Mitochondria, the organelles known to every junior high school student as “the powerhouses of the cell,” go back some two billion years. Although these energy producers were identified in the 1800s, how they became fixtures in cells is still under debate.

Mitochondria's ancestor was a free-living bacterium that another single-celled organism ingested. Most biologists think that the bacterium benefited the host: in one hypothesis, these premitochondria supplied hydrogen to make energy. Other researchers think that when atmospheric oxygen rose sharply in that era, anaerobic cells needed the bacteria to clear out the gas, which is toxic to them. However the match was made, the two lived so harmoniously that they eventually became mutually dependent and formed a long-term relationship.

A new analysis of evolutionary relationships by Martin Wu and Zhang Wang, both then at the University of Virginia, brings up the possibility that the mitochondrial progenitor was actually a parasite. Their claim derives from their recently constructed evolutionary tree for mitochondria, which resolves ancestral relationships among the organelles and their closest living bacterial relatives based on their genomes. Those DNA data led Wu to deduce that mitochondria sit within an order of parasitic and pathogenic bacteria called Rickettsiales and that they evolved from an ancestor that produced an energy-stealing protein. At some point, this parasitic predecessor lost the klepto gene and gained another that enabled it to supply energy to its host, as mitochondria do today. The researchers published their findings in October 2014 in the journal PLOS ONE.

But other scientists take issue with the paper's conclusions. Dennis Searcy, who studies the origin of mitochondria at the University of Massachusetts Amherst, says the authors interpreted their evolutionary tree wrongly when they decided that mitochondria descended from Rickettsiales. Such a miscalculation would clearly corrupt their analysis. And Michael Gray, who researches mitochondrial evolution at Dalhousie University in Nova Scotia, thinks that the rapid evolution of the organelles makes it difficult to say with certainty where the once free-living entities sit within their branch of the tree.

Wu maintains that the study minimized errors as much as possible, while acknowledging that better models are necessary to assign definitive relationships. “There is definitely more work to be done,” he says. “There are still very large gaps in the tree.”