Re: Part 1 (of 3): What are major aspects of evolutionary theory?

> I don't know what you mean by "haplotype block".

Any single contiguous segment of DNA that appears to have evolved as a
single unit for hundreds of thousands of years according to the
preliminary haplotype-map summary report that I saw in _Science_ a few
weeks ago. Meiosis seems to have mixed-and-matched among the different
blocks, yet seems to have kept each single block intact for all that
time. At least that's the impression I got from reading the report.
I had always thought that DNA can break just about anywhere, that
some places it breaks slightly more often than other places, but in
general there's only a small variation between hot spots and cool spots
in regard to breaking and crossing during meiosis. But the report seems
to say that old idea was wrong, that there's such a drastic difference
between hot and cool spots that to a good approximation the chromosomes
are strictly divided into fixed blocks that act as units during meiosis.

SNPs can be tracked through the generations, to detect ancestry along
mixed male/female lines of descent, but because any one SNP has only
four possible states, the same state recurs at random many times, so if
you see the same exact rare SNP in two distant populations you can't be
sure whether there was recent gene flow between them or the same SNP
recurred by chance in both places.

But any single haplotype block has *many* different states, and each
evolutionary line accumulates occasional mutations which don't recur in
the same exact pattern along different evolutionary paths, so seeing
the same identical haplotype block in two remote populations definitely
shows recent gene flow between them, more recent than the last mutation
in that particular block within either population.

So if that's true, haplotype blocks offer a really powerful way to
detect tiny bits of gene flow between distant populations, such as if a
single explorer wandered from one to the other and left one or two
descendents who then left more descendents such that after hundreds or
thoudsands of generations a few of the haplotype blocks from that
invidiual explorer might remain in the population.

(And if two very similar haplotype blocks are found, it indicates
semi-recent gene flow, more recent than the bulk of mutations, but
older than the very latest mutations after the gene-flow occurred.)

The haplotype tool is much more powerful than the mitochondrial or Y
method, because a lot of different haplotype blocks are passed from a
single individual, instead of just one set of intact mitochondria or
one Y chromosome, and then these blocks are split from each other as
they are passed to subsequent generations, so an entire population with
a single common ancestor is likely to each have a few blocks from that
one ancestor mixed in with blocks from the many other ancestors equally
far back in generations. In theory, if we knew the genomes of *every*
individual at a certain time, then by matching haplotype blocks many
generations later we can determine exactly which modern individuals are
descended from exactly which ancestors. But more likely the best we can
do is measure the current genome of all living individuals in each
local population that is tightly inbred, and extrapolate back to the
ancestors who had the same set of haplotype blocks (modulo mutations
since then), and then detect any discrepancies from this neat model
which would indicate any gene flow from one such tight group to/from
another. If we could get DNA from several Neandertals, and thereby
deduce what haplotype blocks they had, we might in fact be able to
detect whether any modern humans have even one such
distinctly-Neandertal haplotype block.

> Ah, here you are apparently talking about frequency of recombination.
> Is "haplotype block" some kind of synonym for "linkage group"?

Checking online definitions:
<http://dictionary.reference.com/search?q=linkage%20group&db=*>

A pair or set of genes on a chromosome that tend to be
transmitted together.
Yes, that's part of what I mean, except I mean they are transmitted
together virtually without exception for tens or hundreds of thousands
of years, not just that they tend to be transmitted together slightly
more often than at randon over a single generation.

: a set of genes at different loci on the same chromosome that except
for crossing-over tend to act as a single pair of genes in meiosis
instead of undergoing independent assortment
No, I mean that *in*spite* of crossing-over they *nevertheless* remain
linked, i.e. crossing-over simply doesn't "ever" happen (not in a
hundred thousand years) between them.

n : any pair of genes that tend to be transmitted together; "the genes
of Drosophila fall into four linkage groups" [syn: linked genes]
No, not at all. That definition simply means they are on the same
chromosome (Drosophila has four chromosomes, as we all know from
high-school biology), so there is *some* (possibly very weak) linkage
between them, that they are not *totally* independent of each other.

