ToxiM is a machine-learning based binary classifier for the prediction of toxicity of molecules. Regression models for CaCo-2 permeability and Aqueous solubility has been provided to support the predicted toxicity for a given molecule. For the batch processing please download standalone version of ToxiM from Dataset page.
Brief decription of the tools available at the webserver is provided below. Toxicity Prediction: This module helps the user to predict the probability values for the potential toxicity of a molecule. Users can submit one or multiple query compound(s) in sdf, smiles, mol or pdb format (except large peptides) or can provide a list of CIDs of compounds. CaCo-2 Permeability Prediciton (LogPapp): This module helps to predict Caco-2 permeability value of query compound(s) in logarithmic form. Using this module, the user can determine the extent of the Caco-2 permeability of the molecule by refering to the colour coding scale provided at the Results page. Aqueous Solubility Prediction (LogS): This module helps to predict the values of aqueous solubility for molecules. If the experimental value for any molecule is not known, a prediction is made. However, if the experimental values are known, the user can validate the predicted values with the experimentally known values of aqueous solubilities in log form of compounds. Similarity Search:This module helps to identify the most similar molecule via similarity search of query molecule against all the molecules present in the toxin database using MACCS, FP2, FP4 fingerprints and the hybrid of all the three.

Citation

Sharma AK, Srivastava GN, Roy A, Sharma VK. Frontiers in Pharmacology. 2017 doi:10.3389/fphar.2017.00880