One aspect of gene nomenclature (often really symbology) that seems not to
have been discussed is the likely future trajectory. If we assume that
the various organismal genome programs will be at all successful we must
begin thinking of the time when there are tens of thousands of genes known
in each of many different organisms. Under those conditions, it will long
have become impossible to embed much mnemonic meaning into any short and
convenient gene symbol.
Another problem from that future perspective is comparative nomenclature.
In vertebrate anatomy there has been a history of attempts to ensure that
equivalent names be given to homologous structures in related organisms.
At the level of basic house-keeping genes, it is not unreasonable to
assume that there will be quite a few homologous genes in all organisms
and others that span very wide taxonomic groups. Determining the most
meaningful nomenclature to capture these homologies will be difficult
until many more loci have been identified and characterized, and the
process will not begin in earnest until the data and tools for comparative
genomics are available. These are coming, but until they arrive there
will be an increasing need for provisional names.
Provisional names are already a problem when orfs are identified and must
be named before anything is known about them, except perhaps some weak
homology to a sequence in a different organism. A few suggestions have
been published on provision genetic nomenclature, but all that I have seen
fall short of meeting even their own avowed goals.
I suggest that care be taken in developing new nomenclature and symbology
systems to use forward-looking rather than backward-looking constraints.
Having to rename many genes is difficult, especially so when all of the
"good" names have already been taken. Computers can always sort out
synonyms, but massive renaming always forces either a move to names of
less and less utility (as previously used names must be avoided) or to
significant homonymy -- something that computers do not handle well at