Molecular topology applied to the recognition of substrates of Breast Cancer Resistance Protein (BCRP)

MELISA GANTNER; MAURICIO E. DI IANNI; ALAN TALEVI; LUIS E. BRUNO-BLANCH

Alejandría

Conferencia; TWAS/BioVisionAlexandria Nxt.2012 International Biennial Conference; 2012

Center for Special Studies and Programs (CSSP) y TWAS Arab Regional Office (TWAS-ARO)

the BCRP is a transmembrane protein that reduces the bioavailability of drugs, so it is associated with multi-drug resistance in many diseases. The aim of this work is the development of computational models based on topological molecular descriptors for early recognition of BCRP substrates. To this purpose, we generated a data set of substrates and non-substrates of human wild type BCRP from literature. This data set was partitioned through a clustering procedure into a training set to train the model and a test set to assess its predictive ability. To obtain the model, Linear Discriminant Analysis was used to derive the corresponding linear discriminant functions that were validated through standard methodologies. The best model we obtained allowed a correct classification of around 92.6% BCRP-substrates and 56.9% of non-substrates, and has an area under the ROC curve of 0.7948 for the test set. This model is the first ?in silico? filter developed capable of early identification of potential substrates and non-substrates of BCRP, hence it can be applied to design drugs that are not recognized by BCRP transporter and is a useful tool to address the issue of multi-drug resistance in a cost-efficient manner.