The second National Toxicology Program comparative exercise on the prediction of rodent carcinogenicity: definitive results

Chemical carcinogenicity has been the target of a large array of attempts to create alternative predictive models, ranging from short-term biological assays (e.g. mutagenicity tests) to theoretical models. Among the theoretical models, the application of the science of structure-activity relationships (SAR) has earned special prominence. A crucial element is the independent evaluation of the predictive ability. In the past decade, there have been two fundamental comparative exercises on the prediction of chemical carcinogenicity, held under the aegis to the US National Toxicology Program (NTP). In both exercises, the predictions were published before the animal data were known, thus using a most stringent criterion of predictivity. We analyzed the results of the first comparative exercise in a previous paper [Mutat. Res. 387 (1997) 35]; here, we present the complete results of the second exercise, and we analyze and compare the prediction sets. The range of accuracy values was quite large: the systems that performed best in this prediction exercise were in the range 60-65% accuracy. They included various human experts approaches (e.g. Oncologic) and biologically based approaches (e.g. the experimental transformation assay in Syrian hamster embryo (SHE) cells). The main difficulty for the structure-activity relationship-based approaches was the discrimination between real carcinogens, and non-carcinogens containing structural alerts (SA) for genotoxic carcinogenicity. It is shown that the use of quantitative structure-activity relationship models, when possible, can contribute to overcome the above problem. Overall, given the uncertainty linked to the predictions, the predictions for the individual chemicals cannot be taken at face value; however, the general level of knowledge available today (especially for genotoxic carcinogens) allows qualified human experts to operate a very efficient priority setting of large sets of chemicals.     

Classification according to chemical structure, mutagenicity to Salmonella and level of carcinogenicity of a further 42 chemicals tested for carcinogenicity by the U.S. National Toxicology Program

This paper is an extension and update of an earlier review published in this journal (Ashby and Tennant, 1988). A summary of the rodent carcinogenicity bioassay data on a further 42 chemicals tested by the U.S. National Toxicology Program (NTP) is presented. An evaluation of each chemical for structural alerts to DNA-reactivity is also provided, together with a summary of its mutagenicity to Salmonella. The 42 chemicals were numbered and evaluated as an extension of the earlier analysis of 222 NTP chemicals. The activity patterns and conclusions derived from the earlier study remain unchanged for the larger group of 264 chemicals. Based on the extended database of 264 NTP chemicals, the sensitivity of the Salmonella assay for rodent carcinogens is 58% and the specificity for the non-carcinogens is 73%. A total of 32 chemicals were defined as equivocal for carcinogenicity and, of these, 11 (34%) are mutagenic to Salmonella. An evaluation is made of instances where predictions of carcinogenicity, based on structural alerts, disagree with the Salmonella mutagenicity result (12% of the database). The majority of the disagreements are for structural alerts on non-mutagens, and that places these alerts as a sensitive primary screen with a specificity lower than that of the Salmonella assay. That analysis indicates some need for assays complementary to the Salmonella test when screening for potential genotoxic carcinogens. It also reveals that the correlation between structural alerts and mutagenicity to Salmonella is probably greater than 90%. Chemicals predicted to show Michael-type alkylating activity (i.e., CH2 = CHX; where X = an electron-withdrawing group, e.g. acrylamide) have been confirmed as a structural alert, and the halomethanes (624 are possible) have been classified as structurally-alerting. To this end an extended carcinogen-alert model structure is presented. Among the 138 NTP carcinogens now reviewed, 45 (33%) are non-mutagenic to Salmonella and possess a chemical structure that does not alert to DNA-reactivity. These carcinogens therefore either illustrate the need for complementary genetic screening tests to the Salmonella assay, or they represent the group of non-genotoxic carcinogens referred to most specifically by Weisburger and Williams (1981); the latter concept is favoured.     

Definitive relationships among chemical structure, carcinogenicity and mutagenicity for 301 chemicals tested by the U.S. NTP.

