The Mycobacterium tuberculosis permeability checker for small molecules uses the physico-chemical descriptors calculated by Schrodinger® Maestro QikProp or the open-source package PaDEL-Descriptor to calculate the coordinate of principal component 1 of a PCA of the training dataset (Ekins et al. 2010). This coordinate is then used as input for a logistic regression model. As a result the user retrieves a probability of permeability for each molecule of interest. To evaluate results each used descriptor is compared to the corresponding distribution within the training data and displayed according to a blue-scale coloring.
The input format (QikProp or PaDEL CSV) is recognized automatically. Please consider that the regression model used by MycPermCheck depends on the input format.
If your file contains several molecules with the same name (e.g. tautomers, protomers, stereo isomers, etc.), you can check the box "Calculate Mean of Isomeric Forms". In this case the values of the descriptors are averaged for all same-named molecules. If your file contains several same-named molecules and the box is not checked, the first molecule of this group is picked unchanged as a representative for all.
Your molecule names should not contain special characters.
The website was optimized for Mozilla Firefox.
If you use MycPermCheck, please cite: Merget et al. (2013) MycPermCheck: The Mycobacterium tuberculosis permeability prediction tool for small molecules, Bioinformatics, 29(1): 62-68.
If you use the PaDEL-Descriptor-based model, please do not forget to also cite: Yap C.W. (2011). PaDEL-Descriptor: An open source software to calculate molecular descriptors and fingerprints, Journal of Computational Chemistry, 32(7): 1466-1474.
How to calculate molecular descriptor data:
To calculate descriptor data with QikProp or PaDEL you need a common 3D-structure coordinate file, like mol2 or sdf. Common chemical sketching programs (like ChemDraw) have an export function to the so-called SMILES format, which looks like this. This format is a linear string, coding your 2D connectivity of your chemical structure including stereochemical information and a title. You can use open-source software like OpenBabel to convert this SMILES-string to a three-dimensional mol2 file or several web services, e.g. Cactus. The resulting 3D mol2 files should be checked for errors with a molecular viewer, like PyMOL, VMD or Avogadro. Eventually, the mol2 file(s) can be used as input for PaDEL-Descriptor or QikProp. But of course, there are countless ways of obtaining 3D structures from chemical sketches.