An Efficient Branch and Cut Algorithm to Find Frequently Mutated Subnetworks in Cancer
Cancer is a disease driven mostly by somatic mutations appearing in an individual’s genome. One of the main challenges in large cancer studies is to identify the handful of driver mutations responsible for cancer among the hundreds or thousands mutations present in a tumour genome. Recent approaches have shown that analyzing mutations in the context of interaction networks increases the power to identify driver mutations.
In this work we propose an ILP formulation for the exact solution of the combinatorial problem of finding subnetworks mutated in a large fraction of cancer patients, a problem previously proposed to identify important mutations in cancer. We show that a branch and cut algorithm provides exact solutions and is faster than previously proposed greedy and approximation algorithms. We test our algorithm on real cancer data and show that our approach is viable and allows for the identification of subnetworks containing known cancer genes.
KeywordsCancer mutations Branch and cut Combinatorial optimization Network analysis
Computation for the work described in this paper was supported by the DeiC National HPC Center, University of Southern Denmark. This work is supported, in part, by MIUR of Italy under project AMANDA and by NSF grant IIS-1247581, and has been done, in part, while FV was a research fellow at the Simons Institute for the Theory of Computing (University of California, Berkeley). The results presented in this manuscript are in whole or part based upon data generated by the TCGA Research Network: http://cancergenome.nih.gov/.
- 6.El-Kebir, M., Klau, G.W.: Solving the maximum-weight connected subgraph problem to optimality. CoRR abs/1409.5308 (2014)Google Scholar
- 8.Fischetti, M., Leitner, M., Ljubic, I., et al.: Thinning out steiner trees: a node based model for uniform edge costs. Math. Progr. Comput. (2015, submitted)Google Scholar
- 13.Johnson, D.S., Koch, T., Werneck, R.F., et al.: The eleventh dimacs implementation challenge. http://dimacs11.cs.princeton.edu/home.html
- 15.Kim, Y.A., Salari, R., Wuchty, S., et al.: Module cover-a new approach to genotype-phenotype studies. Pac. Symp. Biocomput. 135–146 (2013)Google Scholar
- 29.Wang, Y., Buchanan, A., Butenko, S.: On imposing connectivity constraints in integer programs (2015). http://www.optimization-online.org/DB_HTML/2015/02/4768.html