| Thursday,
April 6, 2000 I'm going to move all
this thread to its own page when I get a chance.
Jerry,
Leroy Ortiz writes:
"...human hemoglobin is a string of 574 amino acids of 20
varieties. There are 10E 654, different ways to combine these amino acids.
That is kind of meaningless all by itself, but consider this: It is
thought by evolutionary proponents that life began 2,500 million years
ago, which is still a very long time. However, the number of seconds since
life began is 10e 17. The number of seconds since the big bang 5000
million years ago is 10E 18. The number of stars in the universe is 10E
22. The number of atoms in the universe 10E 80. The improbability of
hemoglobin occurring by random selection is 10E -654."
Short answer from a molecular biologist (me).
It can't have happened by random selection, but
that's not what those of us in the molecular biology community think had
to happen.
First, life arose quickly after the earth's crust
cooled, but still a very long time ago. Cooling of the crust was 4e9 years
ago; convincing microbial fossils in Australia are 3.5e9 years old. Schopf
has given a good popular account of his work on this (_Cradle of Life: The
Discovery of Earth's Earliest Fossils_, by J. William Schopf. Princeton
Univ. Press, 1999).
We have some partial answers about how this could
have happened naturally. Our understanding of the very first steps is
still quite poor -- i.e., how you get from dead mud to *some* very
primitive cell. But we do know the following things that make that cell
somewhat easier to imagine than formerly.
You don't need to jump immediately from nothing to
proteins plus RNA plus DNA. Having RNA alone is a viable basis for early
life, because it turns out that RNA can be both hereditary and an enzyme.
Specifically, it appears very probable that the core of the ribosome (the
anvil on which modern proteins are forged) is actually, even in you and me
now, an RNA enzyme and not a protein one. Other natural RNA enzymes are
found in splicing of messenger RNA and in an E. coli enzyme called "RNAse
P". More generally, it is reasonably easy to create RNA catalysts and
binding sites in the laboratory for a great variety of purposes. For
details, see: _The RNA World_, 2cd. ed., ed. Gesteland et al., Cold Spring
Harbor Press, 1999.
Once you have RNA, you don't immediately need to
construct hemoglobin; you probably need to build up a shell of proteins to
gradually replace preexisting RNA catalysts and extend their functions.
Even today, the total number of proteins required for minimal cellular
function is less than or equal to 350 (Hutchison et al. [1999], Science
vol. 286, pp. 2089-2090). If you have preexisting RNA catalysts you can
build that slowly.
>From 350 proteins to the 14,000 of a fly or the
perhaps 70,000 of a human is much easier to do than it sounds, because
protein-encoding genes can duplicate and the duplicate copies can then
diverge in function. It has been known for many years that the alpha and
beta subunits of hemoglobin are similar (both in amino acid sequence, and
in three-dimensional structure) to both one another and to the simpler
protein myoglobin. Extensive phylogenies have been reconstructed for these
proteins from many species on the basis of protein parsimony (constructing
an evolutionary tree in which modern proteins are given branch lengths to
one another than minimize mutations between them, following Occam's
Razor). Trees constructed in this fashion closely resemble, in their
organismal clustering, the phylogenies previously devised for organisms on
the basis of classical taxonomy. More importantly, such trees show the
subunits of hemoglobin diverging from one another after splitting off from
myoglobin but well before the emergence of modern animals. A fine, though
slightly dated, treatment of this remains Dickerson and Geis' _Hemoglobin:
Structure, Function, Evolution, and Pathology_, by Dickerson and Geis.
While it *is* necessary for a protein to be
constrained in its sequence, it is not true that (say) modern
alpha-hemoglobin in a human being must have exactly one sequence to be
functional at all. We know from comparisons of well-characterized protein
families that very large divergences in amino acid sequence (up to 50%
dissimilarity) are entirely consistent with conserved structure and
function. At the same time, while mutation is random, selection is not:
bad mutations are "destroyed" immediately (with the deaths of
their carriers), neutral ones are statistically likely to be lost through
drift alone, and rare good mutations may or may not end up being fixed in
a future species. The "motion" of a protein through
"sequence space" via evolution is therefore neither an
infinitely narrow track nor a purely random one. Again, Dickerson's work
is a good review of this applied to the protein Ortiz cites. A more recent
and general treatment is Graur and Li's _Fundamentals of Molecular
Evolution_, 2cd. ed., Sinauer Press, 1999.
