Text 697, 200 rader
Skriven 2004-11-07 21:59:00 av Tinyurl.Com/uh3t (1:278/230)
Ärende: Re: Metabolism Forced
=============================


> From: Brett Aubrey <brett.aubrey@shaw.ca>
> I would have assumed that more than "very simple chemicals" could be
> cooked up due to solar, volcanic (and vents), lightning, impacts,
> etc.(?), and that this might possibly happen far more often than once
> every few million years.

Maybe I should have omitted the word "very". The great bulk would be
truly VERY simple things like water, ammonia, formaldehyde, cyanide,
etc., but a significant fraction would be semi-simple things like
simple sugars and amino acids. Sometimes two amino acid molecules might
by chance link up to form a dimer, and if that happened often enough
there might be millions of copies of each particular two-acid sequence.
Longer sequences would likewise form by chance, or clumps not in a
chain, but any particular combination of ten or twenty units would be
so unlikely to happen that only one of it would exist at any one time,
in the absence of any replicator that liked to repeatedly make a
particular combination in preference to all the other possibilities.

> Life replicates but not all that replicates is likely to be life (not
> sure of your meaning).  And I specifically said that I think that
> there could be chemical replicators.

I consider successful replication (fecundity greater than one, so it
grows exponentially to consume all available resources of the type it
needs) to be the essential thing that defines life, so I would consider
the first successful replicator to be the first just-barely-life, and
it seems much more likely that the first replicator evolved (by
modification and/or replacement) to yield present-day life than that
the first replicator had nothing much to do with eventual life. In
particular I think life as we know it would have been a lot more
difficult without that first replicator leading the way by generating
large quantities of specific very-complicted chemicals in place of the
mostly very-simple and somewhat-simple molecules that existed before
the replicator started working.

> > Or you don't belive that such a simple chemical replictor,
> > once formed and exponetially grown to have so many copies that it
> > persists indefinitely,
> Yes, this is a big potential stumbling block, IMO.  Especially if we
> can't make replicators (and I don't think that's yet been done).

I haven't heard that anybody has devised a self-replicating molecule
that works in a typical Miller-Urey environment, so I agree it's still
a speculation whether it can be done at all. Some agency with lots of
money needs to fund biochemical research to explore (via computer
simulation and/or micro-fluidics chemical apparatus) all the possible
chemical reactions between basic chemicals (water, carbon dioxide,
methane, ammonia, formaldehyde, cyanide, hydrogen sulfide, inorganic
metal and non-metal ions from dissolved salts, etc.) and chemical
fragments of such chemicals formed from high-energy events (UV,
electric sparks, high temperatures), to build up a large database of
everything that gets made that we can figure out the path to, starting
with whatever appears in largest yield and working down to rarer and
rarer products. As this database is built up, we need to perform
statistical analysis of it to find out how often any chemical that is
produced has a catalytic effect in helping produce other cheimcals.
Then analyse the rate at which one catalyst produces some other
catalyst. Once we have numbers for all these statistics, we can
calculate how often we would expect catalytic chains of various lengths
to spontaneously appear in the early Earth's oceans. We also need to
get statistics as to the number of different types of catalysts, that
is the number of specific kinds of chemical reactions being catalyzed.
Putting those together mathematically we should be able to calculate
the expected time before the pidgenhole principle forces a chain of
catalysts to "re-invent the wheel" i.e. form some product that has
already been listed in the catalog of catalysts.

Let me tighten up my math a little now: In building the catalog of
species of molecules that can be created spontaneously by these kinds
of chemical reactions, we would include *any* molecular species that is
likely to appear even once in a hundred million years. Any such
molecular species need only appear once to trigger a successful
catalytic bloom.

But in counting lengths of catalytic chains and consequent sizes of
catalytic blooms, we stop counting as soon as the one-way fecundity
from the start of the chain/bloom to the particular end-product drops
below 1. So we count chains/blooms only so far as one molecule of the
starting species is likely to produce one or more of the product
species. So in counting chains/blooms only in this way, we have a
precise definition of a chain or bloom, so we can then calculate the
statistical distribution of sizes of such. Given the rate at which any
catalyst spontaneously appears with one-step fecundity at least one,
i.e. n molecules of catalysts produce more than n molecules of product
on the average, and given the statistics that tell, relative to that
rate, how often two-step, three-step, etc. chains (of end-to-end
fecundity greater than one) are formed, we do simple arithmetic to
determine how often in real-time chains of any given length are formed.
For example, if one-step chains form a billion times per second in the
Ocean, and n+1 step chains form only one hundredth as often as n step
chains, then that means two-step chains form ten million times per
second, three-step chains form a hundred thousand times per second,
four-step chains form a thousand times per second, etc.

