicals and 19 of 24 test set chemicals. An individual category was formed for each of the chemicals by extracting its corresponding k-analogs that were identified by k-NN classification. These categories were used to perform the read-across study for prediction of the chronic toxicity, i.e., Lowest Observed Effect Levels (LOEL). We have successfully predicted the LOELs of 54 of 70 training set chemicals (77%) and 14 of 19 test set chemicals (74%) to within an order of magnitude from their experimental LOEL values. Given the success thus far, we conclude that if the k-NN model predicts LD₅₀ classes correctly for a certain chemical, then the k-analogs of such a chemical can be successfully used for data gap filling for the LOEL. This model should support the in silico prediction of repeated dose toxicity.