Most proteins are to some degree replaceable by
other proteins. Recently a team of computational biologists in Bethesda,
MD did a study of some two-dozen microbial genomes to try finding what
proteins were *absolutely* conserved in all of them, without a single
species exception. The final number was 81 proteins:
http://www.ncbi.nlm.nih.gov/cgi-bin/COG/palog?phy=amtkyqvcebrhujgpolinx
Most of these are involved in protein translation,
and some of them could well have been duplicates of one another very early
on (e.g., two different GTPases involved in protein translation).
So the task of explaining very early life really
involves explaining how a primitive RNA life arose and then how it
acquired a set of 80-350 proteins -- but with the caveat that the proteins
could be acquired incrementally, and that it was not at all necessary that
their sequences be "perfect" or "exact". It is a hard
problem, but it is not a problem that actually involves explaining the
*random* occurrence of *one* configuration out of 1e654. It is also a
problem that we have made real progress on: absolutely nothing that I have
described was known, or even accessible to science, a mere century ago,
and much of it was not known until a few decades ago at most.
--Erich Schwarz
Well, yes; but as we find out more, we also see
more problems. From the outside it looks as if every new discovery posits
at least three more stumbling blocks while clearing out one or two old
one; the clearing out raises the level of faith that the new ones will
also go away. Perhaps I read it wrong, but I have been watching this for
about fifty years now.
Jerry,
Having some knowledge of the theory and with some
experience modeling evolutionary systems, I can provide my perspective on
the problem.
Question: How fast is evolution? Answer: Very, very
fast. Suppose you have a wild-type gene and an alternative allele with 1%
selective advantage in a specific environment. The alternative will evolve
to fixation in a few hundred generations. For prokaryotes, this could be a
few days; for humans, 2500-10000 years. And that isn't much of a selective
advantage. The only thing that keeps this process from going to completion
globally is geographic or very rapid temporal variability.
Question: What about Lamarckian evolution? Answer:
Actually, cultural coevolution is Lamarckian. Darwinian evolution goes to
fixation very quickly, even when there is no selectively advantageous
allele, simply because the strongest selection gradient is associated with
mutation rate--low mutation rate alleles waste less zygotes on monster
mutations. Lamarckian evolution is dynamic. Mutation is not random, and
the system can evolve chaotically. The evolutionary game underlying
territoriality happens to be chaotic (expressed in bluffing strategies)
when allowed to evolve culturally. So humans have been evolving
chaotically for a long time. Given that mammals learn, first from their
parents and later from their social group, I suspect some aspects of
mammal behavior have been evolving 'culturally' for a long time. Most
people are unaware of the sophistication of basic mammalian intelligence.
Question: What about macroevolution? Answer: There's
evidence that life started up very rapidly--within 400 MYr of the end of
the era of bombardment. There is fossil evidence in the genome and in
basic cellular metabolism that suggest life emerged around a set of
chemical hypercycles that may have been driven by elevated levels of
carbon dioxide in the environment in a direction reverse to that seen now.
That is, the metabolic cycle originally went in reverse and was associated
with efficient chemical breakdown of more complex carbon compounds to
carbon dioxide and water. Also, interestingly, the genetic code maps in a
very natural way to the metabolic cycle, with amino acids being produced
in similar ways corresponding to similar codons. Hence, life is probably
not a random phenomenon, either in origin or in history. This universe may
have properties that encourage the emergence of simple life in many common
environments. I believe Harold Morowitz may have written something on
this.
Whether intelligence is also inevitable is another
issue. I suspect not, based on currently emerging theories about the
Cambrian transition.
