Different combinations of many processes may be responsible for the different levels of somatic instability. To investigate the levels of contributions from significant pathways in each tissue, we identified highly correlative probes (absolute Pearson coefficients > 0.6) in the two most significantly positively correlated gene sets, and plotted relative expression levels against measured instability indices in the 16 training set tissues. (A) 'UDP-galactose beta-N-acetylglucosamine beta-1,3-galactosyltransferase activity' gene set was the most significant pathway in the gene set analysis (positive correlation). This gene set is composed of 21 probes, and 7 probes were highly correlative (correlation coefficient > 0.6). The expression levels of these 7 probes in liver were low compared to those in striatum although instability indices are similar in these tissues. (B) 'Adrenoceptor activity' gene set was the second most significant pathway (positive correlation), and has 25 probes as members, of which 7 were highly correlative (correlation coefficient > 0.6). Interestingly, cortex or hippocampus showed similar expression levels of the highly correlative probes in this gene set to those in striatum, although instability indices in cortex or hippocampus were significantly lower than that of striatum. Graphs show 7 highly correlated probes for each gene set, and IDs are Affymetrix MG430 2.0 probe set IDs.
Lee et al. BMC Systems Biology 2010 4:29 doi:10.1186/1752-0509-4-29