<http://www.biochem.northwestern.edu/holmgren/Glossary/Definitions/Def-L/linkage_group.html>
A group of gene loci known to be linked; a chromosome. There are as
many linkage groups as there are homologous pairs of chromosomes.
No! That's most definitely not what I mean here. (Note that "syntenic"
means the same thing, but is unambiguous, applies to *any* two loci on
the same chromosome, regardless of whether the linkage is strong enough
to detect by standard tests, presumably Mendel-type breeding
experiments.)

<http://opbs.okstate.edu/~melcher/MG/MGW1/MG111122.html>
Ah, a fine Web site with lots of quotes to mine, such as:
* Markers that have measurable recombination frequencies are said to
be linked.
* Markers related through a chain of linkage constitute a linkage
group.
Ah, so a "linkage group" is a maximal connected set of markers (loci)
via the pairwise "linked" property (presumably meaning not easily
separable by recombination, which is easier for me to understand than
the verbage in the definition quoted). In most cases that would imply
that such a maximal connected set is a whole chromosome. It's roughly
analagous to a "ring species" which is a maximal connected set via
pairwise ability-to-mate, or maximal-connected-graph in any other model
that has binary links between nodes.
* As more and more markers are studied, linkage groups become larger
(and the number of groups smaller) until the number of linkage
groups equals the number of chromosomes.
So technically, "linkage group" is not globally defined for a species,
rather it's defined contingent on a given set of markers under study.

(SIde remark: When two ancestral chimp/human chromosomes joined
end-to-end to yield a single human chromosome, two "linkage groups"
were merged to form just one. In particular, genes that were near the
joining ends of the original chromosomes, suddenly changed from being
totally unlinked to being very closely linked. I wonder how such new
linkage affected subsequent evolution of those loci? In particular, I
wonder whether those two joined ends are now parts of a single
haplotype block or parts of two adjacent haplotype blocks?)

<http://66.102.7.104/search?q=cache:APrIlzHRp6wJ:hal.weihenstephan.de/genglos/asp/genreq.asp%3Fnr%3D519+linkage+disequilibrium&hl=en&ie=UTF-8>
Nice definition:
Linkage - An association in inheritance between characters such that the
parental character combinations appear among the progeny more often
than the non-parental.
Note that any linkage, no matter how weak, qualifies for that definition.
Connected in that way is not at all what I mean by "haplotype block"
wherein all characters (or loci or markers) are virtually 100% linked
over many generations.

And for fun, slightly side topic, see this attempt to clear confusion
regarding "linkage" vs. "linkage disequilibrium":
<http://linkage.rockefeller.edu/wli/lld.html>

Due to the ambiguity of definition, and the fact that the centroid of
the various definitions simply defines linkage group as the contents of
a chromosome, which is *not* what I mean, I think I'll avoid using
that standard term when I mean the more specific kind of longterm
linkage discovered by the haplotype-map project, where there are
several different such blocks on each single chromosome.

I'm going to cut my reply here, and respond to later remarks in another
posting.
..

.

Relevant Pages

  • Re: Making sense of HLA (I)
    ... A morgan is a length of recombination. ... had determined the 'linkage map' of the chromosome ... Did you know the word "linkage group" is the old ... Suppose I have a gene which is an allele at locus D allele 1 ...
    (sci.archaeology)
  • Re: New Scientist: The Neanderthal within
    ... in Piscataway, New Jersey, looked at a haplotype within a gene called PDHA1. ... which modern humans derive from a single population," says Garrigan. ... of selection wholly internal to the H. sapiens genome. ...
    (sci.bio.paleontology)
  • Re: Link to intriguing article about yet another leap within the genetics-explotion
    ... Haplotype ... Are Hamilton et al going to redefine one gene to be one ... > going to be able to force altruism at the Darwinian organism level? ... blocky genophenotype) to become the original ancestor of prolifically ...
    (sci.bio.evolution)