Structure-mutagenicity modelling using counter propagation neural networks

Computational identification of structural factors affecting the mutagenic potential of aromatic amines: study design and experimental validation

A grid-based, alignment-independent 3D-SDAR (three-dimensional spectral data-activity relationship) approach based on simulated ¹³C and ¹⁵N NMR chemical shifts augmented with through-space interatomic distances was used to model the mutagenicity of 554 primary and 419 secondary aromatic amines. A robust modeling strategy supported by extensive validation including randomized training/hold-out test set pairs, validation sets, "blind" external test sets as well as experimental validation was applied to avoid over-parameterization and build Organization for Economic Cooperation and Development (OECD 2004) compliant models. Based on an experimental validation set of 23 chemicals tested in a two-strain Salmonella typhimurium Ames assay, 3D-SDAR was able to achieve performance comparable to 5-strain (Ames) predictions by Lhasa Limited's Derek and Sarah Nexus for the same set. Furthermore, mapping of the most frequently occurring bins on the primary and secondary aromatic amine structures allowed the identification of molecular features that were associated either positively or negatively with mutagenicity. Prominent structural features found to enhance the mutagenic potential included: nitrobenzene moieties, conjugated π-systems, nitrothiophene groups, and aromatic hydroxylamine moieties. 3D-SDAR was also able to capture "true" negative contributions that are particularly difficult to detect through alternative methods. These include sulphonamide, acetamide, and other functional groups, which not only lack contributions to the overall mutagenic potential, but are known to actively lower it, if present in the chemical structures of what otherwise would be potential mutagens.

QSAR modeling for predicting mutagenic toxicity of diverse chemicals for regulatory purposes

The safety assessment process of chemicals requires information on their mutagenic potential. The experimental determination of mutagenicity of a large number of chemicals is tedious and time and cost intensive, thus compelling for alternative methods. We have established local and global QSAR models for discriminating low and high mutagenic compounds and predicting their mutagenic activity in a quantitative manner in Salmonella typhimurium (TA) bacterial strains (TA98 and TA100). The decision treeboost (DTB)-based classification QSAR models discriminated among two categories with accuracies of >96% and the regression QSAR models precisely predicted the mutagenic activity of diverse chemicals yielding high correlations (R ²) between the experimental and model-predicted values in the respective training (>0.96) and test (>0.94) sets. The test set root mean squared error (RMSE) and mean absolute error (MAE) values emphasized the usefulness of the developed models for predicting new compounds. Relevant structural features of diverse chemicals that were responsible and influence the mutagenic activity were identified. The applicability domains of the developed models were defined. The developed models can be used as tools for screening new chemicals for their mutagenicity assessment for regulatory purpose.

The index of ideality of correlation: A criterion of predictive potential of QSPR/QSAR models?

The index of ideality of correlation (IIC) is a new criterion of the predictive potential of quantitative structure-property/activity relationships (QSPRs/QSARs). This IIC is calculated with using of the correlation coefficient between experimental and calculated values of endpoint for the calibration set, with taking into account the positive and negative dispersions between experimental and calculated values. The mutagenicity is well-known important characteristic of substances from ecological point of view. Consequently, the estimation of the IIC for mutagenicity is well motivated. It is confirmed that the utilization of this criterion significantly improves the predictive potential of QSAR models of mutagenicity. The new criterion can be used for other endpoints.

The way to cover prediction for cytotoxicity for all existing nano-sized metal oxides by using neural network method

The regulatory agencies should fulfil the data gap in toxicity for new chemicals including nano-sized compounds, like metal oxides nanoparticles (MeO_x NPs) according to the registration, evaluation, authorisation and restriction of chemicals (REACH) legislation policy. This study demonstrates the perspective capability of neural network models for prediction of cytotoxicity of MeO_x NPs to bacteria Escherichia coli (E. coli) for the widest range of metal oxides extracted from Periodic table. The counter propagation artificial neural network (CP ANN) models for prediction of cytotoxicity of MeO_x NPs for data sets of 17, 36 and 72 metal oxides were employed in the study. The cytotoxicity of studied metal oxide NPs was correlated with (i) χ-metal electronegativity (EN) by Pauling scale and composition of metal oxides characterised by (ii) number of metal atoms in oxide, (iii) number of oxygen atoms in oxide and (iv) charge of metal cation in oxide. The paper describes the models in context of five OECD principles of validation models accepted for regulatory use. The recommendations were done for the minimal number of cytotoxicity tests needs for evaluation of the large set of MeO_x with different oxidation states. The methodology is expected to be useful for potential hazard assessment of MeO_x NPs and prioritisation for further testing and risk assessment.

