This group did, in my opinion, a reasonably good job of modeling mitochondrial membrane proteins in a way that utilizes all the information (likely) available.  I would like to be able to do such for all known mitochondrial membrane proteins, or even the whole mitochondrion… matrix proteins as well, but lack the skill and knowledge (working on the latter) to do so at this time.  I would have made only one adjustment, and that would be to have depicted the inner and outer mitochondrial membranes as actual molecular structures as well. It is still possible to detect the areas the transmembrane areas in some of the molecules.  It is a little confusing to have the outer mitochodrial membrane as a stiff line, as we know it curves with, sometimes closely to (as at the pores) and other times further apart, and the difference in the curvature of the inner mitochondrial membrane is sort of cumbersome and doesn’t flow through the transmembrane domains comfortably.  But a great depiction, none-the-less.

Here is their citation: Medlock, Amy & T Shiferaw, Mesafint & Marcero, Jason & Vashisht, Ajay & Wohlschlegel, James & Phillips, John & Dailey, Harry. (2015). Identification of the Mitochondrial Heme Metabolism Complex. PloS one. 10. e0135896. 10.1371/journal.pone.0135896.