Can Genetic Loss Increase Diversity?

Great OP, Jazz.However, I have been mulling over Faith's (and to a much lesser extent MJ's) arguments in the previous two threads, and would like to start out with an observation.I'm not sure you've captured Faith's entire argument. Admittedly, she occasionally has some difficulty in getting across what she means. However, although it appears superficially that she and MJ are reading from the same sheet of music - and in fact she mentioned something along those lines herself at one point - there are quite substantial (albeit subtle) differences between their formulations.Faith has not argued consistently that loss of alleles leads to speciation - i.e., is a causal factor. I think she may have stumbled across that line once or twice, but the overall trend has been her claim that loss of alleles is an inevitable result of speciation. Leaving aside for the moment the mutation argument (which you have said doesn't really apply in this thread), her formulation is only wrong in her insistance on the "inevitable" part. In other words, there IS a mode of speciation that automatically results in could be considered "loss" of alleles, but in most other cases this is not the "inevitable", immediate result.Another subtle point on which she is partially correct, and partially incorrect, is her insistence on the time required for "beneficial" genetic sequences to appear in diverging populations. She seems to be saying that "beneficial" sequences are required for speciation to take place as the ToE suggests, and that there is insufficient time for such "beneficial" sequences to appear to off-set the "loss of diversity" caused by speciation. Both points are interwoven in her formula. Since she insists that only recombination of existing sequences can account for speciation, the ToE is wrong. So, restating her concept:Speciation = loss of alleles + requirement for beneficial sequences + insufficient time => an intrinsic barrier to evolution.My problem in discussing her ideas has been that I have consistently missed the subtlties. Simply dismissing her formulation out of hand as "creationist nonsense" doesn't work, because there are parts that she is actually correct on. It's taken me two whole threads to twig to my mistake. So what I'd like to do is discuss why her formulation doesn't work - and doesn't reflect the reality of speciation. Unless someone really needs them, I will avoid using the technical literature in support, and cover concepts in this thread rather than specific examples, which seem to serve only to take us off on tangents. What's right: What Faith is describing is one of the particular cases of the peak shift mode of speciation known as peripatric speciation - more specifically the subform of peripatric speciation known as the founder effect. In peripatric speciation, a portion of a source population "colonizes" a new habitat, becoming geographically isolated from the source. (Peripatric is derived from the word peripatetic - meaning wandering or traveling about; itinerant - and is very descriptive of what happens). There are any number of reasons why this might occur, from accidental dispersal to intraspecific competition, but in any case the "colony" represents a statistical sampling of the allelic diversity in the source population. If the new population is large enough, the population can contain all of the diversity present in the source. Smaller "bud" populations, on the other hand, may "miss" some rare alleles due simply to what is known as sampling error.The founder effect is an extreme, and rare, example of peripatric speciation. In this case, the “bud” is represented by at most a few organisms and often by a single individual. Obviously, this extreme example represents the ultimate bottleneck - and a statistically very small sample of the available diversity. This is the form of speciation that Faith is generalizing from.What’s wrong: Peripatric speciation doesn’t equate to "loss of alleles" as Faith insists - the alleles are still present in the source, and may be also present in the colony. Nor does the act of colonization automatically lead to speciation, and is in fact a normal dynamic of almost any population with any kind of dispersal ability (for those interested, a general discussion of source-sink dynamics can be found here). Speciation does not occur due to recombination of pre-existing alleles in the new population, especially in the founder case that Faith is insisting on (considering the dearth of alleles to begin with). Only in the specific instance of what Ernst Mayr categorized as “instantaneous speciation” can cytological changes create a new species (for example, by polyploidy or chromosome rearrangement). Even here, we’re talking at least the F1 generation - not the first, parent colony generation - and usually several generations. However, since these types of changes are mostly limited to plants and a very few animals, we can safely ignore them for the purposes of this discussion. I’m also ignoring asexual organisms, so don’t ask.What else is right? Loss of alleles CAN be the result of speciation. In small populations, the “drunkard’s walk” of genetic drift alone can cause the loss of alleles that are not under stabilizing selection. Drift can also increase the frequency of rare alleles. In fact, genetic drift has been postulated as one method for reproductive barriers to arise in a small population separated from the source (i.e., may be a cause of speciation itself). Since drift is essentially random (but see epistatic selection), it can result in loss of alleles. In addition, given enough generations, vicariant selection due to different environmental pressures, and in the absence of heterozygote incompatibility (which occurs in hybrid populations, for instance), alleles can be lost from a population even without them being different species. Ultimately the differences between isolated populations become great enough through this and the other speciation processes that we proclaim them varieties, subspecies, semi-species or true species - one or both populations has lost alleles over the generation to the point that they are incompatible.So what else is wrong? There is no requirement, however, that the “new” population lose the alleles - loss can occur in the ancestral population even if the colony retained all of the original alleles! Peripatric speciation writ large requires that the sampling error in the initial population be accompanied not just by changes in frequency (Faith’s claim), but by changes in frequency that lead to reproductive incompatibility between source and colony. In essence, then, either:1. Drift drags enough alleles over the generations to fixation that epistatic (basically, “linked”) effects at other loci cause incompatibility between the populations;2. Vicariant selection (adaptive divergence) due to the action of natural selection emphasizing or penalizing existing or new alleles does the same.Beyond the sampling error of peripatric speciation and the action of genetic drift, only over extended numbers of generations can loss of alleles be related to speciation. Hopefully this explains what is right - and what isn’t - in Faith’s formula.Making a long story short: whereas speciation may cause loss of alleles, loss of alleles cannot (to my knowledge) cause speciation. Mere changes in the frequency of alleles, except as noted, do NOT drive speciation.I'll save the "beneficial" part for another post if it really is necessary. The short critique is that "beneficial" is not a requirement - merely differences that lead to incompatibility. Indeed, rare deleterious traits can become fixed in a population through drift (as long as the pleiotropic effects are net positive - i.e., hitchhikers).Oh yeah. You may have noticed I didn't bring in MJ's formulation. Primarily because whereas Faith is right in some parts (at least as far as she goes), MJ is irrevocably wrong in just about every aspect - beginning with loss of genes (not alleles) causing speciation and passing through a complete lack of understanding of what speciation actually is. He's not in the same league as Faith, to be honest.Edited by Quetzal, : No reason given.

