Genetic and clinical information converge
However, due to the special features of microarray data, traditional statistical analysis methods can not be applied directly to analyze such data. For example, measurement error, missing observations, and high-dimensionality – in which the number of observations (less than a hundred patients or samples) is small relative to the number of variables (tens of thousands of genes) – present considerable challenges in developing strategies for the analysis of microarray data.
“My recent project involves building models to predict survival probability of cancer patient using gene expression data,” says Grace. “Traditionally, clinical data are used to study patient survival. Converging genetic and clinical information may help to better comprehend cancer patient disease progression and build more powerful prediction models.”
University of Waterloo Mathematics, Annual Report 2006