I've been rather halfheartedly drafting a reply to Zauis
Message 44 for the past few days, but the discussion seems to have moved on quite a bit in that time and there were some sections relevant to the current direction, and particularly Tangle's question in
Message 80, that I thought I would extract and present here.
We were discussing peripherally the action of the
FGF4 retrogene insertion in short legged dogs and the fact that the genetic makeup necessary for such an insertion was not found in other modern dogs but rather in a grey wolf population. Zaius suggests that this may be a frequently occurring spontaneous mutation in the wolf population, which is merely strongly selected against, and so cannot really be considered novel.
Zaius writes:
This does not sound unreasonable because if the retrogene appeared in a wild wolf that wolf might not be alive long enough to enjoy benefits of short legs.
That is a perfectly reasonable argument and raises the question of how novel a feature can be considered when it has a low but significant rate of spontaneous appearance in a parent population, even if it is subsequently strongly selected against as in Zaius' example. Should we consider it novel simply if it was absent in the founding population? Taken to an extreme the alternative approach rapidly takes us to a position where our imperfect knowledge of the genetic makeup of previous generations makes it impossible to definitively consider any mutation novel as we can't confidently state that it never occurred transiently in any ancestral population.
I'd say that for pragmatic purposes we would have to look at simply the extant variation that we can see and in the case of a line where we actually known the genetic makeup of the founders it is reasonable to consider a genetic feature novel that was not present in that founder population, I say a genetic feature rather than a phenotype since recessive traits could easily be masked for few generations.
I think this is relevant of the discussion of the melanic mice because there are several other examples of adaptive colouration in other mouse species where the mutation has gone in the other direction from dark to light (
Hoekstra et al., 2006;
Linnen et al., 2009;
Steiner et al., 2009).
Without a much larger scale phylogeny taking in the relatedness of these several rodent species, it is very hard to say which trait is truly ancestral and which derived and certainly not if the trait is novel in a broader sense. I would say that we
can pragmatically determine given a certain depth of genetic information whether a trait is novel to a particular population as I outlined above.
The above papers are also interesting as they touch on Zaius' challenge to me to show him some examples of adaptive mutations that weren't 'degradative'. As is often the case this is problematic as Zaius hasn't given us a clear definition of what he considers 'degradative', indeed he has made comments that suggest he might be using the term so broadly that
any genetic change will fit it regardless of its effect.
So within these various papers we have the initial example of the light to dark mutation set in the pocket mice. The function of the melanocortin receptor is to mediate the signalling from certain hormones which bind to the receptors, this mediation is done via the generation of cyclic-AMP (cAMP), the level of cAMP then determines whether the target cell produces eumelanin (Darker) or phaeomelanin (Lighter). So in terms of the genetic concepts of loss- or gain-of-function the melanic pocket mouse mutations in
Mc1r are gain-of-function mutations giving rise to a more sensitive receptor balanced more towards eumelanin production (
Nachman, 2005).
However when we look at the other papers we see a light phenotype arising both from gain-of-function mutations increasing the expression of the
Agouti gene, which inhibits the binding of melanocortin to its receptor, and through loss-of-function mutations where the binding specificity of the receptor for melanocortin is reduced and so is the level of cAMP, tipping the balance towards phaeomelanin production.
So we have a very heterogeneous mixture of mutations, albeit all centered on one or two genes, some are gain-of-function mutations, some are loss-of-function, In some populations the derived mutations are those producing the light colouration, in others the dark and in some there are a mixture of derived alleles in both genes giving rise to a spectrum of melanic phenotypes.
So given all this messy complexity how do we address Tangle's question, how do you distinguish the new, derived allele from the ancestral one? This usually comes down to being able to compare to another population considered to have common ancestry (which is undoubtedly an unfounded assumption in Zaius opinion). Polymorphisms/mutations which are found to be present in either of your populations of interest and your more distant outgroup are considered to be ancestral and those found in only one population to be derived. Since there is no way to be sure that you have captured the entire genetic diversity of a whole population in these sorts of cases these are always probabilistic inferences.
TTFN,
WK
Edited by Wounded King, : Spelling
Edited by Wounded King, : No reason given.
™ Version 4.2
(1)