Conserved genetic interactions of transcribed unitary pseudogenes. (a) We estimated the correlation in expression between orthologous mouse (green) or human (blue) loci (Mi and Hi, respectively, ovals) and other protein-coding genes (rectangles) annotated in mouse or human genomes. We retained only genes exhibiting significantly (empirical P-value < 0.05) positive (+) and negative (-) correlation with Mi or Hi tissue expression (dashed line). We identified all mouse genes (for example, 'a' boxed) whose expression profiles are significantly correlated with that of Mi, and whose human one-to-one ortholog (for example, 'A') is also significantly correlated in the same direction with Hi. (b) f is the fraction of mouse genes, m, that are significantly correlated in expression with a unitary pseudogene, Mi, whose human orthologs are also significantly correlated in expression (and in the same direction, either positively or negatively) with Hi, the human ortholog of Mi. f for human protein-coding gene and mouse unitary pseudogene pairs (green) is significantly higher (***P < 0.001) than for random non-orthologous mouse and human gene pairs (grey) and significantly lower (*P < 0.05) for mouse and human one-to-one orthologs. No constitutively expressed exons (required to measure gene expression) were identified for EXD3, THAP9, RNF175, DBF4B and ZBED. Hence these genes were not considered in this analysis. (c) Unitary pseudogenes (green) share significantly (***P < 0.001) more MREs with their human protein-coding ortholog 3' UTR than random pairs of mouse and human protein-coding genes (grey). MRE predictions were not available for ZBED5 and this locus was not considered in this analysis. (d) Twenty-two percent of conserved significantly positively (++) correlated quartets (Mi-A-a-Hi, dark grey) share at least one MRE for the same miRNA family across the four loci. This is a significantly higher fraction than found for quartets that are significantly negatively correlated (--) or significantly correlated (+- and -+) in different directions (light grey).
Marques et al. Genome Biology 2012 13:R102 doi:10.1186/gb-2012-13-11-r102