On the debate of recognising mitochondria as bacteria

I recently read Mark Pallen's provocative opinion article on whether we should recognise mitochondria as bacteria, and therefore give them their own taxonomic classification (now they are simply considered an organelle of the host cell). This paper has been featured and discussed in other blogs such as that of Jonathan Eisen.

 This issue is close to my heart, since I did my PhD precisely in tracing the origin and evolution of this organelles and their inter-mingling with their hosts. I agree with many of the arguments raised by Pallen,
 but I do not necessarily conclude that this should lead us to create a distinct taxonomic class for mitochondria. In practice, thinking of "mitochondria as bacteria" can coexist with its current classification as an organelle, as long as we are aware of their bacterial ancestry. Indeed I think that this is the dominant view of most people doing research on mitochondria, so I do not think that classifying mitochondria as a new bacterial class will radically enhance our possibilities of understanding or manipulating this organelles. On the other hand one could argue that strictly considering mitochondria as bacteria will close our eyes to the critical organellar properties of mitochondria, thereby hampering our potential to understand them. I imagine a future opinion paper entitled "time to recognise that Mitochondriaceae are organelles?", with the same kind of arguments arguing for recognising mitochondria as a true organelle from the host cells, and listing its many similarities with other membrane-bounded organelles in the cell.

 Since evolution has crossed the border from free living bacterium to organelle at least a couple of times, it follows that these two stages are united by a continued evolutionary time line, of small stages separated by a discrete number of changes. The current diversity only allows us to infer some of these intermediate stages and usually these are used to base our "categories". Making a separation is of course arbitrary but useful to describe to things that appear different to us. Despite the parallelisms mentioned with some reduced endosymbiont or pathogens, there is a quantitative jump in the level of integration with the host cell that we can easily recognize. Setting there a divide between what we call an organelle or what we call a bacterial species seems to me reasonable, although we could think of other operational definitions, is clear. 

 All the rest is deciding what degree of purity we want to apply to our definitions. As a biologist I am used to the limitations of our central concepts such as the ones of "species" or "genes", which similarly have to accomodate exceptions and different interpretations depending on what organisms we are dealing with. We humans have a natural tendency of classifying things into simple schemes, and we have to recognize the advantage of using operational classifications that are "generally correct" while not becoming too uneasy when understanding that the actual complexity is much bigger. The important thing is to be aware of the exceptions and of the "provisional and approximate nature" of practical definitions, until we find better ones.

 In summary, I am in favour of changes of our current paradigms to newer ones that better fit our current knowledge, but I am of the opinion that simply giving mitochondria the level of a taxonomic family is not solving anything, nor improving our understanding of these "highly derived bacteria" or "bacterial-derived organelles", as you prefer to call them.