October 27, 2005

Chromosomes and Evolution

I have just learned something new. Or, rather, become newly aware of the implications of some things that have been rattling around in my mind for a while. Greg Cochran linked (indirectly) to this quote:

The loci in question are so tightly linked that rare recombinants practically never arise - this explains why the different multi-locus genotypes appear, when crossed, to segregate like single locus genotypes. A set of genes so tightly linked that they behave like a single locus has been termed a supergene.

This was a eureka moment for me. I have sometimes wondered about the evolutionary implications of chromosomes. I'm sure that there's a molecular reason for them - certainly, it would be hard to imagine a diploid genetic architecture, necessary for sexual reproduction, without them! But having said that, it would seem that chromosomes only get in the way of sexual reproduction: If sexual reproduction is about facilitating genetic recombination, then more would certainly be better than less, and we know that many other species have many more than our 23 pairs: horses have 32, dogs have 49, ferns have 630! So why haven't we evolved the maximum possible number of chromosomes?;It's certainly possible to have a lot more chromosomes than we have.

Clearly, it seems to me, the answer is that sexual reproduction is not always a good thing. Rescrambling our genes every generation has the effect of breaking up favorable combinations of genes, so it must be that a small number of chromosomes is an adaptive response to this. Genes on the same chromosome get rescrambled not every generation, but once out of many generations, with genes closer together getting rescrambled less often than genes farther apart. The infrequency of the rescrambling makes time for selection to weed out unfavorable linkages as they arise. 

Some predictions:

1. Linkage disequilibrium is not necessarily a sign of recent positive selection - it could also be a sign of coadapted gene complexes.

2. Coadapted gene complexes that involve genes on different chromosomes would have to be much more advantageous than those involving genes on the same chromosome, in order to be maintained.

3. The advantage necessary to maintain coadapted gene complexes varies according to the physical distance on the chromosome of the genes involved. (I know it's a bit more complicated than that, but roughly.)

4. A reduction in the number of chromosomes could be adaptive if it locks-in a favorable gene complex.

5. A coadapted gene complex could also explain this, as these genes don't recombine.

6. This could be part of the answer why we have sex so often (i.e. more often than models would predict)!

PS: This is another example of the importance of tradition.

(Cross-posted at Gene Expression.)

Posted by David Boxenhorn at October 27, 2005 10:57 AM | TrackBacks
Comments & Trackbacks

I'm afraid your idea is competely wrong. In fact, it's diametrically opposite to the truth.

Have you heard of Barbara McClintock? She won a Nobel Prize for Medicine for her breeding experiments with corn. What she was really doing was to track descent of different genes and trying to correlate them.

What she demonstrated statistically has now been understood biochemically. When a cell divides, its chromosomes are torn apart and the DNA is duplicated. Then the DNA is reassembled into twice as many chromosomes.

But none of the new chromosomes is exactly the same as any of the parent chromosomes. Each child chromosome of a particular kind (e.g. human chromosome 7) is made up of pieces of both parent chromosomes of that kind. The interesting fact is that no two cells in your body have identical chromosomes; every one is different -- and not just because of mutation.

Genes which are physically near one another on the same chromosome tend to stay together in this process. Genes on opposite ends of a given chromosome have no better than a 50:50 chance of remaining together on any of the daughter chromosomes.

The "supergene" complexes to which this article refers are ones where the genes are all extremely close together, which means they tend to stay together.

By the way, this effect means that in evolutionary terms the number of chromosomes doesn't really actually matter much in terms of reproductive variability.

Posted by: Steven Den Beste at November 3, 2005 03:57 AM Permalink

The chance of crossover (50/50, you say) is higher than I thought, but I don't see how that invalidates anything I said.

Posted by: David Boxenhorn at November 3, 2005 07:21 AM Permalink

I just left this comment at Gene Expression:

Steven, Thanks for pointing that out. Although it would mean that the specific number of chromosomes not very important (though if 50:50 is correct it would still have a noticeable statistical impact), it doesn't invalidate anything else I said. Specifically, that arranging genes in chromosomes results in a spectrum of gene-gene relationships from totally sexual to almost totally asexual (for really close genes). And it doesn't invalidate any of my predictions (some of which have been confirmed by commenters).

Posted by: David Boxenhorn at November 3, 2005 07:40 AM Permalink

× Network: