Subject: Fwd: gene nomenclature
From: Paul Sternberg
Date: Fri, 24 Oct 2003 16:33:27 -0700
CC: Paul Sternberg

Begin forwarded message:

From: Eric Haag <>
Date: Fri Oct 24, 2003 3:06:56 PM US/Pacific
To: "David H. A. Fitch" <>, Ronald E Ellis <>,,,
Subject: Re: gene nomenclature

Wow, so many good points made.  I must retreat!  In fact, there are many cases along the lines of what David describes:  briggsae has two nearly identical but widely separated lin-12 duplicates (Dave Rudel's work, verified by the genome project).  One is in the syntenic place, but the other one is off by itself.  And as we heard in LA, there is no clear fog-2 orthologue, but many FTR genes in both species.  How do we handle these?   And this is clearly just the tip of the iceberg--God forbid someone tack nuclear hormone receptors or serpentine proteins.

But what should I do in this hypothetical (and hoped-for) case:  I get a mutant that cleanly masculinizes  XX briggsae animals, and it doesn't map to a known Tra or Sdc gene.  What do I call it?  I think we all agree that it should be named Cb- or Cbr-xyz-n.  Following Ron and David's suggestion, I might make up a new name, "mas-1," for "masculinized"  Then I never have to change the name, and that is great.   But I live with the fact that it sounds like it ought to have a different function than the  Cb-tra homologues with the same phenotype.  We don't do this sort of mixing of names within a class in elegans (at least not on purpose), and it really helps.

This brings me to two wildly  divergent options:  either abandon acknowledgement of homology with elegans, or make it the central organizing principle.  With the history and utility of the elegans names, I hate to see them tossed.  But perhaps it is better to have no expectation of naming similarity between species than have it be unreliable (as Ron pointed out).

Given that, I'm willing to consider the idea of parallel naming--one specifically relative to elegans, another independent.  We do need to chew on the practicality of this, though.  Do we really want to have to look at a correspondence table to see what a gene does? If we collectively feel it's best, I'm OK with it.  If we do have a species-specific nomenclature, then for sequenced genomes might we consider getting out of phenotypic descriptors altogether and adopting  a purely physical map-based system, using the gene models in wormbase as the guide?  This creates totally nonintuitive names, but it's internally consistent, is effectively already done for us in briggsae, and doesn't lead to false similarities or dissimilarities in relation to elegans.  We'll also never run out of classes, because there won't be any.

Is the baby still in the bathtub?



                                          Eric S. Haag, Ph.D.        ~
  Assistant Professor                                   ~
  Department of Biology                               ~ ~
  University of Maryland
  College Park, MD  20742               ~                       ~   ~
  phone:  (301) 405-8534      fax:  (301) 314-9358    ~         ~
                                    "I'd rather be here now."

Paul W. Sternberg
HHMI and Division of Biology 156-29, Caltech
1200 E. Calif Blvd., Pasadena CA 91125
tel 626-395-2181 fax 626-568-8012 email
Admin. Assistant Mary Alvarez 626-395-3990
Lab webpage:
Worm Literature Search Engine: