Topographic independent component analysis of gene expression time series data

Abstract

Topographic independent component analysis (TICA) is an interesting extension of the conventional ICA, which aims at finding a linear decomposition into approximately independent components with the dependence between two components is approximated by their proximity in the topographic representation. In this paper we apply the topographic ICA to gene expression time series data and compare it with the conventional ICA as well as the independent subspace analysis (ISA). Empirical study with yeast cell cycle-related data and yeast sporulation data, shows that TICA is more suitable for gene clustering.