Development of computational drug repositioning approach

Computational drug repositioning aims to discover new therapeutic fields by analysing FDA-approved drugs. Due to the high cost of new drug development, the reuse of an approved drug for the treatment of another disease has become an important technique, especially for the cancer treatment which causes the most deaths in the world. This thesis study aimed to find candidate drugs that can change the expression profiles of cancer-causing proteins by developing a drug repositioning approach based on network theory. While creating the network structure, the protein-protein interaction network information obtained from the literature was used. The transcriptome profiles of the disease proteins on the created network structure and the drug candidates that are likely to suppress these molecules were taken from the LINCS L1000 project and mapped separately on this network structure. A drug repositioning approach was created by calculating the overlap scores of these two networks using statistical methods. The developed drug repositioning approach has been applied to lung and breast cancer. Five different metrics (shortest path, weighted jaccard, euclidean, manhattan, canberra) were used in the calculations. When the results were compared, it was seen that the shortest past algorithm included the compounds predicted by other metrics and the results of the shortest path algorithm were evaluated. When all compounds were ranked according to their normalized similar scores in the lung cancer disease network, 14 candidates were proposed for the treatment of lung cancer and 36 compounds for breast cancer. Comparing proposed candidate compounds with literature review, clinical phase studies and similar network-based approaches, common compounds and phase-II or phase-III clinical trials (mitoxantrone, dinaciclib and foretinib for lung cancer, alvocidib, AT-7519, mocetinostat, entinostat, mitoxantrone, and sunitinib for breast cancer) were found. Literature searches have shown that the compounds predicted by the developed approach may be candidate compounds for the treatment of both types of cancer. However, confirmation of these compounds by both in-vitro and clinical trials is recommended.