FreeFall (webcomic) meets ecological genetics.
Cross-time Cafe chatI think I've mentioned FreeFall - a webcomic here previously. … No, maybe not - it's not in my tag list. Well, it's an SF webcomic that has been running for ... decades now? I think found it because got mentioned for having a use of "on the third hand" which is an obvious reference to Niven & Pournelle's "Moties". Is good. You read. (Don't be upset by the hosting site not doing HTTPS:// properly.)
In discussion of the genetics of one of the species involved ("Bowman's Wolves", below), a "rule of thumb" from ecology was mentioned, which I followed up. Someone called "Franklin" had a rule that for short term survival of a species, you need a population of at least 50 individuals, and for longer terms survival, at least 500.
I followed that logical thread, and Read The Friendly Paper. Below are my contemproaneous notes from RTFP (translated from the original PHPBB markup.
cite="MaverickMopete post_id=840776 time=1735693995 user_id=1060"
at least 500 Bowman's Wolves [the gene-engineered organism], according to Franklin's 50/500 Rule.
I don't particularly remember going "what?!" on seeing that the first time (whenever that was). Now I'm going "what?!"
The idea of effective population size (my emphasis) being different for short-term and long term survival sounds vaguely right. But it also flies in the face of the experience of every species that has originated (be that sympatrically or allopatrically), and most extension-of-range developments in mobile species. Living during a period of repeated ice age versus rain age conditions, sweeping across the globe and extinguishing one species making space for another to move in ... that's very challenging. Unless the emphasised effective population size comes into play quite strongly. One wolf making it across 500 km of hostile terrain per generation being enough to couple the isolated pack of a few dozen to a larger gene pool of hundreds or thousands means that the effective population size is thousands, not dozens.
Reference : I. R. Franklin, “Evolutionary Changes in Small Populations,” In: M. E. Soulé and B. M. Wilcox, Eds., Conservation Biology: An Evolutionary-Ecological Perspective, Sinauer, Sunderland, 1980, pp. 135-149.
PDF of the paper available at https://downloads.regulations.gov/FWS-R4-ES-2014-0065-0208/content.pdf
Some soundbites :
("Ne" is the effective population size, in contrast to the "census size") :
Before developing this argument in detail, we should ask if evolutionary change is what we want. Do we wish to conserve the elephant, or ensure the survival of its elephant-like descendants?
Speciation is discussed in Chapter 9.
(I spotted that connection immediately - but maybe that's my history of trying to inject brains into Creationists using the soggy end of a ripped-out thigh bone.)
We can distinguish between three modes of selection for quantitative traits. These are: (a) stabilizing selection, or selection against extreme values; (b) directional selection, in which one extreme is at an advantage; and (c) disruptive selection, or selection for extreme values, and against intermediates.
(a) was a big problem for Darwin, when the discrete, "atomic" (sense "undividable", not sense "components of molecules") nature of the "particle of inheritance" wasn't understood. Cue a monk in a garden in Brno. Most of the classic examples of evolution-in-action (and particularly artificial selection) operate in mode (b). Everyone who has tried hard to impress a "potential mate" only to see him/ her/ it/ them go off with "pug-ugly" has seen mode (c) in action.
Someone mentioned "uncanny valley" recently. Mode (c) poised to leap into action.
It is well known that populations which have been through a bottleneck or which have been maintained at a small population size do not show as great a response to artificial selection as do large populations.
This is one of the things that makes me dubious about the regular claims of "Mitochondrial Eve" (an unavoidable statistical event) and a human "population bottleneck" at (various times, around Toba (a major volcanic eruption about 70 kyr BP, or pick-a-number around 100 kyr BP) Both of which were certainly after the dispersal of Homo erectus across Asia, and one is almost certainly after Homo sapiens had reached towards Australia. What actually happened most likely wasn't a single, simple event. Which doesn't make for neat headlines. (Note : this paper is approaching it's 45th birthday.)
Skipping some statistics :
Effective number is decreased by increased variation in progeny number and, conversely, Ne, is maximized when all families contribute equally to the next generation.
That's one to put into your "Practical Starship Genetics" guide book. It also, effectively means making not having children illegal and immoral. "You don't own you genome, you borrow it from your grandchildren, and if they don't exist, you don't. Into the recycling tanks!" Hark, I hear the sound of a rotating Heinlein in his grave. No spreading your genes all over the galaxy!
Crow and Morton (1955) calculated from observed distributions of progeny number in a variety of species that the effective size ranged from 0.6 to 0.85 of the census number.
