number of alleles of a gene

Dr N.I.Leaves nleaves at
Tue Dec 18 09:18:30 EST 2001

I think that logically I have to disagree with Brian's assertion that  -
"variation isn't directly related to length" although I have no 
expermental data to support my argument. 

If you take the extremes of the argument...I would say that there would
almost always be more variation in genes that occupy say 100kb of genomic
space, than genes which occupy 5kb. So there IS a fuzzy but direct
relationship between the number of variants of a gene (alleles) and
the gene size. Of course this is unpredicatable on a gene by gene basis
but on a genomic scale I would guess this is probably true. 

I dont think you could ever see a table which non-specifically correlates
gene size with number of alleles because of the variable selection
pressures which are exherted by coding DNA. Clearly, genes do change at
different rates and there will be plenty of examples of the small highly
variable genes which you could compare to big highly conserved
genes. But, logic tells me... that on average the reverse will be
true and that big genes (however you define them - big coding, big genomic
- with ~30000, definitions probably dont mattter) are more variable than


On 17 Dec 2001, Osborne, Brian wrote:

> Lin,
> There are two processes operating that generate the variation we see,
> mutation and selection. Mutation is thought to be random, more or less,
> though mutagens vary in their specificity. So the incidence of "silent"
> mutations, mutations that don't change protein sequence, could be roughly
> proportional to length, if by "gene" you mean coding sequence. However, it's
> clear that different genes evolve at different rates, meaning the effects of
> selection are different. So I'd say no, genes evolve at different rates and
> these rates aren't related to their length, meaning variation isn't directly
> related to length.
> Brian O.

Dr N I Leaves
Mouse Sequencing
MRC HGMP Resource Centre 
Cambridge CB10 1SB 
tel: 01223 494557 (office) or 01223 494541 (lab)
email: nleaves at

More information about the Autoseq mailing list