Suplementary material for manuscript submitted to Nature

From Genome to Phenome:
Here is Looking at You and Your Cancers

Antonio Reverter, Wes Barris, Sean McWilliam, Greg Harper and Brian Dalrymple

Bioinformatics Group, CSIRO Livestock Industries
306 Carmody Rd., St. Lucia, QLD 4067, Australia


ABSTRACT

The recent explosion in gene expression experiments has generated a large amount of data, which is predominantly analysed on an experiment by experiment basis although integrative analyses are being to emerge. What kinds of integrative analyses of gene expression datasets might we expect to see over the next few years? Here we examine two different analyses of one such large data set, the entire National Cancer Institute, Cancer Genome Anatomy Project, SAGE database, utilizing a range of currently available sophisticated, yet widely described mathematical and computational statistical tools, that are rarely, if at all, applied to gene expression data.. In the first analysis we ask the question; is there a relationship between the gene expression profile of a tissue and the expression pattern of it's neighbours, such that we can reconstruct an organism in three dimensions based on gene expression patterns? By applying multidimensional scaling techniques to gene co-expression similarity measures, we have generated a picture of a 'normal' and a 'cancer-affected' humanoid. We show that these pictures are consistent with the gross anatomy of the human being. In the second analysis we ask; can a single integrative analysis shed light on many of the key processes associated with the generic aspects of neoplasia, previously identified by many different experimental methodologies? Here, we identify and discuss the biological interpretation of the top 100 up- and down-regulated genes identified in our analysis