An analysis is presented in which are evaluated correlations among chemical structure, mutagenicity to Salmonella, and carcinogenicity to rats and mice among 301 chemicals tested by the U.S. NTP. Overall, there was a high correlation between structural alerts to DNA reactivity and mutagenicity, but the correlation of either property with carcinogenicity was low. If rodent carcinogenicity is regarded as a singular property of chemicals, then neither structural alerts nor mutagenicity to Salmonella are effective in its prediction. Given this, the database was fragmented and new correlations sought between the derived sub-groups. First, the 301 chemicals were segregated into six broad chemical groupings. Second, the rodent cancer data were partially segregated by target tissue. Using the previously assigned structural alerts to DNA reactivity (electrophilicity), the chemicals were split into 154 alerting chemicals and 147 non-alerting chemicals. The alerting chemicals were split into three chemical groups; aromatic amino/nitro-types, alkylating agents and miscellaneous structurally-alerting groups. The non-alerting chemicals were subjectively split into three broad categories; non-alerting, non-alerting containing a non-reactive halogen group, and non-alerting chemical with minor concerns about a possible structural alert. The tumor data for all 301 chemicals are re-presented according to these six chemical groupings. The most significant findings to emerge from comparisons among these six groups of chemicals were as follows: (a) Most of the rodent carcinogens, including most of the 2-species and/or multiple site carcinogens, were among the structurally alerting chemicals. (b) Most of the structurally alerting chemicals were mutagenic; 84% of the carcinogens and 66% of the non-carcinogens. 100% of the 33 aromatic amino/nitro-type 2-species carcinogens were mutagenic. Thus, for structurally alerting chemicals, the Salmonella assay showed high sensitivity and low specificity (0.84 and 0.33, respectively). (c) Among the 147 non-alerting chemicals less than 5% were mutagenic, whether they were carcinogens or non-carcinogens (sensitivity 0.04).     

Structure-based methods for predicting mutagenicity and carcinogenicity: are we there yet?

There is a great deal of current interest in the use of commercial, automated programs for the prediction of mutagenicity and carcinogenicity based on chemical structure. However, the goal of accurate and reliable toxicity prediction for any chemical, based solely on structural information remains elusive. The toxicity prediction challenge is global in its objective, but limited in its solution, to within local domains of chemicals acting according to similar mechanisms of action in the biological system; to predict, we must be able to generalize based on chemical structure, but the biology fundamentally limits our ability to do so. Available commercial systems for mutagenicity and/or carcinogenicity prediction differ in their specifics, yet most fall in two major categories: (1) automated approaches that rely on the use of statistics for extracting correlations between structure and activity; and (2) knowledge-based expert systems that rely on a set of programmed rules distilled from available knowledge and human expert judgement. These two categories of approaches differ in the ways that they represent, process, and generalize chemical-biological activity information. An application of four commercial systems (TOPKAT, CASE/MULTI-CASE, DEREK, and OncoLogic) to mutagenicity and carcinogenicity prediction for a particular class of chemicals—the haloacetic acids (HAs)—is presented to highlight these differences. Some discussion is devoted to the issue of gauging the relative performance of commercial prediction systems, as well as to the role of prospective prediction exercises in this effort. And finally, an alternative approach that stops short of delivering a prediction to a user, involving structure-searching and data base exploration, is briefly considered.     

Computer-aided rodent carcinogenicity prediction

The potential of the computer program PASS (Prediction Activity Spectra for Substances) to predict rodent carcinogenicity for chemical compounds was studied. PASS predicts carcinogenicity of chemical compounds on the basis of their structural formula and of structure-activity relationship analysis of known carcinogens and non-carcinogens. The data on structures and experimental results of 2-year carcinogenicity assays for 412 chemicals from the NTP (National Toxicological Program) and 1190 chemicals from the CPDB (Carcinogenic Potency Database) were used in our study. The predictions take into consideration information about species and sex of animals. For evaluation of the predictive accuracy we used two procedures: leave-one-out cross-validation (LOO CV) and leave-20%-out cross-validation. In the last case we randomly divided the studied data set 20 times into two subsets. The data from the first subset, containing 80% of the compounds, were added to the PASS training set (which includes about 46,000 compounds with about 1500 biological activity types collected during the last 20 years to predict biological activity spectra), the second subset with 20% of the compounds was used as an evaluation set. The mean accuracy of prediction calculated by LOO CV is about 73% for NTP compounds in the 'equivocal' category of carcinogenic activity and 80% for NTP compounds in the 'evidence' category of carcinogenicity. The mean accuracy of prediction for the CPDB database is 89.9% calculated by LOO CV and 63.4% calculated by leave-20%-out cross-validation. Influence of incorporation of species and sex data on the accuracy of carcinogenicity prediction was also investigated. It was shown that the accuracy was increased only for data on male animals.     

