Fig. 2

Evaluation of conservation of 30 tissue-specific gene sets. The statistical procedure to obtain the results for panels a and b is identical and based on a Mann-Whitney test between the values of a gene set and the values of the remaining genes. For a we used the number of CCG and for b we used the values of differential connectivity. The analysis has been performed both on the entire set of homologs (bars in black) and on one-to-one orthologs only (bars in grey) with asterisks indicating the significant results (FDR <0.05). For the panels c and d we used the Fisher’s exact test. In c we reported the odd ratio of homologous genes that underwent duplication (one-to-many and many-to-many homologs), and in d we reported the odd ratio of non-homologous genes (for further details refer to Additional file 2: Figure S5 and Methods). The tissues are ranked according to the level of conservation in terms of common co-expression and 20 out of 30 resulted to have conserved patterns (a). The differential connectivity is biased towards the human species (b) that can be partially explained by the fact that the mouse species has a greater annotated number of non-homologous genes than humans that in addition have a higher connectivity compared to the human counterpart (Additional file 2: Figure S4). In panel c the brain and bone tissues have the lowest proportion of duplicated genes, opposite to blood, liver, bone marrow, mammary gland, placenta and skin. Lastly, panel d shows that there is a global preservation of the usage of homologs for tissue-specificity gene expression