Fig. 2

The scoring algorithm for determining most relevant miRs, genes and gene ontology terms. a Disease conditions were collated into a group of cardiovascular and metabolic as indicated. MiRs from literature were assigned an individual score as indicated in the box. Individual scores were dependent on whether miRs were found to be up-or down regulated, and dependent on the statistical significance of the finding as reported in literature. b Scores of miRs were calculated by balancing scores from each publication relevant to the condition under study. In addition, absolute values of miR scores for each condition were summarized over all disease conditions of a group, allowing us to assess its overall importance for either cardiovascular or metabolic diseases. c Disease specific scores from miRs were used to calculate a ranked list of target genes that are likely up-regulated or down-regulated in a certain condition. Gene-miR associations were obtained from miRTarBase, whereby only experimentally confirmed, strong relations were considered (hsa_MTI_strong). A ranking score for each gene was calculated summarizing the scores from all miRs of which the gene is a target. To correlate positive gene regulation with positive score, the sign of the score was inverted. The ranking list was then used for subsequent GO analyses using the tool Gorilla [9]