Carcinogenicity of the aromatic amines: from structure-activity relationships to mechanisms of action and risk assessment

Aromatic amines represent one of the most important classes of industrial and environmental chemicals: many of them have been reported to be powerful carcinogens and mutagens, and/or hemotoxicants. Their toxicity has been studied also with quantitative structure-activity relationship (QSAR) methods: these studies are potentially suitable for investigating mechanisms of action and for estimating the toxicity of compounds lacking experimental determinations. In this paper, we first summarized the QSAR models for the rodent carcinogenicity of the aromatic amines. The gradation of potency of the carcinogenic amines depended firstly on their hydrophobicity, and secondly on electronic (reactivity, propensity to be metabolically transformed) and steric properties. On the contrary, the difference between carcinogenic and non-carcinogenic aromatic amines depended mainly on electronic and steric properties. These QSARs can be used directly for estimating the carcinogenicity of aromatic amines. A two-step prediction is possible: (1) estimation of yes/no activity; (2) if the answer from step 1 is yes, then prediction of the degree of potency. The QSARs for rodent carcinogenicity were put in a wider context by comparing them with those for: (a) Salmonella mutagenicity; (b) general toxicity; (c) enzymatic reactions; (d) physical-chemical reactions. This comparative QSAR exercise generated a coherent global picture of the action mechanisms of the aromatic amines. The QSARs for carcinogenicity were similar to those for Salmonella mutagenicity, thus pointing to a similar mechanism of action. On the contrary, the general toxicity QSARs (both in vitro and in vivo systems) were mostly based on hydrophobicity, pointing to an aspecific mechanism of action much simpler than that for carcinogenicity and mutagenicity. The oxidation of the amines (first step in the main metabolic pathway leading to carcinogenic and mutagenic species) had identical QSARs in both enzymatic and physical-chemical systems, thus providing evidence for the link between simple chemical reactions and those in biological systems. The results show that it is possible to generate mechanistically and statistically sound QSAR models for rodent carcinogenicity, and indirectly that the rodent bioassay is a reliable source of good quality data.     

Genomic Models of Short-Term Exposure Accurately Predict Long-Term Chemical Carcinogenicity and Identify Putative Mechanisms of Action

Background

Despite an overall decrease in incidence of and mortality from cancer, about 40% of Americans will be diagnosed with the disease in their lifetime, and around 20% will die of it. Current approaches to test carcinogenic chemicals adopt the 2-year rodent bioassay, which is costly and time-consuming. As a result, fewer than 2% of the chemicals on the market have actually been tested. However, evidence accumulated to date suggests that gene expression profiles from model organisms exposed to chemical compounds reflect underlying mechanisms of action, and that these toxicogenomic models could be used in the prediction of chemical carcinogenicity.

Results

In this study, we used a rat-based microarray dataset from the NTP DrugMatrix Database to test the ability of toxicogenomics to model carcinogenicity. We analyzed 1,221 gene-expression profiles obtained from rats treated with 127 well-characterized compounds, including genotoxic and non-genotoxic carcinogens. We built a classifier that predicts a chemical''s carcinogenic potential with an AUC of 0.78, and validated it on an independent dataset from the Japanese Toxicogenomics Project consisting of 2,065 profiles from 72 compounds. Finally, we identified differentially expressed genes associated with chemical carcinogenesis, and developed novel data-driven approaches for the molecular characterization of the response to chemical stressors.

Conclusion

Here, we validate a toxicogenomic approach to predict carcinogenicity and provide strong evidence that, with a larger set of compounds, we should be able to improve the sensitivity and specificity of the predictions. We found that the prediction of carcinogenicity is tissue-dependent and that the results also confirm and expand upon previous studies implicating DNA damage, the peroxisome proliferator-activated receptor, the aryl hydrocarbon receptor, and regenerative pathology in the response to carcinogen exposure.     

A survey of the Predictive Toxicology Challenge 2000-2001

MOTIVATION: The Predictive Toxicology Challenge (PTC) was initiated to stimulate the development of advanced techniques for predictive toxicology models. The goal of this challenge was to compare different approaches for the prediction of rodent carcinogenicity, based on the experimental results of the US National Toxicology Program (NTP). RESULTS: 111 sets of predictions for 185 compounds have been evaluated on quantitative and qualitative scales to select the most predictive models and those with the highest toxicological relevance. The accuracy of the submitted predictions was between 25 and 79 %. An evaluation of the most accurate models by toxicological experts showed, that it is still hard for domain experts to interpret the submitted models and to put them into relation with toxicological knowledge. AVAILABILITY: PTC details and data can be found at: http://www.predictive-toxicology.org/ptc/.     

