http://genomebiology.com/2001 /2/4/comment/1 005.1commentreviewsreportsdeposited researchinteractionsinformationrefereed researchCommentAn apology for orthologs - or brave new memesEugene V KooninAddress: National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD 20894, USA.E-mail: koonin@ncbi.nlm.nih.govPublished: 6 April 2001Genome Biology 2001 , 2(4):comment1 005.1 –1 005.2The electronic version of this article is the complete one and can befound online at http://genomebiology.com/2001 /2/4/comment/1 005© BioMed Central Ltd (Print ISSN 1465-6906; Online ISSN 1465-6914)Over the last few months, I have learned to look forward toGregory Petsko?s comments in Genome Biology. Beyondenjoying the witticisms, I tend mostly to agree with hismessage. When I started reading the recent article ?Homo-loguephobia? (Genome Biology 2001, 2:comment1002), itwas no different. Like Petsko, I hate that ?ue? at the end of?homologue?, and for that matter, all the other obnoxious?e?s?. I never consider having a drink in any establishmentthat has ?Olde? in its name, or buying as much as a hair combin a ?Shoppe?. And coming back to the ?homologous?problem, I think I have made a small contribution to leaner,meaner spelling by getting away with ?homolog? in manypublications, albeit accepting the forced ?homologue? inothers (even as I type this, the impervious little red wavefrom my spell-checker is, of course, right here, under?homolog?). So I was very much with Petsko on this momen-tous issue. As I read on, however, my happiness started towaver, and after reaching the invective against orthologs andparalogs, which Petsko says ?add nothing to the subject?, Ifelt that I had to lift my self-imposed ban on writing com-ments and respond to his article. Let me put it bluntly: I am confident that orthologs and par-alogs not only ?add something to the subject? but are criticalfor the development of evolutionary genomics (and as soonas two genomes were sequenced, all genomics became evolu-tionary). These are not fancy words (nor new, by the way:the notion of orthology versus paralogy was introduced byWalter .itch in a seminal 1970 paper; Syst Zool 1970, 19:99-113), but are essential designations for two distinct types ofevolutionary relationships. In a nutshell, orthologs are directevolutionary counterparts derived from a common ancestorthrough vertical descent; whenever we speak of ?the samegene in different species?, we actually mean orthologs. Incontrast, paralogs are genes within the same genome thathave evolved by duplication. Distinguishing between orthoand para is critical if we strive to describe evolution with anysemblance of accuracy. It is equally important for inferringgene function. Conservation of function is not part of thedefinition of orthology but rather its consequence. The distinction is not only logical but also very practical,because it is quite common for the same function in differentorganisms to be performed by proteins that are notorthologs nor even homologs (defining homologs as anygenes that have common ancestry, including both orthologsand paralogs). Hence another neologism of comparativegenomics that might induce a cringe even in some strongersouls who put up with orthologs and paralogs: non-ortholo-gous gene displacement, when unrelated - or at least notorthologous - genes perform analogous functions (seeKoonin EV, Mushegian AR, Bork P: Trends Genet 1996,12:334-336). I do maintain, however, that this one alsohelps us to speak more, rather than less, accurately andcomprehensibly about what is really going on duringgenome evolution. There is, however, yet another wrinkle that becomes appar-ent when one tries to think this through. Look at the trivial-ized schematic in .igure 1. Clearly, genes A1 and A2 areorthologs, and so are B1 and B2; and without hesitation wewill call A1 and B1 (or A2 and B2) paralogs, just as A and Bwere paralogs in the ancestral species. But what about A1and B2? These are not orthologs - they are not directly con-nected by vertical descent, not ?the same gene in differentspecies? - but neither are they paralogs, at least not accordingto the formal definition, because they reside in differentgenomes. Are we in need of yet another term? Perhaps meta-logs? This is not an idle concern. Imagine that B1 and A2have been lost during evolution and A1 and B2 are all thatremain of this gene family. We need to be able adequately todescribe the relationships between them, and at present thebest way to do so seems to be through the vague statementthat they are ?homologs but not orthologs?. Personally, Iwould prefer a new term.So what?s the issue with all these new terms (or ?exapted?ones, to use a favorite term of Stephen Jay Gould?s to indi-cate something pre-existing that has been recruited for anew function)? Or, for a good measure, with all the mush-rooming ?-omes? - transcriptome, proteome, metabolome,
