TubPred 1.1


1. What is TubPred?

TubPred is a webserver which is based on proteochemometric modeling (PCM) for the prediction of the bioactviity between a small compound and a given tubulin rececptor.

2. What is PCM?

PCM is a quantitative bioactivity modelling technique that is used to predict the bioactivity of compound-protein (ligand-target) pairs by using the compounds and related protein features in the construction of a single machine learning model.

3. How does TubPred works ?

The enhanced user query interface enables users to perform the following actions:
  • Tubulin receptor uniprot id entry
    • - Tubulin proteins: e.g. P02554
  • Small compound entry in smiles format
    • - Compound: e.g. for Epothilone B (CC1CCCC2(C(O2)CC(OC(=O)CC(C(C(=O)C(C1O)C)(C)C)O)C(=CC3=CSC(=N3)C)C)C)
  • Simply click to compute prediction
  • Results of predictions will be outputted

4. What is the current version of TubPred?

The current version 1.1 is an update to the original TubPred server.

5. Will there be any future updates or modification to TubPred?

Yes, we would like to enhance the capabilities of Tubpred predictions by retraining model with more dataset. Additionally, we will integrate an insilico target prediction to Tubpred for computing new plausible tubulin targets for small compounds.

6. Are there any limitations with TubPred?

Yes, the model was trained on a limited dataset of tubulin-small compound interaction space.

7. What is the meaning of a query with an output, for example, active with 0.82 confidence score ?

Confidence score ranges from 1.0 to 0.0. Close to 1.0 is high confidence whilst close to 0.0 is low. So active with 0.82 confidence score means it is predicted as active but not completely certain as having anti-tubulin activity. Only experimental validation can confirm bioactivity.

8. Can I download prediction results from TubPred?

Yes, users may download the results in a tab delimited format at the results section

9. Where should I send my complaints?

Please send a complaints via the Contact menu item.

10. How to cite TubPred 1.1?

Agyapong, O., Miller, W.A., Wilson, M.D. & Kwofie, S.K. (2021). Development of a proteochemometric-based support vector machine model for predicting bioactive molecules of tubulin receptors. Mol Divers (2021). https://doi.org/10.1007/s11030-021-10329-w