Systematic Artifacts in Support Vector Regression-Based Compound Potency Prediction Revealed by Statistical and Activity Landscape Analysis

Support vector machines are a popular machine learning method for many classification tasks in biology and chemistry. In addition, the support vector regression (SVR) variant is widely used for numerical property predictions. In chemoinformatics and pharmaceutical research, SVR has become the probably most popular approach for modeling of non-linear structure-activity relationships (SARs) and predicting compound potency values. Herein, we have systematically generated and analyzed SVR prediction models for a variety of compound data sets with different SAR characteristics. Although these SVR models were accurate on the basis of global prediction statistics and not prone to overfitting, they were found to consistently mispredict highly potent compounds. Hence, in regions of local SAR discontinuity, SVR prediction models displayed clear limitations. Compared to observed activity landscapes of compound data sets, landscapes generated on the basis of SVR potency predictions were partly flattened and activity cliff information was lost. Taken together, these findings have implications for practical SVR applications. In particular, prospective SVR-based potency predictions should be considered with caution because artificially low predictions are very likely for highly potent candidate compounds, the most important prediction targets.     

Classification according to chemical structure, mutagenicity to Salmonella and level of carcinogenicity of a further 39 chemicals tested for carcinogenicity by the U.S. National Toxicology Program

This paper is an extension of compilations published previously in this journal. (Ashby and Tennant, 1988; Ashby et al., 1989). A summary of the rodent carcinogenicity bioassay data on a further 39 chemicals tested by the U.S. National Toxicology Program (NTP) is presented. An evaluation of each chemical for structural alerts to DNA-reactivity is also provided, together with a summary of its mutagenicity to Salmonella. Chemicals with an aliphatic nitro group (-C-NO2) have been added to the composite structure of DNA-reactive sub-groups. The 39 chemicals were numbered and evaluated as an extension of the earlier analysis of 264 NTP chemicals. The activity patterns and conclusions derived from the earlier studies are followed by these 39 chemicals, albeit a detailed analysis of the total database of 301 chemicals is reserved for the succeeding paper.     

Distributed structure-searchable toxicity (DSSTox) public database network: a proposal.

The ability to assess the potential genotoxicity, carcinogenicity, or other toxicity of pharmaceutical or industrial chemicals based on chemical structure information is a highly coveted and shared goal of varied academic, commercial, and government regulatory groups. These diverse interests often employ different approaches and have different criteria and use for toxicity assessments, but they share a need for unrestricted access to existing public toxicity data linked with chemical structure information. Currently, there exists no central repository of toxicity information, commercial or public, that adequately meets the data requirements for flexible analogue searching, Structure-Activity Relationship (SAR) model development, or building of chemical relational databases (CRD). The distributed structure-searchable toxicity (DSSTox) public database network is being proposed as a community-supported, web-based effort to address these shared needs of the SAR and toxicology communities. The DSSTox project has the following major elements: (1) to adopt and encourage the use of a common standard file format (structure data file (SDF)) for public toxicity databases that includes chemical structure, text and property information, and that can easily be imported into available CRD applications; (2) to implement a distributed source approach, managed by a DSSTox Central Website, that will enable decentralized, free public access to structure-toxicity data files, and that will effectively link knowledgeable toxicity data sources with potential users of these data from other disciplines (such as chemistry, modeling, and computer science); and (3) to engage public/commercial/academic/industry groups in contributing to and expanding this community-wide, public data sharing and distribution effort. The DSSTox project's overall aims are to effect the closer association of chemical structure information with existing toxicity data, and to promote and facilitate structure-based exploration of these data within a common chemistry-based framework that spans toxicological disciplines.     

Putting the Predictive Toxicology Challenge into perspective: reflections on the results

MOTIVATION: Chemical carcinogenicity is of primary interest, because it drives much of the current regulatory actions regarding new and existing chemicals, and its experimental determination involves time-consuming and expensive animal testing. Both academia and private companies are actively trying to develop SAR and QSAR models. This paper reviews the new Predictive Toxicology Challenge (PTC) results, by putting them into the context of previous attempts. RESULTS: A marked dependency of the prediction ability of the different algorithms on the training sets was observed, pointing to a still insufficient coverage of the chemical carcinogens 'universe'. A theoretical treatment of the possible developments of the Artificial Intelligence approaches is sketched.     

