ProTox: a web server for the in silico prediction of rodent oral toxicity.

Drwal MN, Banerjee P, Dunkel M, Wettig MR, Preissner R. · 2014

Read ProTox: a web server for the in silico prediction of rodent oral toxicity. by Drwal MN, Banerjee P, Dunkel M, Wettig MR, Preissner R. (2014) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4086068.txt.

Introduction

Malgorzata N. Drwal, Priyanka Banerjee, Mathias Dunkel, Martin R. Wettig, and Robert Preissner built ProTox, a free web server that predicts how toxic a compound would be to rodents if swallowed. Their paper appeared in Nucleic Acids Research in 2014. Computer prediction aims to cut the time, cost, and animal testing in early drug and cosmetic safety work.

Given only a compound's two-dimensional structure, ProTox compares it with chemicals of known lethal dose and scans for toxic fragments. It also flags possible protein targets of harm, using models built from known drug-reaction data. On an outside validation set, the server balanced catching toxic compounds with few false alarms, performing strongly against formula-based rivals. No registration is needed to use it.

Catalog record

CREATORDrwal MN, Banerjee P, Dunkel M, Wettig MR, Preissner R.
YEAR2014
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC4086068

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

ProTox: a web server for the in silico prediction of rodent oral toxicity By Malgorzata N Drwal, Priyanka Banerjee, Mathias Dunkel, Martin R Wettig, Robert Preissner Published in: Nucleic Acids Research DOI: 10.1093/nar/gku401 ABSTRACT Abstract Animal trials are currently the major method for determining the possible toxic effects of drug candidates and cosmetics. In silico prediction methods represent an alternative approach and aim to rationalize the preclinical drug development, thus enabling the reduction of the associated time, costs and animal experiments. Here, we present ProTox, a web server for the prediction of rodent oral toxicity. The prediction method is based on the analysis of the similarity of compounds with known median lethal doses (LD50) and incorporates the identification of toxic fragments, therefore representing a novel approach in toxicity prediction. In addition, the web server includes an indication of possible toxicity targets which is based on an in-house collection of protein–ligand-based pharmacophore models (‘toxicophores’) for targets associated with adverse drug reactions. The ProTox web server is open to all users and can be accessed without registration at: http://tox.charite.de/tox. The only requirement for the prediction is the two-dimensional structure of the input compounds. All ProTox methods have been evaluated based on a diverse external validation set and displayed strong performance (sensitivity, specificity and precision of 76, 95 and 75%, respectively) and superiority over other toxicity prediction tools, indicating their possible applicability for other compound classes. INTRODUCTION About 100 million rodents are used annually in animal trials, many of them for toxicity testing in the pharmaceutical or cosmetic industry. European regulations require in vivo testing of novel industrial chemicals, biocides, pesticides and cosmetic ingredients for acute oral toxicity (1). Similar regulations are in place in other countries, too. However, following the Replace, Reduce, Refine (3Rs) principle (2), the development and optimization of alternative methods including in vitro assays and in silico predictions have been put into focus in the recent years. According to estimations, drug candidates entering clinical trials have only an 8% chance of becoming marketed drugs and about 20% of failures in the late drug development are caused by occurrences of toxicities (3,4). Computational predictions offer many advantages in the drug development process. They can be applied early in the drug discovery process and therefore save the amount of time as well as the in vitro and in vivo experiments required. Predictions of pharmacokinetic properties including the absorption, distribution, metabolism, excretion and toxicity (ADMET) also provide a method to prioritize the selection of compounds to be synthesized and evaluated as drug candidates. Toxicity predictions take place at different levels of complexity, ranging from predictions of toxic endpoints such as acute toxicity or carcinogenicity to predictions of the underlying mechanisms of toxicity development, e.g. the identification of targets involved in adverse drug reactions and toxic effects. Predictions of toxicity endpoints are mainly divided into statistical models and so-called expert systems (5). Whereas expert systems incorporate human knowledge and expert toxicity rules such as structural alerts (4), statistical models include quantitative structure–activity