Harry Erwin, Ph.D., Computational Neuroscientist
(Bat Behavior) and Adjunct Professor of Computer Science, George Mason
University
I note that while everything about microevolution
-- within species -- is pretty well consistent with what's known, when it
gets to macroevolution there's a lot more handwaving, and speculation
about what might have been (and in the case of hard core evolutionists,
what must have been). What we don't have is much in the way of hard
evidence. Some species paths are easy enough to see (although even there
we don't have much in the way of intermediate forms). Some transitions
from one species to another go through stages that hardly seem viable at
all, much less viable in a competitive environment; yet we don't see how
to get from one to the other without passing through those disadvantaged
forms, nor have we any evidence those intermediates ever existed, or ever
could have existed. This is a non-trivial problem, one that has driven
people like Hoyle to postulate "evolution from space" and the
like.
The interesting part is the number of people who
insist that this knotty problem is settled one way or another, and who
will use political mechanisms to have their particular view taught to the
exclusion of others. Me, I'm quite happy to realize there are many
thing I don't know and may never know. But that view isn't popular with
either camp...
Hi Jerry,
Sorry to keep hammering on this, but, well, I do
this stuff for a living and in some sense I'm obliged to make a case for
it.
Wil Spruth writes:
"Repeat this 10 times, and the homo sapiens genome differs from
the australopithecus genome by only 350 bits."
Actually, that's pretty close to being the
scientific view: DNA sequence comparisons of human with *chimp* genes are
easy to do now, and the results of my own random checking is that they are
routinely 99% identical. Many of the differences are actually likely to be
neutral polymorphisms fixed in populations by statistical drift, not
phenotypically significant. The mutations required to make a human from a
chimp are probably closer to 350 bits than 3500 bits.
Which specific mutations *are* needed should be
known once we have the genome sequences of both humans and great apes, and
can compare them. Very little work as yet has been done to pin this down,
but two cases have been turned up: a luteinizing hormone in chimps has a
higher affinity for its receptor than in humans (and thus we humans may
have a "tardy" puberty, allowing greater time for parents to
teach their offspring), and humans lack a sialic acid present in all other
primates. There will be other differences, but they aren't going to be a
vast number -- we already know from DNA that they can't be.
Science is scary because the more we know about our
world, the more precarious and precious our existence seems. There is a
very small gap between us and "brute" apes, and not really much
more of a gap between us and lifeless matter. If there were interventions,
fine -- but I don't think you need to postulate them to explain the world.
I think the world may just be beautiful and terrible, on its own steam.
--Erich Schwarz
We have no argument about that last paragraph.
And thanks for the summaries.
I find this discussion interesting precisely
because there is not enough evidence to "settle" it and say that
those who don't agree fall in the same category as the Flat Earth Society
believers (although I gather most members of that outfit don't really
believe...) Yet, although smart people can disagree on
macroevolution, there are those on BOTH sides who would simply mandate
that their view, and only their view, shall be taught and discussed in
schools. THAT is what I find interesting. As if the fate of the world
depends on choosing the right answer. Well, a day.
Now for something complete different:
Jerry, we've just put up a new feature on PC Pitstop,
called Disk Health. Once you've downloaded the control (which we don't
expect to change often) it's a 10-second check of your PC's disks for FAT
copy mismatches, FAT chain errors, cross-linked files, and fragmentation.
We developed this in partnership with Golden Bow, of Vopt fame.
The URL for the Disk Health quick check page is http://www.pcpitstop.com/pcpitstop/diskhealth.asp.
We've also incorporated the Disk Health checks into the full PC Pitstop
test sequence, at http://www.pcpitstop.com/pcpitstop/.
We recommend that people check their PC's Disk Health at least weekly, and
run a full PC Pitstop test at least monthly.
As is the case for the PC Pitstop tests, the Disk
Health test uses an ActiveX control and JavaScript, which means that it
requires IE 4+. Right now it works on Windows 95, Windows 98, and Windows
Me.
Best, Martin Heller
Web Developer and Founder, PC Pitstop LLC Owner,
Martin Heller &; Co. Senior Contributing Editor, WINMAG.com and
BYTE.com Reviewer, PC WEEK and Government Computer News
http://www.pcpitstop.com
http://wintune.winmag.com http://www.byte.com
Personal/Consulting mail: meh@mheller.com PC Pitstop
mail: martin@pcpitstop.com BYTE.com mail: mheller@cmp.com
I haven't tried this, but I will. If you've been
working with Barry at Golden Bow we can be sure of one thing, it won't do
any harm. I've used VOPT for maybe twenty years without losing one byte of
data because of it. Thanks/
Subject: Jury duty, FootFall, and more.