The unique role of rodents in the detection of possible human carcinogens and mutagens

Some of the probable reasons underlying the observation that not all chemicals shown to be genotoxic in vitro are capable of eliciting tumours in rodents or humans are discussed using appropriate examples. It is suggested that a substantial proportion of the resources currently available for conducting rodent carcinogenicity bioassays should be employed in the short-term evaluation in vivo of some of the many hundreds of chemicals recently defined as genotoxic in vitro, rather than in the protracted evaluation of a few chemicals, often of unknown activity in vitro, for carcinogenicity. A decision tree approach to the evaluation of chemicals for human mutagenic/carcinogenic potential is presented which is at variance with the construction and philosophy of many of the current legislative guidelines. The immediate need for the adoption of one of the available short-term in vivo liver assays, and/or the development of a short-term in vivo rodent assay capable of concomitantly monitoring different genetic end-points in a range of organs or tissues is emphasized.     

Evaluation of the ability of a battery of three in vitro genotoxicity tests to discriminate rodent carcinogens and non-carcinogens I. Sensitivity, specificity and relative predictivity

The performance of a battery of three of the most commonly used in vitro genotoxicity tests--Ames+mouse lymphoma assay (MLA)+in vitro micronucleus (MN) or chromosomal aberrations (CA) test--has been evaluated for its ability to discriminate rodent carcinogens and non-carcinogens, from a large database of over 700 chemicals compiled from the CPDB ("Gold"), NTP, IARC and other publications. We re-evaluated many (113 MLA and 30 CA) previously published genotoxicity results in order to categorise the performance of these assays using the response categories we established. The sensitivity of the three-test battery was high. Of the 553 carcinogens for which there were valid genotoxicity data, 93% of the rodent carcinogens evaluated in at least one assay gave positive results in at least one of the three tests. Combinations of two and three test systems had greater sensitivity than individual tests resulting in sensitivities of around 90% or more, depending on test combination. Only 19 carcinogens (out of 206 tested in all three tests, considering CA and MN as alternatives) gave consistently negative results in a full three-test battery. Most were either carcinogenic via a non-genotoxic mechanism (liver enzyme inducers, peroxisome proliferators, hormonal carcinogens) considered not necessarily relevant for humans, or were extremely weak (presumed) genotoxic carcinogens (e.g. N-nitrosodiphenylamine). Two carcinogens (5-chloro-o-toluidine, 1,1,2,2-tetrachloroethane) may have a genotoxic element to their carcinogenicity and may have been expected to produce positive results somewhere in the battery. We identified 183 chemicals that were non-carcinogenic after testing in both male and female rats and mice. There were genotoxicity data on 177 of these. The specificity of the Ames test was reasonable (73.9%), but all mammalian cell tests had very low specificity (i.e. below 45%), and this declined to extremely low levels in combinations of two and three test systems. When all three tests were performed, 75-95% of non-carcinogens gave positive (i.e. false positive) results in at least one test in the battery. The extremely low specificity highlights the importance of understanding the mechanism by which genotoxicity may be induced (whether it is relevant for the whole animal or human) and using weight of evidence approaches to assess the carcinogenic risk from a positive genotoxicity signal. It also highlights deficiencies in the current prediction from and understanding of such in vitro results for the in vivo situation. It may even signal the need for either a reassessment of the conditions and criteria for positive results (cytotoxicity, solubility, etc.) or the development and use of a completely new set of in vitro tests (e.g. mutation in transgenic cell lines, systems with inherent metabolic activity avoiding the use of S9, measurement of genetic changes in more cancer-relevant genes or hotspots of genes, etc.). It was very difficult to assess the performance of the in vitro MN test, particularly in combination with other assays, because the published database for this assay is relatively small at this time. The specificity values for the in vitro MN assay may improve if data from a larger proportion of the known non-carcinogens becomes available, and a larger published database of results with the MN assay is urgently needed if this test is to be appreciated for regulatory use. However, specificity levels of <50% will still be unacceptable. Despite these issues, by adopting a relative predictivity (RP) measure (ratio of real:false results), it was possible to establish that positive results in all three tests indicate the chemical is greater than three times more likely to be a rodent carcinogen than a non-carcinogen. Likewise, negative results in all three tests indicate the chemical is greater than two times more likely to be a rodent non-carcinogen than a carcinogen. This RP measure is considered a useful tool for industry to assess the likelihood of a chemical possessing carcinogenic potential from batteries of positive or negative results.     

