Supplementary Material

"Biological Assessment of Robust Noise Models for Micro-array Data Analysis"

by Posekany, Felsenstein and Sykacek

Gene Ontology terms

''GO_description_table.pdf'' provides a complete list of the significant Gene Ontology terms found by the robust and the Gaussian model along with their respective p-values.

Additionally we selected some of the differing genes of 5 data sets and plotted group-wise boxplots to gain a visual impression of the differential expression setting. This visualisation only made sense for those experiments with a more than 4 observations per group. The file ''boxplot_analysis.pdf'' contains the data and the respective plots.

Details of the Model and the Algorithm

This file contains additional details of

Using the Matlab program

The file ''matlab.tar'' provides the actual MCMC sampler, scripts for analysis of results and a normalised test data set.

Unpack it to a new directory, then open Matlab and call the script for analysing the Melanoma data by typing 'melanoma'. The scans in the data, performs inference of the noise model and selects the one with highest posterior probability. Then inference of gene expression is performed for the t model and the Gaussian model. To create parallel chains for comparison run the script with different filenames. Scripts for comparing genelists and Gene Ontology lists resulting from the inference are included and run as well.