Behavioral Ecology Blog

About bacterial species concepts– or the lack thereof…

The prevailing definition of a bacterial species is that of a coherent group of organisms more similar to themselves than they are to other things.. this is primarily thought of as in terms of phenotype or gene sequence. The degree of similarity is currently measured as >70% DNA-DNA hybridization, >97% sequence similarity when looking at 16 rRNA, or more modernly- >94% ANI. But we do know some phenotypically distinct organisms are pretty similar when looking at housekeeping or 16S- so the usefulness of basic similarity methods are also in question..

Coherence is eroded by lateral gene transfer and homologous recombination**, both of which can happen between distantly related groups. These processes are so common that it is likely that there are very few groups that maintain themselves coherent enough to be called species. In addition, limits on LGT and recombination are few and far in between..i.e both mechanisms are quite promiscuous…

OK, and even if you do successfully form a cluster of related individuals based on 16S or whatever other marker- there is the issue of gene content versus gene sequence. For instance- in a study of 33 genome clusters that are >94% ANI- there are between 1-4,044 genes per cluster that are present in one genome but not another. Clearly- within “species” variability is a big problem. How can you call 2 things the same species when they have 4,000 additional genes??

So- coherence is unlikely to happen in bacteria- and even when it does, we’re still not all that sure how overall similarity will fall out. This clearly does not bode well for ANY bacterial species concept.

Perhaps this is an issue of animal-envy.. Microbiologists want that tight species concept like animals have (if they only knew how confused we are ourselves), and perhaps it is just not so. As Doolittle 2006 suggests- maybe species boundries in Proks are simply fuzzy at best, with exceptions going both ways- some that are clearly coherent groups- some that clearly aren’t. The question for me at this point is this:

How much weight does each group have: Will we be able to have fuzzy species boundaries for the majority of bacteria? Will the majority be “without borders”, will there be an even distribution???

** homologous here is meant pretty loosely- as it can occur even in highly divergent (>20%) sequences…