Comparison of criteria used to access carcinogenicity in CPANN QSAR models versus the knowledge-based expert system Toxtree

The primary goal of this study was to describe and compare the criteria used to assess carcinogenic activity. The statistically-based predictive quantitative structure-activity relationship (QSAR) models based on the counter propagation artificial neural network (CPANN) algorithm, and knowledge-based expert systems based on a decision tree structural alert (SA) approach (Toxtree application), were considered. The integration of the QSAR (CPANN models) and SAR (Toxtree SA application) approach contributed to the mechanistic understanding of the QSAR model considered. The mapping technique inherent to CPANN Kohonen enables us to relate the similarities or dissimilarities within a congeneric set of chemicals with particular SAs for carcinogenicity. The focus of our investigations was the similarities and dissimilarities of the features used in the QSAR and SAR methods. Due to the complexity of the carcinogenic endpoint, the integration of different approaches allows the models to be improved and provides a valuable technique for evaluating the safety of chemicals.

Integration of QSAR and SAR methods for the mechanistic interpretation of predictive models for carcinogenicity

The knowledge-based Toxtree expert system (SAR approach) was integrated with the statistically based counter propagation artificial neural network (CP ANN) model (QSAR approach) to contribute to a better mechanistic understanding of a carcinogenicity model for non-congeneric chemicals using Dragon descriptors and carcinogenic potency for rats as a response. The transparency of the CP ANN algorithm was demonstrated using intrinsic mapping technique specifically Kohonen maps. Chemical structures were represented by Dragon descriptors that express the structural and electronic features of molecules such as their shape and electronic surrounding related to reactivity of molecules. It was illustrated how the descriptors are correlated with particular structural alerts (SAs) for carcinogenicity with recognized mechanistic link to carcinogenic activity. Moreover, the Kohonen mapping technique enables one to examine the separation of carcinogens and non-carcinogens (for rats) within a family of chemicals with a particular SA for carcinogenicity. The mechanistic interpretation of models is important for the evaluation of safety of chemicals.

Integrated decision-tree testing strategies for mutagenicity and carcinogenicity with respect to the requirements of the EU REACH legislation

Liverpool John Moores University and FRAME recently conducted a research project sponsored by Defra, on the status of alternatives to animal testing with regard to the European Union REACH (Registration, Evaluation and Authorisation of Chemicals) system for the safety testing and risk assessment of chemicals. The project covered all the main toxicity endpoints associated with the REACH system. This paper focuses on the prospects for using alternative methods (both in vitro and in silico) for mutagenicity (genotoxicity) and carcinogenicity testing--two toxicity endpoints, which, together with reproductive toxicity, are of pivotal importance for the REACH system. The manuscript critically discusses well-established testing approaches, and in particular, the requirement for short-term in vivo tests for confirming positive mutagenicity, and the need for the rodent bioassay for detecting non-genotoxic carcinogens. Recently-proposed testing strategies focusing on non-animal approaches are also considered, and our own testing scheme is presented and supported with background information. This scheme makes maximum use of pre-existing data, computer (in silico) and in vitro methods, with weight-of-evidence assessments at each major stage. The need for the improvement of in vitro methods, to reduce the generation of false-positive results, is also discussed. Lastly, ways in which reduction and refinement measures can be used are also considered, and some recommendations are made for future research to facilitate the implementation of the proposed testing scheme.

Anticancer activity of selected phenolic compounds: QSAR studies using ridge regression and neural networks

Phenol and its congeners are known to induce caspase-mediated apoptosis activity and cytotoxicity on various cancer cell lines. Apoptosis, scavenging of radicals, antioxidant, and pro-oxidant characteristics are primarily responsible for the antitumor activities of phenolic compounds. Quantitative structure-activity relationship studies on the cellular apoptosis and cytotoxicity of phenolic compounds have been investigated recently by Selassie and colleagues (J Med Chem; 48:7234, 2005) wherein models were developed for various carcinogenic cell lines. These quantitative structure-activity relationship models are based on few experimentally obtained physicochemical parameters such as Verloop's sterimol descriptor, hydrophobicity, Hammett electronic parameter, and octanol/water partition coefficient. The paper deals with structure-activity relationships of phenols and its derivatives for the development of predictive models from the standpoint of theoretical structural parameters and ridge regression methodology. The quantitative structure-activity relationship studies developed here for the caspase-mediated apoptosis activity and cytotoxicity on murine leukemia cell line (L1210), human promylolytic cell line (HL-60), human breast cancer cell line (MCF-7), parenteral human acute lymphoblastic cells (CCRF-CEM), and multidrug-resistant subline of CCRF-resistant to vinblastine (CEM/VLB) cells utilize physicochemical molecular descriptors calculated solely from the structure of phenolic compounds under investigation along with the descriptors used by Selassie and group. It is seen that such quantitative structure-activity relationships can provide a better quality predictive model for the phenolic compounds. The biological activities of the nine sets of phenolic compounds have been calculated based on ridge regression analysis that clearly gives a better significant correlation compared to the activities predicted by Selassie and co-workers. Counter-propagation artificial neural network studies have been introduced in the present investigation for a better understanding of multidimensional rational patterns in more complex data sets. The counter-propagation artificial neural network studies were performed on the same data set and with the same descriptors as have been carried out in developing ridge regression models and the result of counter-propagation neural network models produces very interesting findings in terms of leave-one-out test. Finally, an attempt has been made for a comparative study of the relative effectiveness of linear statistical methods versus nonlinear techniques, such as counter-propagation neural networks in modeling structure-activity studies of the phenolic compounds.

