From: Robert J. Bradbury (bradbury@aeiveos.com)
Date: Wed Aug 21 2002 - 11:51:04 MDT
On Mon, 19 Aug 2002, Anders Sandberg wrote:
> It will still be many years before we have a catalogue sizeable enough.
> It will happen, but it will take time.
Agreed, but starting from highly conserved regions you can get a toehold
on the genome and start sequencing from there. Shotgun sequencing doesn't
even require that if you can culture the culprit.
> Yes, but the sheer variety is staggering. They seem to be coevolving new
> toxins very quickly.
I saw a reference that seemed to suggest that conotoxins were evolving
through an antibody like diversity generation mechanism. Clever or nasty
depending on how you look at it.
> How good are the retrosynthesis programs today?
I think the academic versions can do up to 500 atoms and people are
working on a commercial version to go to 1000 atoms. Thats large
enough for most drug molecules.
> Isn't having the crystal structure a bit optimistic? What if the toxin is
> one of those membrane-spanners where the active form is very different
> from the crystaline form?
Thats why one has to use both crystal structures and molecular modeling.
The advantage we have over terrorists is that we have *much* greater
capacity to do the both the hardware and software molecular analysis.
They would have to be either very clever or very lucky to come up
with something very different from what is already known that we
couldn't crack quickly.
> There are a lot of
> human delays inherent in any such project - discovering that there is an
> attack might take a while, getting the signal out to mobilize the
> emergency response and organising the shift in activity in hundreds or
> thousands of independent research facilities.
True. But we have lived in an era where people practiced "emergency drills"
before.
> Even with perfect tech
> there are serious logistic problems here (central decision-making nodes
> must avoid being swamped with information, error checking is necessary
> etc).
Design it so it can naturally be run in a distributed fashion (use the
"cell" concept against the enemy).
> I think much of this could be solved, but it would require a
> serious political effort that managed to get support from the involved
> research institutions.
Granted.
Robert
This archive was generated by hypermail 2.1.5 : Sat Nov 02 2002 - 09:16:19 MST