Preferential co-expression of genes belonging to the same OXPHOS complex. (a) (i) Pearson correlation values (PCV) were calculated between the expression patterns of each OXPHOS subunit against all the others. In this example, the PCV for complex I subunits are coded white and the rest in black. For simplicity, the PCV of one complex I subunit (vertical) against all the others OXPHOS subunits are shown. (ii) Next, for each subunit (on the vertical axis), the genes were vertically ordered by their PCV values. A 30% cutoff was selected to calculate the percentage of Complex I subunits (in this example) over all other subunits within this cutoff. In this case, 35% of genes in the 30% cutoff are complex I subunits. (b) Schematic representation of the vertically ordered PCV for all OXPHOS subunits coded for each complex I subunit (each vertical line). The same approach was repeated for all other OXPHOS subunits. Notice, the top left quadrant (30% most similar genes to complex I subunits) there is a higher likelihood of finding complex I subunits (white) than subunits belonging to any other complex (black). (c) The same analysis but in this case for complex V subunits. See table 1 for a statistical analysis of these diagrams. These diagrams were constructed using data from the mouse GNF tissue compendium. CI, complex I genes; CII, complex II genes; CIII, complex III genes; CIV, complex IV genes, CV, complex V genes.