Not necessarily. The idea here is that frequencies of existing variation within a population can change due to drift of alleles that are generally neutral or even mildly deleterious. Let me see if I can explain it a bit better.Let’s say we have a population that has two loci: A and B. There are two alleles for each locus: A₁ and A₂, and B₁ and B₂. Let’s say the homozygous combination A₁A₁/B₁B₁ is “normal” with respect to fitness, but say the occasional A₂A₂/B₂B₂ is present but mildly deleterious in the ancestral habitat (or as far as that goes, could be simply different without either negative or positive effects). However, heterozygote combinations/hybrids are invariably deleterious in the ancestral “sub-1” environment. This would tend to keep the frequency of the latter fairly low. Then let’s say a random combination of a bunch of A₁A₁/B₁B₁ and A₂A₂/B₂B₂ found a colony. IF the A₂A₂/B₂B₂ homozygotes are favored in the new environment, then the frequency of the two alleles would change due to the action of natural selection, all other things being equal. Eventually, the “sub-2” population will dominate, and the “sub-one” alleles may eventually disappear from the “sub-2” population. This isn’t speciation, of course (since there’s no real barrier to reproduction with the ancestral population), and doesn’t include drift. Here’s the rub: what if A₁ is neutral with respect to fitness on a “sub-2” background? That it’s only when A₁ is combined with B₁ in a “sub-2” environment that the combination is deleterious? Since A₁ is neutral (not under selection), eventually some A₂ are going to be replaced by A₁ on a relatively random basis - through drift (alternatively, of course, A₁ could simply disappear the same way). IF the combination A₂A₁ is linked to other genes epistatically, then the frequency of those linked genes increases as well. The new genetic composition of the colony through drift may be the cause of incompatibility between the two populations - ancestor and daughter. It isn’t because natural selection is favoring the heterozygote A’s in the “sub-2” population - simply that the ancestral homozygous karyotype is no longer fertile with the “new”, epistatically-driven frequency. This is the ONLY conceivable way for a recombination of existing alleles to create the effect Faith is insisting occurs IN ALL CASES.

You got it. And this is in fact the issue. The case for this type of speciation is purely hypothetical - to the best of my knowledge, there has been no unequivocal observation showing that speciation by genetic drift is anything but theoretical. There have been some tantalizing hints, especially in insect studies, but no "for sure" example. So not only is this type of speciation not "the norm", but may not even really exist.

Hi Jazz, I just realized that this may require some clarification to make sure that all the bases are covered. In addition to what you wrote, the following are also necessary for speciation by genetic drift (at least using the hypothetical example I gave):1. The A₁ allele must be neutral with regards to fitness in the new (sub-2) environment. When the A₁ is linked to a B₂ backbone, it is not disvafored. During the time when the B₁ alleles are being eliminated, the A₁ must replace an A₂ allele.2. The A₁ allele must reach fixation in the sub-2 population. This is where the random walk of genetic drift comes in - the allele could just as easily be eliminated through this means.3. The A₂A₁ heterozygote must be linked to other suites of genes. As and if A₁ increases in frequency, the frequency of these other genes also increases willy-nilly.When these four (yours and the three I noted) conditions are met, the genetic composition of the colony MAY have changed enough to create a reproductive barrier with the ancestral population. And again, it may not.Hope this clarifies, rather than obscures.Edited by Quetzal, : No reason given.