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

anon1@xxxxxxx wrote:

>>>I've looked at the FAQ again just now.
>>
>>"The" FAQ? There are dozens (hundreds) of FAQs. Huh?
>
> The talk.origins FAQ, you stupid ***.
> The FAQ for the newsgroup where this thread is posted.
> Which FAQ for which newsgroup do you think I was referring to anyway?

Sorry, I thought we were talking about the talk.origins archive. That's
what I was talking about. No need to get so testy. Take a deep breath
and count to 10.

>>>I have a problem: I don't believe there is yet any evidence for
>>>Universal Common Descent, only for very-large common descent. We might
>>>be able to trace each of the three domains from a single origin (each)
>>>to all modern species of such domain, but there's no clear evidence
>>>that the trunks of those three large clades can be traced back to a
>>>single all-encompassing common ancestor.
>>
>>Yes there is. The evidence is of a somewhat different sort than the
>>evidence for the rest of the tree (roots are always a problem), but it's
>>there.
>
> Are you claiming that a single species was the LCA of all domains of
> life, that the tree can indeed be traced (with current evidence) back
> to a single (one, uno, ein) root species (or strain of bacteria)?

No.

> Or are you allowing that the idea of a tree breaks down beyond the
> three separate domain roots, that there may have been a mixmash of
> multiple ancestors into *each* domain beyond that point? For example,
> could there have been pre-prokaryotes A B C D E F G and then
> the three domains formed by merger of those genomes:
> - archea from A + B + F + G
> - eubacteria from C + D + E + F + G
> - pre-eukaryotae from B + C + E + G
> Then much later endosymbiosis created the true eukaryotes as we know them.

Yes, that's possible.

But whichever, if there were an A-G or not, they were all related by
common ancestry. That much seem clear because of the various universal
features.

>>And without arithmetic, we can't do radioactive dating. Is it therefore
>>necessary to discuss the basics of arithmetic?
>
> Perhaps I should include one extra preliminary question, as to whether
> the person has graduated from eighth grade or not. If so, then
> presumably the person knows arithmetic and reading and writing and
> other basic stuff needed to even understand what we mean when we ask
> about atoms and isotopes and radioactive decay. I wouldn't break down
> the question into each particular pre-8th-grade subject since virtually
> everyone would answer YES to this question.

I really don't think you have a very clear idea of what you are trying
to do. If you had such an idea, it would enable to focus your aims a bit
better.

>>>I personally think a four-taxa unrooted tree is the best starting
>>>example, and whoever is maintaining that example should change it to
>>>four taxa if not already done.
>>
>>Nobody is maintaining it. It's a POTM, frozen in the form in which it
>>first appeared.
>
> Then perhaps I might undertake to rework it to use only four taxa, with
> only three possible ways to make the 2:2 split for the unrooted tree.
> Or do you assert copyright on the work and refuse to grant me license
> to create a derivative work?

I don't assert copyright. But I already turned it into a 4-taxon
example. Don't ask me to find it, but it's somewhere in Google groups.

>>And you think the way I did it was too complicated?
>
> Yes. For a very first exercise in using DNA/genome sequences to support
> an unrooted tree, I think using five species was unnecessarily
> complicated. Four species would have been less work on the part of the
> reader, and increase likelihood of reader-success.
>
> Too bad I missed the vote. I would have voted for four instead of five
> species. Do you happen to know the message-ID of the call-for-votes so
> I use Google to look it up and see what I missed?

No.

>>>Or maybe nobody suggested my fine idea before. I don't remember whether
>>>anybody thought of exactly that idea before.
>>
>>Yes, it's been suggested before. Any tree can be assembled from a set of
>>quartets. There's a program that tries to do that, called
>>"quartet-puzzling". Here, though, it's silly, one reason being that one
>>of the two nodes just isn't well resolved by the data at hand.
>
> So you're saying that even though we know from other studies that
> *both* 3:2 splits are well supported:
> chi+hum+gor / ora+gib
> ora+gib+gor / chi+hum
> the first of those two is supported by that particular sample evidence
> whereas the second isn't supported there and would require some
> different evidence?

Yes. And in fact there are sequences that strongly contradict the second
split, and put either gorilla with human or gorilla with chimp. This
suggests that the branch involved was short enough for differential
sorting of ancestral polymorphisms (which you might want to look up);
however, the clear majority of loci that make any claim support the
chimp-human clade.

> If you delete one of the three African apes from
> the set, can you use that particular evidence to support the one 2:2
> split that remains from the first of the 3:2 splits above?

Yes.

> AfApe#1+AfApe#2 / ora+gib
> Which of the three African apes is best to delete from the set, so that
> the evidence best supports the remaining two as distinct from ora+gib?

There is no difference. But of course you would want to keep the human
and eliminate one of the other two, because otherwise the title of the
piece is wrong.

> IMO, after that first four-species example is presented, there should
> be ideas for followup experiments, such as picking a different African
> ape to omit, or using different portions of the genome for comparison.
> By the way, I know that virtually 100% of the human genome has been
> sequenced, and most of the chimp genome likewise, but how much of the
> other three species have been sequenced so-far?

A lot of sequences, probably totalling hundreds of thousands or even
millions of bases, but still a small proportion of the genome. Thus
according to you (in another post) there can't be any real evidence for
human evolution coming from DNA sequences so far.

> And for how many DNA
> bases is there unique alignment across all five genomes?

Never looked, but I would suspect it's at least in the hundred
thousands, if anyone cared to put it all together, which I don't know
that anyone has ever done.

> (After I get those answers, then I have followup questions such as if
> you divide the fully-five-aligned portions of the genomes into sections
> of reasonble size, such as breaking the sequence whenever you've
> accumulated one hundred SNPs, what fraction of these segments show
> support for one or the other of the two 2:3 splits (or 2:2 splits with
> one species omitted)?)

Sorry, not clear which node you are talking about here. Pretty much all
of them will split the African apes from the others. My guess is that
around half would split the human and chimp from the others, another
quarter would show one of the two other resolutions of that node, and
the remaining quarter wouldn't show anything much for the node (as is
the case with the mtDNA data I posted). But that last is really just a
guess.

.