Global pattern and local variation in species–area relationships

Aim We conducted a meta‐analysis of species–area relationships (SARs) by combining several data sets and important covariates such as types of islands, taxonomic groups, latitude and spatial extent, in a hierarchical model framework to study global pattern and local variation in SARs and its consequences for prediction. Location One thousand nine hundred and eighteen islands from 94 SAR studies from around the world. Methods We developed a generalization of the power‐law SAR model, the HSARX model, which allows: (1) the inclusion of multiple focal parameters (intercept, slope, within‐study variance), (2) use of multiple effect modifiers based on a collection of SAR studies, and (3) modelling of the between‐ and within‐study variability. Results The global pattern in the SAR was the average of local SARs and had wide confidence intervals. The global SAR slope was 0.228 with 90% confidence limits of 0.059 and 0.412. The intercept, slope and within‐study variability of local SARs showed great heterogeneity as a result of the interaction of modifying covariates. Confidence intervals for these SAR parameters were narrower when other covariates in addition to area were accounted for, thus increasing the accuracy of the predictions for species richness. The significant effect of latitude and the interaction of latitude, taxa and island type on the SAR slope indicated that the ‘typical’ latitudinal diversity gradient can be reversed in isolated systems. Main conclusions The power‐law relationship underlying the HSARX model provides a good fit for non‐nested SARs across vastly different spatial scales by taking into account other covariates. The HSARX framework allows researchers to explore the complex interactions among SAR parameters and modifying variables, to explicitly study the scale dependence, and to make robust predictions on multiple levels (island, study, global) with associated prediction intervals. From a prediction perspective, it is not the global pattern but the local variation that matters.     

Thalidomide and metabolites: indications of the absence of 'genotoxic' carcinogenic potentials

Because of the reintroduction into human therapeutics of thalidomide, a recognized developmental toxicant in humans, there has been concern about its potential for inducing other health effects as well. The present study is concerned with the possible mutagenicity and carcinogenicity of this chemical. Using the expert system, META, a series of putative metabolites of thalidomide was generated. In addition to the known or hypothesized metabolites of thalidomide (N=12), a number of additional putative metabolites (N=131) were identified by META. The structures of these chemicals were subjected to structure-activity analyses using predictive CASE/MULTICASE models of developmental toxicity, rodent carcinogenicity and mutagenicity in Salmonella. While thalidomide and some of its putative metabolites were predicted to be developmental toxicants, none of them were predicted to be rodent carcinogens. Putative metabolites containing the hydroxamic acid or hydroxylamine moieties were predicted to be mutagens. None of the 'known' metabolites of thalidomide contained these reactive moieties. Whether such intermediates are indeed generated or whether they are generated and are either unstable in the presence of oxygen or react rapidly with nucleophiles is unknown.     

Knowledge-based battery design of short-term tests based on dose information

A construction of batteries of short-term tests (STTs) is described which is based on a classification of 73 chemicals in regard to their carcinogenicity. The 73 chemicals were studied within the U.S. National Toxicology Program (Ashby and Tennant, 1988). The batteries are validated using the classification of 35 additional chemicals. They are defined by logically structured combinations of rules. The single rules are defined by the z-scores of the logarithmic values of the limiting doses obtained from the 4 in vitro STTs used in the study by Ashby and Tennant. The limiting dose is defined as the lowest effective dose or the highest ineffective dose (Waters et al., 1987). The batteries are constructed by minimizing the number of disagreements with the classification by Ashby and Tennant. Compared with the results obtained from single STTs, 2 batteries of 3 STTs have higher concordances with the carcinogenicity data, namely 70% for the NTP data and 74-77% for the independent test data. In addition, a theoretical result shows that the proposed battery design, for a large enough learning set of chemicals, leads to results which are replicated with high probability on a large enough validation set. Based on the first results obtained with a limited number of chemicals it is concluded that the knowledge-based battery design is worth further development.     

Mathematical parameters for quantification of mutational responses in bacteria

This paper introduces a new parameter, derivable from dose-response data for induced mutagenesis in bacteria, that can be used to quantify mutational responses in short-term tests. We called this parameter the mutational response of the bipartite experimental system (agent plus cells). We define it as being jointly proportional to the efficiency of the mutagen and the sensitivity of the test. We show how this quantity can be used to rank order chemical carcinogens on the basis of their mutagenicity and to determine the strength of any quantitative correlation that may exist between mutagenicity in bacteria and carcinogenicity in rodents. We find that this particular measure of mutational response for 10 direct-acting monofunctional alkylating agents correlates remarkably well with the rodent carcinogenicity of these chemicals measured in terms of their reciprocal TD₅₀ values.