How to use BChemRF-CPPred

Please read the following steps to perform a prediction with BChemRF-CPPred.

  1. Go to the ‘Prediction‘ page by clicking on the appropriate link at the navigation bar.
  2. Scroll down the page until you see the ‘Upload your peptides to begin.‘ section.
  3. Select the desired input peptide type under ‘Input type‘. Two formats are supported, which are FASTA and PDB. Note that the models contained in this application are specialized to each input type.
  4. If you selected FASTA as input type, paste the peptide sequences in FASTA format in the box under Peptides. If you selected PDB as input type, click on ‘Select‘ and select the desired PDB files. As another option, you can also enable ‘Demonstration mode‘ to use examples of peptide structures to test the framework.
  5. Select the desired Feature Composition (FC) under ‘Feature composition‘. Four FCs are available according to respective descriptors. For more information on them, read the ‘About’ page.
  6. Select the desired model version under ‘Model Selection’. Version 1.0 corresponds to the models trained with 600 peptides, whereas Version 2.0 corresponds to the models trained with 750 peptides.
  7. When you are done, press on ‘Predict‘ to perform the prediction of peptide permeability using BChemRF-CPPred. You can also press on ‘Reset‘ if you want to clear the form and remove the previously selected structures.

After submission and processing, the application will expose the results. The ‘Predictions’ table contains the predicted class label of loaded peptides in CPP or non-CPP, followed by probability of the predicted class label. Below this table, other tables containing the values of the descriptors calculated for the uploaded structures will be shown, in accordance to the selected FC. These descriptors can be downloaded pressing ‘Download as .csv’.