Figure 1

Principle of the AGEP method, comparing microarray data from one test sample against a large reference database of different tissue/cell types. A) The expression profile of a test sample is first normalized to be compatible with reference data. Density estimates are then calculated for expression levels of each gene in each reference tissue type. B) Data for each gene in the test sample is aligned with the density estimates of all the normal reference tissue types to calculate an tissue match score (tm-score). This defines the likelihood that the expression of the gene originates from the reference values, with the score of 1 indicating that the gene in the input sample had the best match with the levels for that tissue type. A tm-score of 0 means that the input sample had an expression level that did not match the reference tissue type at all. C) Tissue specificity scores (ts-scores) for each gene of the test sample for each tissue in the reference database are then calculated from the tm-score matrix (see methods). Ts-scores range from -1 to 1 and indicate how uniquely the test sample resembles a certain tissue type according to the gene's expression level. D) Scatter plot visualization of tm- and ts-scores of all genes for a single query sample against one reference tissue type. Genes highlighted with the green box have matching and tissue specific expression level in the reference tissue and the query sample, genes highlighted with a blue box do not have tissue specific expression level in this tissue but the expression level of the query sample matched that, genes highlighted with a red box have tissue specific expression level in this tissue but the expression level of the query sample did not match that E) Based on the mean of ts-scores for all genes for each reference tissue type, similarity of the test sample against all the reference tissue types is displayed as a bar graph.