Dear Dr. Pournelle, You said "They actually
wanted me on a jury", are you on it? If so, when will we hear about
your experiences? Even if the case wasn't interesting, your observations
probably would be.
I have been re-reading FootFall for the nth time
this week and got to thinking. It's set in the mid eighties so, by now,
some of the fi who had surrendered would have had children. Who would have
been born in the U.S. Thus making them citizens. After all they would be
persons born in the United States, no? Possibly interesting idea for a
short-to-medium story sequel. Also, related to FootFall, I noticed how out
of date it was in terms of international politics. References to the
Soviet Union seem so outdated today, and I wonder why no one saw the
collapse coming. RAH wrote a report on a trip that he took to the USSR,
which I read in "Expanded Universe", where he reported on
economic and other failures of the Soviet system. Why, given that the
problems were well known, were we all so surprised by the sudden collapse?
Not seeing the forest?
On Microsoft. There seems to be quite a bit of
/anecdotal/ evidence that Microsoft stifled innovation. Not directly, but
through fear that, if Microsoft entered a market, they would kill the
companies that were already in it, thus keeping people from investing in
companies that Microsoft /might/ compete with someday. And, also, keeping
companies from developing products that Microsoft might want to compete
with. Call it the "Netscape Lesson". I feel that the AT&;T
and IBM antitrust cases were, in the long run, good for consumers.
Certainly the markets are more confused, I sometimes miss the convenience
of being able to call the "Phone Company" to get a problem taken
care of instead of trying to determine which phone company to call. And
the bills were certainly simpler, also higher. Would we have multiple
phone lines, DSL, Caller ID, Call Waiting, etc. without the AT&;T
antitrust case? Would the PC revolution have happened, as quickly, without
the IBM antitrust case? The antitrust case has acted as a restraint on
Microsoft which, even if overturned on appeal, will be good for us.
Finally, the recent column by George Will, on
capital punishment, seems to indicate that the mainstream Republican
hierarchy may be developing some serious doubts about the utility of
capital punishment. Here's the URL http://washingtonpost.com/wp-dyn/opinion/columns/willgeorge/A20906-2000A
pr5.html
Interesting.
Sincerely, Kit Case kitcase@netutah.com
The legend is that the automobile industry
"stifled innovation" by preventing the Tucker and the Pogue
Carburetor from coming to market. Perhaps. But I am not sure just how much
more innovation this industry could have handled: we went from 64K systems
with CP/M and command lines (PIP A:=B:*.*) to big graphical
interface systems in a decade. Apple had lots of innovation.
What concerns me about the Microsoft case is not
the "health of the industry" but the lesson we taught: pay
tribute to Imperial Washington or wish you had. And every one of Penfield
Jackson's findings of harm to consumers talks about what was done to a
competitor. Every bit of it. The only "harm" to consumers is his
speculation that Windows might have sold for $40 instead of $80 at some
points in time. I hadn't realized that the Constitution gave Washington
the power to fix prices except as part of a wartime situation.
It's easy to speculate what might have happened
without Microsoft. But I can speculate that absent Microsoft the IBM model
in which our small machines are "entry level" systems,
deliberately CRIPPLED (recall the PC Junior?) so that we would "Move
Up" to "Real" computers might have prevailed. IBM was a BIG
outfit with much power in those times. Gates was an upstart with a queer
vision of "a computer on every desktop and in every home and in every
classroom". This was considered insane by all the computer
professionals. Now it's the standard model for us all. THAT is what I call
innovation.
As to the jury, we went in the next morning, sat
in the hall for 2 hours, and were called in to be told that the defendant
had pleaded guilty to a reduced charge. No trial, we were dismissed, and
that's it for 6 months at least....
TOP
|
CLICK HERE for
a Special Request. 
