This is a follow on from a previous question I recently had... In a previous post on microevolution, Allistair posted (quoting part of his answer):

"Populations harbour a lot of genetic variance arising ultiomately from mutations. Mutations are arising all the time and ultimately proceed to fixation (the mutant allele becomes the only variant i the population) or loss (it disappears from the population). Selection plays a big role here but so does random chance. At any one time there will be lots of mutations at lost of genes at all sorts of different frequencies in a population. So it's definitely not a sequential process at all!"

How are variance and mutation distinguished? Is fixation the same as speciation? From wiki article on speciation it says that speciation follows from the extent to which a particular population is isolated. How often (and at what stage) does isolation need to occur to see significant differences in phenotype? Is there a case study for such a process? i.e. not analogy or plausibility argument.

Furthermore, I mean 'sequential' in the sense that an organism doesn't develop a (e.g.) limb after a single mutation, it will, of course, take many. At how many stages does isolation need to take place for the inherited trait to dominate a population? And whats currently the best evidence available for this?

(disclaimer!: i used the term 'macroevolution' in my previous post totally in reference to the berkeley 'understanding evolution' site that I was reading.)

p.s. is there a way to respond to an expert's answer as a dialogue?

Many thanks :)

Mutation is the process whereby DNA becomes altered.

Genetic variation among individuals is ultimately down to mutation but they are not the same thing. Imagine a population of 1000 individuals all with a genotype AA at a gene. A mutant then arises causing a new version (allele) of the gene, lets call it "a" and one individual carries a copy of this new version

We now have 999 xAA and 1 xAa and no aa genotypes. So there is some variation at this gene, but not much since nearly all individuals we might sample from the population are genetically identical.

Now imagine that some (maybe 10's-100's of) generations later, we look at the population again. It is stable in size so still 1000 individuals - but now the two alleles are equal in frequency - so we might find we have 250xAA, 500xAa and 250x aa. Clearly there is now more variation among individuals in their genotypes, and - if this gene actually influences something we can measure (e.g. body size) we might see more variation in that trait now as well.

So mutation is the ultimate source of (genetic) variation, but the amount of variation present will depend on the allele frequencies. Most mutations are lost by chance (a process called genetic drift), but some may be fixed (go to 100% frequency). So if in the above example all individuals at some stage in the future have genotype aa - then the original allele A has been entirely (and been replaced by the mutation). A mutation is more likely to be lost if it is deleterious (selected against) than if it is advantageous (selected for), but important to bear in mind that chance can be much more important than selection in determining the fate of new mutations in anything but the largest populations.

Fixation is not speciation. Fixation refers to the relacement of an existing allele (or alleles) at a genetic locus by a mutant (i,e, new allele) so that individuals in a single population become genetically identical (at that locus). Speciation occurs when these population-specific genetic processes (selection, mutation, drift) result in so much differentiation between populations (e.g. some pops fixed for some alleles, others for different ones, selection favouring some genes in one population, different genes in another) that individuals from the populations can (or will) no longer succesfully reproduce with each other. This is not generally thought to be a consequence of a mutation being fixed at a single gene, but rather an accumulation of genetic differences between populations at lots of loci (although there are likely exceptions to this generalism). Importantly gene flow (movement and breeding of individuals) among populations acts to reduce among-population genetic differentiation (and hence reduce the likelihood of speciation). This is why isolation is so important - isolated populations will tend to diverge genetically, but gene flow among populations is often sufficient to limit (or even completely nullify) this - meaning speciation will not occur.

Hope that's useful. I realise I have not answered all your questions so will try and respond to more later if other's don't beat me to it!

>p.s. is there a way to respond to an expert's answer as a dialogue?

Nope. Many other sites are based on interactive conversation, ours is a deliberately different model (q+a)!

Last edited by Alistair Wilson (17th May 2016 18:55:47)

How often (and at what stage) does isolation need to occur to see significant differences in phenotype? Is there a case study for such a process? i.e. not analogy or plausibility argument.

This is almost impossible to answer since there is no single "rule" here. Detection of "significant differences in phenotype" between popoulations is something we do routinely but is of course a function of how hard you look (e.g. how many individuals do you compare between each of two populations, how many traits do you look at).

We then use "common garden experiments" - bringing individuals from both populations into a single (usually lab) environment and ideally letting them breed (within-population) for a few generations so ensure the population differences persist. This tells us the phenotypic differences we see are genetically based (i.e. not totally explained by differerent environmental conditions experienced by the two populations). This type of experiment has been done literally thousands of times on hundreds of different organisms (if you want examples just google "common garden experiment" or "transplant experiment").

However, in any given case the answer to your question depends on a) the degree of gene flow between populations (and whether it is bidirectional), b) the stength of selection on the trait in the two populations, c) the (effective) sizes of the two populations (critical in determining the rate of genetic drift) and d) the genetic architecture (how much variation among individuals in a population is caused by genes, how many genes are there underpinning this, and what are their effect sizes) of the trait you are interested in. We have mathematical models that bring all this information together, but since all parameters vary across any context (set of populations of traits) you might focus on there is no single answer. Also worth remembering that evolutionary time is in generations - so all else being equal divergence happens much quicker in real time (e.g. years) for organsism with short generation times.

Last edited by Alistair Wilson (17th May 2016 18:57:39)