relationships (QSAR) and machine learning algorithms trained on compounds with known toxic activity. On the other hand, target-based toxicity predictions mostly apply methods also used in computer-aided drug design to identify novel hit compounds. In particular, pharmacophore models have emerged as a fast technique which can be easily applied for profiling purposes (6–8). In silico toxicity predictions are integrated into several commercial software environments including Discovery Studio's TOPKAT® (Toxicity Prediction by Komputer Assisted Technology; Accelrys, Inc., USA), ADMET Predictor™ (Simulations Plus, Inc., USA) or ADME-Tox Prediction (Advanced Chemistry Development, Inc., Canada). Freely accessible tools are also available for a variety of endpoints and are mainly based on QSAR models and expert systems. Examples of QSAR-based methods which present prediction of acute oral toxicity include the Toxicity Estimation Software Tools (T.E.S.T.) developed by the U.S. Environmental Protection Agency or the web server AdmetSAR which includes a variety of ADMET prediction models (9). Although QSAR models can show good hit rates in toxicity prediction, it is not clear how they will perform for structurally novel compounds not included in the training set. While the integration of newly available data requires the redevelopment of new QSAR models, other methods like similarity searches in databases are fast and can easily be extended to include new information. Similarity methods are based on the assumption that structurally similar molecules should manifest similar biological activities and many successful applications of similarity-based methods have been reported, including the identification of novel hits for drug discovery, drug repositioning studies or the prediction of drug–target interactions, side-effects and therapeutic indications (10–14). Thus, the use of similarity-based methods might also represent a promising approach for the prediction of toxic effects. Here, we present ProTox, a web server for the prediction of median oral lethal doses (LD50 values) and toxicity classes in rodents. The prediction method is based on the analysis of the two-dimensional (2D) similarity to compounds with known LD50 values and the identification of fragments over-represented in toxic compounds. In addition to the oral toxicity prediction, the web server indicates possible toxicity targets based on a collection of protein–ligand-based pharmacophores (‘toxicophores’) and therefore provides suggestions for the mechanism of toxicity development. The prediction methods are fast and the underlying database structure allows simple and fast upgrades of the prediction server when new data becomes available. A validation on an external dataset suggests good performance of ProTox in comparison to several other in silico toxicity prediction tools. PROTOX INTERFACE Input The ProTox web server has an easy-to-use interface and the only requirement is the 2D structure of the molecule for which the toxicity is to be predicted. The user has the possibility to draw the structure with an embedded chemical editor or use the integrated PubChem search (https://pubchem.ncbi.nlm.nih.gov/) which allows the search for chemical structures using the compound name. Additionally, ProTox allows the upload of files containing one or more compounds in the mol format. This is particularly interesting for early drug discovery projects, where drug candidates could be evaluated using ProTox with respect to their potential toxicities. A possible input for ProTox would be, for example, a hit list obtained from high-throughput or virtual screening and the predicted toxicities could be used for prioritization of compounds for further developments. Output Typically, a toxicity prediction report for one compound is generated within seconds. The report can be divided into two parts: the prediction of the acute oral toxicity and the indication of possible toxicity targets (Figure 1). The oral toxicity prediction results are based on the analysis of 2D similarities and the recognition of toxic fragments, as described below. In addition to the prediction of the LD50 in mg/kg, the input compound is classified into a toxicity class ranging from I to VI, according to the globally harmonized system of classification of labelling of chemicals (GHS, United Nations, first revised edition 2005). The prediction accuracy derived from cross-validation results is also given. Additionally, the structure of the input compound as well as its physicochemical properties are displayed. In the similarity section, the three most similar compounds of the database which are used for the prediction of the oral toxicity, their properties and classification are shown. Furthermore, the fragment section indicates if and

Read the complete text