Chordoma low-template microarray analysis

Abstract

Chordoma is a malignant bone cancer that develops from notochordal remnants. The MUG-Chor1 chordoma cell line shows a phenotypic anomaly, as it comprises two subpopulations, which were separated by micromanipulation and analysed with Affymetrix microarrays. Such a low-template microarray analysis, however, poses a difficult statistical problem, as the variances between biological replicates are high, whereas the biological differences are modest. Hence statistical approaches to increase the detection power are required. Firstly, we have applied a filtering step, which could slightly improve the detection power. Secondly, we have conducted a gene set test shifting the analysis from single gene level to gene set level, as such techniques are more sensitive in detecting moderate changes when these changes affect a set of related genes.