Figure 3

Ortho2ExpressMatrix case study A. RUNX1 data were extracted from two different public resources. Data in GEO series GSE17311 (related literature, see text) include log10-ratios of two channel differential expression data in four experiments using the human megakaryoblastic leukaemia cell line Meg-01; GSE2592 comprises expression data for Runx1 knock-out and wild-type mice at two time points and for all conditions done in duplicate. We extracted particular experiments from both data series for comparison with O2EM. Data for the human cell line were prepared for upload into O2EM by pre-calculation of the median from four biological replicates (combining data of biological or technical replicates are always in the user's control). Mouse data were separately averaged for the four knock-out and for the four wild-type experiments. The annotation function of O2EM associates probe information to gene and protein identifiers (transcripts and proteins stand, according to Ensembl, in a 1-to-1 correlation; transcript annotation is generally ignored in O2EM matrices) and finally to the families. This Figure, SYSTERS protein family 149072, in detail: annotation for RUNX genes of mouse (number 1) is displayed on the vertical axis, human orthologs are displayed on the horizontal axis. Gene names including links to Ensembl and gene description (near number 2) are located in the blue fields (additional links to KEGG genes and ATLAS experiments are truncated). The two items of the matrix borders comprise gene annotations in outer sphere (number 3) and, in the inner sphere (number 4), expression values of transcripts and information on related proteins; log2-ratios of differential expression intensities are averaged across all probe sets and translated to red/green colours (see section: Utility/Online visualization). Additionally, identifiers and intensities for each single probe set are given in a separate column (number 5; mouse-over). Background colours of the matrix cells (number 6) indicate similarity of the protein sequence pairs in the respective columns and rows. The reader should note the violet cell background that indicates the most similar proteins and respective genes and hence the respective closest orthologs, here distinguishing RUNX1-Runx1 genes from RUNX2-Runx2 and RUNX3-Runx3 genes and vice versa.