Pharmacoinformatic approaches to predict substrates of ABC-transporters

Supervisor: Gerhard F. Ecker, Medicinal Chemistry

Student: Floriane Montanari

Almost one-third of all compounds in the drug discovery pipeline fail due to improper ADMET behavior which renders in silico prediction of ADMET properties an important issue. Within the past decade, ATP binding cassette (ABC) transporters have been shown to influence the pharmacokinetics of many drugs. They both confer a multidrug-resistant (MDR) phenotype to cancer cells by expelling chemotherapeutic agents out of the tumor cells and play a crucial role in drug availability, metabolism and toxicity. Substrates of selected ABC-transporter are considered to show lower oral absorption, have limited capabilities to pass the blood-brain barrier, and are susceptible to drug-drug interactions, especially when combined with inhibitors.

Within the thesis we will develop and validate in silico models for prediction of substrate properties for the 5 major ABC-transporters involved in multidrug resistance and ADME: ABCB1, ABCB11, ABCC1, ABCC2, and ABCG2, focusing on natural products. Validated models will be further used to predict ABC-transporter profiles of compounds available in collaborating groups of the consortium.

Methods applied will comprise classical classification algorithms, support vector machines, decision trees, counter propagation neural network, random forest and rule fit. As validation is crucial for assessing the quality of the classification models obtained, we will also use the compound data sets present in the groups of our partners as external test sets and predict their profiles. Selected compounds will be biologically tested in the labs of our external collaborators.

The final goal of the project is to have a series of in silico models for prediction of ABC-transporter substrates. These models can be implemented in KNIME workflows and made available to the whole consortium. In particular, in silico profiles of compounds identified and analyzed in the projects of C. Hartinger, D. Marko, W. Jäger and V. Somoza will be established.