Proposed integrated decision-tree testing strategies for mutagenicity and carcinogenicity in relation to the EU REACH legislation

Liverpool John Moores University and FRAME recently conducted a research project sponsored by Defra, on the status of alternatives to animal testing with regard to the European Union REACH (Registration, Evaluation and Authorisation of Chemicals) system for the safety testing and risk assessment of chemicals. The project covered all the main toxicity endpoints associated with the REACH system. This paper focuses on the prospects for using alternative methods (both in vitro and in silico) for mutagenicity (genotoxicity) and carcinogenicity testing - two toxicity endpoints, which, together with reproductive toxicity, are of pivotal importance for the REACH system. The manuscript critically discusses well-established testing approaches, and in particular, the requirement for short-term in vivo tests for confirming positive mutagenicity, and the need for the rodent bioassay for detecting non-genotoxic carcinogens. Recently-proposed testing strategies focusing on non-animal approaches are also considered, and our own testing scheme is presented and supported with background information. This scheme makes maximum use of pre-existing data, computer (in silico) and in vitro methods, with weight-of-evidence assessments at each major stage. The need for the improvement of in vitro methods, to reduce the generation of false-positive results, is also discussed. Lastly, ways in which reduction and refinement measures can be used are also considered, and some recommendations are made for future research to facilitate the implementation of the proposed testing scheme.

Biopartitioning micellar chromatography to predict mutagenicity of aromatic amines

Mutagenicity is a toxicity endpoint associated with the chronic exposure to chemicals. Aromatic amines have considerable industrial and environmental importance due to their widespread use in industry and their mutagenic capacity. Biopartitioning micellar chromatography (BMC), a mode of micellar liquid chromatography that uses micellar mobile phases of Brij35 in adequate experimental conditions, has demonstrated to be useful in mimicking the drug partitioning process into biological systems. In this paper, the usefulness of BMC for predicting mutagenicity of aromatic amines is demonstrated. A multiple linear regression (MLR) model based on BMC retention data is proposed and compared with other ones reported in bibliography. The proposed model present better or similar descriptive and predictive capability.

Prediction of tissue: air partition coefficients--theoretical vs. experimental methods

Predictive QSAR models for rat and human tissue : air partition coefficients, namely blood : air, fat : air, brain : air, liver : air, muscle : air, and kidney : air were developed utilizing experimentally determined partition coefficients for 131 chemicals obtained from the literature and molecular descriptors based solely on chemical structure. The descriptors were partitioned into four hierarchical classes, including topostructural, topochemical, 3-dimensional, and ab initio quantum chemical. Three types of regression methodologies--ridge regression, principal components regression, and partial least squares regression--were used comparatively in the development of the structure-based models. In addition to the structure-based models, ordinary least squares regression was used to develop comparative models based on experimentally determined properties including saline : air and olive oil : air partition coefficients. The results of the study indicate that many of the structure-based models are comparable or superior to their respective property-based models. This is an important result considering that structural descriptors can be calculated quickly and inexpensively for both existing chemicals and those not yet synthesized. It was also found that ridge regression outperformed principal components regression and partial least squares regression, with respect to the structure-based models, and that generally the topochemical descriptors alone produced models of good predictive ability.

Validation of counter propagation neural network models for predictive toxicology according to the OECD principles: a case study

The OECD has proposed five principles for validation of QSAR models used for regulatory purposes. Here we present a case study investigating how these principles can be applied to models based on Kohonen and counter propagation neural networks. The study is based on a counter propagation network model that has been built using toxicity data in fish fathead minnow for 541 compounds. The study demonstrates that most, if not all, of the OECD criteria may be met when modeling using this neural network approach.