Drug ranking and repositioning through network projections and kernelized score functions
Drug repositioning is the problem of finding novel therapeutic indications for existing drugs.
By exploiting similarities between drugs in the "pharmacological space", constructed through the integration of drug chemical similarity networks and network projections based on bipartite graphs (Re et al, 2012), we rank drugs with respect to DrugBank therapeutic categories, using methods based on kernelized score functions, thus allowing to find candidates for drug repositioning (Re and Valentini, 2012).
Kernelized score functions apply a double semi-supervised learning strategy (Re and Valentini, 2013):
1. Local learning based on properly designed score functions that take into account the neighborhood of each node/drug of interest.
2. Global learning strategies that take into account the overall topology of the pharmacological integrated network through the selection or the proper design of graph kernels embedded in the score functions.
M. Re, and G. Valentini, Network-based Drug Ranking and Repositioning with respect to DrugBank Therapeutic Categories, IEEE ACM Transactions on Computational Biology and Bioinformatics 10(6), pp. 1359-1371, Nov-Dec 2013 IEEE link Supplemental Material
M. Re, G. Valentini Large Scale Ranking and Repositioning of Drugs with Respect to DrugBank Therapeutic Categories, slides In: L. Bleris et al. (Eds.): International Symposium on Bioinformatics Research and Applications (ISBRA 2012), Dallas, USA, Lecture Notes in Bioinformatics vol.7292, pp. 225-236, Springer, 2012.
M. Re, M. Mesiti, G. Valentini, Drug repositioning through pharmacological spaces integration based on networks projection, EMBnet.journal, vol 18, Supplement A, pp.30-31, 2012.