An Iterative Approach for Phylogenetic Analysis of Tumor Progression Using FISH Copy Number

Abstract

Copy number variants are an underlying factor in human evolution and in many diseases, especially in cancer. Tumors generally contain cells with a varying number of gene copies, and the variance in the number of gene copies follows a pattern formed by an evolutionary process. The Fluorescence in situ hybridization (FISH) provides researchers a reliable technique to measure the copy numbers of preselected genes in a group of cells. Recently, Chowdhury et al. successfully modeled the progression of tumor progression using FISH copy number to the Rectilinear Steiner Minimum Tree (RSMT) problem, and proposed both exact and heuristic algorithms to reconstruct phylogenetic trees modeling the development of cancer cell patterns [1]. We proposed new heuristics to attack the RSMT problem, which is inspired by iterative approaches to approximate solutions to the Steiner tree in the “small phylogeny” problem [2,3]. Experimental results from both simulated and real tumor data show that our approach outperforms the previous heuristic algorithm in approximating better solutions for the RSMT problem.