The BioZyne P-gp server uses a Support Vector Machine classifier to predict, with high accuracy, whether a compound is likely to be a substrate of the P-gp drug efflux pump.
Moreover, the server outputs the results of simple, interpretable decision tree models, as well as highlighting 'effluxophores', molecular fragments that enhance or abolish P-gp efflux. Read more about the methodology in our paper in Journal of Medicinal Chemistry.
The use of BioZyne P-gp server is free for a single compound at a time. A batch mode BioZyne P-gp server is available for screening large compound libraries on a commercial basis.
Try an example compound:
If you found this web service useful in your academic work, please cite:
Levatić J, Ćurak J, Kralj M, Šmuc T, Osmak M and Supek F. Accurate Models for P-gp Drug Recognition Induced from a Cancer Cell Line Cytotoxicity Screen. Journal of Medicinal Chemistry, 2013. DOI: 10.1021/jm400328s
To download the dataset of 934 P-gp substrates and nonsubstrates from our paper, click here.

Running P-gp substrate specificity model, please wait...