Time Series Analysis Of Cancer Metastasis Gene Expression Data Using Predictive Clustering Trees (PCT)

Abstract

Time series microarray data analysis provides an invaluable insight into the genetic progression of metastasis. In this paper we analyzed time series microarray data of cancer metastasis. Predictive clustering using decision trees on time series data integrates clustering and prediction. We present an implementation of PCT on cancer metastasis gene expression data to obtain high level descriptions of clusters of gene expression samples in terms of clinical variables in an unparalleled serendipitous manner. We evaluate time series data from microarray experiments. Each data set records the change over time in the expression level of epithelial cells to facilitate dispersion by induction of Epithelial Mesenchymal Transition (EMT). Implementation of PCT with CLUS-TS on the cancer metastasis gene expression data enables us to understand the early event during tumor metastasis. Summing up, we applied predictive clustering trees to the cancer metastasis gene expression data, where the goal was to obtain integrated high-level descriptions of clusters of gene expression samples in terms of clinical variables. Our preliminary results hint at the potential to understand the molecular pathogenesis of tumor metastasis through predictive clustering using decision trees approach.