That gels with what I worked out from census data - in any normal year, between 10 and 15 % of women who reach menopause have had no children (and one infers the proportions are probably similar for males ; just not recorded with confidence until recently). That's using British data from a period when homosexuality was a serious criminal offence most of the time, so it's prevalence wasn't meaningfully recorded.
Unequal numbers of the two sexes. (...) A breeding population in which there are 90 effective females but only 10 effective males has a total effective size of 36, not 100.
The grave of Lazarus Heinlein would like to add the following : [rotating noises, subterranean]. Particularly since Heinlein (through Lazarus and other characters) was a great one for lecturing the readers on the genetic consequences of breeding choices.
Fluctuation in population size. If population size varies from generation to generation, the effective number is the harmonic mean:[snip maths] Suppose, for example, that a population which normally maintains an effective size of 1000 drops for one generation to 50. Then, over a ten generation interval, Ne = 345
WW1 was not that severe, but it's generation of "doomed youth" will probably have effects for generations to come. If you put it into an "American Indian etc" or "Australian Aborigine" context of repeated generations with maybe half the population of the previous generation ... yeah, that's going to be visible for generations to come too.
Immigration of unrelated individuals into an inbred population reduces the level of inbreeding dramatically
The "isolated Swedish wolves" example, or my suggestion for distributing 100 Bowman's Embryos through a slowly growing population seeded from 14 Bowman's Wolves.
Ah, this (p6) is where he gets to the "50" part of his rule of thumb. It's to keep the inbreeding coefficient below about 1% per generation. He doesn't cite any sources for his claim that Animal breeders accept inbreeding coefficients as high as a one percent increase per generation.
So someone should really check on that. I don't have a, ehemm, dog in this fight, so over to someone who cares more.
Hence, Wright suggests, the splitting of species into isolated subpopulations promotes evolutionary change
What I referred to as "allopatric speciation" above, from a palaeontological point of view.
The "500" figure for his "long term survival" Ne seems pulled pretty much from thin air with a hand-waving justification that a population that size will have reduced rates of genetic drift.
Ah - next page he gives a little more justification for "500" :
Let us ignore for the present the effects of natural selection and consider only the rates of gain and loss by mutation and drift. In very small populations, the loss of variability by sampling will be greater than the gain by mutation, and there will be a net loss of variation. Conversely, in very large populations, mutation will dominate the process, and we expect a steady gain. For each trait there will be a population size at which the rates of gain and loss are equal, and there will be no net change in the existing level of variability.
He then cites some 1976 numbers for the rate of decrease of variability in a Drosophila population to support his "500" number.
It's an argument.
Considering it involves biology (the "plus or minus 10%" science), it's got about the right number of significant digits (one). It might even be "correct". For some taxa in some environments.
Selection and linkage complicate this simple picture enormously, but I will attempt to show that even strong selection does not result in dramatic reductions of additive variance.
Tease!
Hammond (1973) collected some data which illustrate the effect of small founder populations very well. He established populations of Drosophila from one, 10 and 50 pairs, and then measured the response to selection (..) over 10 generations.
(...) realized heritability (a measure of the observed change) in populations established from 10 pairs differs very little from that in populations established from 50 [pairs]. As expected, populations founded from one pair, but not maintained at this level, showed three quarters of the maximum response
The metaphorical situation of a single pregnant human female landing on a raft on the coast of Australia and thereby founding the Aboriginal population - with the addition of, say, one further genome per generation - becomes far more reasonable than some people hold it.
Finally there is the question of whether to maintain a single large population or to split the species into a number of smaller breeding units. Such decisions will be primarily made on political or ecological grounds, but the latter course seems to have distinct genetical advantages. If a species is maintained in a number of small populations, not only is the danger of accidental extinction (for example, by disease) reduced, but an opportunity for local adaptation exists which may increase the chance of ultimate survival. Genetic drift can be countered by allowing occasional migration.
Which would be assisted, in the Jeanverse [the SF universe in which this takes place] , by the relative difficulty of moving between Petri dishes using the DAVE FTL drive. [Dangerous And Very Expensive ; Faster Than Light drive]
Well that was a very worthwhile read. Sorry for monopolising this part of the thread - I find that papers stick in my head better if I make notes on them as I'm going along.
Somehow, I suspect the Mark [author of the comic] has read this paper through several times, and it has deeply influenced his world-building.
And now that I've copied this self-conversation to my own space, I can edit the original to oblivion if requested.
Is there such a thing as a PHPBB to HTML converter? Maybe.No, not really. Oh well, it wasn't too bad.
No comments:
Post a Comment
Please add any useful comments you have. Some HTML allowed. All comments go through moderation (because : spambots).