Now in building our catalog of *all* chemicals formed, and scanning for
any catalysts that have even one step fecundity greater than one, we
classify such catalysts according to what kind of reaction they
catalyze. We build up a side list of such kinds of reaction. If we
discover there are only twenty different kind of catalyzed reaction,
and we compute from the statistics of catalytic-chain lengths that a
20-step chain (21 catalysts counting both starting and ending point)
would form about once every ten million years, then we can conclude
that such a 21-catalyst chain would, by the pidgenhole principle,
include two catalysts that are in the same class, thereby closing the
catalyst-type loop, about once every ten million years. At that point
we'd almost have proven this method of abiogenesis (as something that
definitely would have happened if no other way of making a replacator
already happened faster). We'd then need to study what kinds of
catalysts usually make the longest chains, and try to actually find a
20-step chain. At that point we'd see which two catalysts among the 21
are of the same type, i.e. catalyze the same type of reaction, comare
them, and see whether starting from the second one we trace the same
path through types of catalyst as happened from the first to the
second, to see whether the loop repeats. I.e. suppose we find:
C1a -> C2a -> C3a -> ... -> C19a -> C20a -> C3b
where C3a and C3b are the repeats (among 21 catalysts of 20 different
types). Now running from C3b we expect a chain like this:
C3b -> C4b -> C5b -> ... -> C19b -> C20b -> C3c
In some cases, that chain has such low fecundity that it's no good and
we have to start from scratch with another 20-link chain that might
appear another 10 millions later. In other cases, that chain might
actually converge with the original chain, for example instead of
each catalyst in the second chain being different from the first,
we might have:
C2b -> C4b -> ... -> C7b -> C8a -> C9a -> ... -> C20a -> C3b
at which we have an actual catalytic cycle with fecundity greater than
one already, a replicator! More expectedly, we'd see each chain
slightly different from the previous, for several times around the
loop, where the type of catalyst repeats but the specific catalyst is
slightly different each time around. But given only a few chemicals as
input that the catalysts can act on, I'd expect the loop to start
repeating some exact catalyst from earlier, pretty soon, after just two
or three times around the catalytic-type cycle. It could converge to a
single loop, or it could jump back to some earlier cycle and thus have
a catalytic-species loop that is two or three times as long as the
catalytic-type loop.

The nice thing about approaching the research in this way, cataloging
the chemicals and then computing statistics of the catalog, is that we
can get a mathematical model of the statistics, showing how often
chains of each length occur, with a nice scaling law to extrapolate the
statistics beyond the current catalog, and thereby be able to calculate
a good estimate of how often a catalytic-type loop would be expected to
occur in the early Earth's oceans, and thus be able to say whether this
would almost surely happen quickly in our history, or wouldn't happen
in all the time before the Universe dies hence probably didn't happen
on Earth or anywhere else in our galaxy so probably doesn't explain
life on Earth, or is likely to happen every fifty billion years or so
in each planet's ocean in which case we might have gotten lucky, or
more likely some random planet out there in our galaxy got lucky and
panspermed every other planet in the galaxy.

> if I had to make a bet at this point, it'd be for panspermia (extra
> solar system).  That is, if Mars is involved, it also was seeded from
> outside.  I recognise that this only puts the OOL problem back to some
> other planet, but I think life is pretty improbable and likely quite
> rare if it has to emerge independently on "each" planet (where it
> exists).

That argument all depends on how commonly abiogenesis occurs. If
abiogenesis occurs, on the average, only once in fifty billion years
per planet, most planets would get panspermed long before they'd die
out, and most of the rest wouldn't abiogen in their entire time before
their star dies out. On other hand if abiogenesis occurs about every
ten million years on each planet, by the time a pansperm comes along
it'd find life already going quite nicely and ready to destroy the
pansperm the moment it arrived. If the Miller-Urey experiment worked
because a replicator actually was formed in just a few hours time in
such a small apparatus, and we just didn't happen to find it because we
had no idea to even look for it (by the multi-generation infection
experiment I proposed recently), it may be that in highly reducing
regions close to a few volcanic vents abiogenesis is actually happening
all the time even today, but before it can fill the ocean, any molecule
of replicator that ventures too far from the vent will encounter oxygen
in the water which destroys it, and any that remain close to the vent
encounter tubeworms ready to gobble it up, so regardless of which way
it drifts in the current it never gets a chance to show us that it
exists.

> But I will say that your agruments (and Tom's and Richard Dawkins')
> are having an affect and I'm far more ready to accept your scenario
> than I was a month ago.

Hmm, I know Dawkins wrote a lot about evolution and selfish gene and
apparent altruism explained by replication of copies of genes rather
than individuals, but I haven't seen anything he wrote specifically
about chemical-replicator abiogenesis. Do you have a summary of what he
said, or a relevant quote, or an online citation? In any case, I'm
somewhat flattered that you compare me with Dawkins in regard to
ability to convince you of the plausability of rather novel ideas
related to evolution.
---
ū RIMEGate(tm)/RGXPost V1.14 at BBSWORLD * Info@bbsworld.com

---
 * RIMEGate(tm)V10.2á˙* RelayNet(tm) NNTP Gateway * MoonDog BBS
 * RgateImp.MoonDog.BBS at 11/7/04 9:59:05 PM
 * Origin: MoonDog BBS, Brooklyn,NY, 718 692-2498, 1:278/230 (1:278/230)