[RC5] Gene sequencing next d.net project?

Essam Metwally emetwall at ucalgary.ca
Mon Dec 6 09:47:27 EST 1999

> > I myself have been toying with the idea of 3D Protein Structure
> > analysis.
> I'd love to help on both - but you need to define your problem better,
> too. What exactly did *you* have in mind?

Well, I work with small peptide sequences.  I am looking for the
receptor for these sequences.  In the interim we have done quite a bit
of 3D structure analysis in vacuo.  While in vacuo makes for simplified
analysis, its not the realistic scenario (in vacuo = vacuum) where
peptides are more often found in association with varying amounts of
water molecules.  Ideally you would study structure in a "box of water"
ie in an excess amount of water.

However, to start in vacuo would be easiest.  The problem comes down to
this.  What would a peptide's structure look like given an absence of
outside influence.  Or in other words, what would the internal bond
angles, dipole moments, and attractions between the various atoms
present in a peptide yield as a final structure.  So you have a
multiplicity of problems.  The basics have already been defined to some
extent, each of the actual bonds have known or at least hypothesized
structures and characteristics.  This would form the basis of any
analysis.  But how do the interactions between atoms which are nearby in
a spatial arrangement affect the orientations of other atoms.  This is
all to be taken given the fact that all of these atoms are constrained
in there 3D arrangement by the underlying peptide backbone.

I know this is just a skeletal analysis but if you truly consider the
problem, its possible to see that there are a multiplicity of
arrangements possible given just minor shifts of single atoms.  By
increasing the number of atoms, the complexity of the problem likewise
increases exponentially.

> What are we analyzing?

Ultimate output would be the location of all associated atoms in 3D

> What kind of calculations are involved in analyzation?

Typically this is done through the use of Gibbs free energy.  Given all
of the final positions, this is the energy which is left.  By minimizing
this number you in effect have a more stable molecule.  Hence it is
thought that by understanding what the most stable conformation is.

> What kind of computers are used now to perform this now?

Any computer can be used, there are various software packages available
which do this type of work.  All are commercial.  I think there are a
couple which support parallel distribution.  Now I can see how to do it
given a continuously connected network, but how would you do it when
machine only plug into the net every few days?

> What kind of programs are used now to perform this now?
> What structure do these programs have?

The majority are single process programs.  Essentially they rotate,
shorten, lengthen and do other 3D transformations and recalculate the
total free energy at every step.  You can restrict the locus of rotation
or you can give the computer free reign.  Typically the end result is a
more stable conformation.  You can then plug this conformation in and do
some tweaking and find a more stable confromation.  Typically this goes
on until nothing  more stable is found.

What would you propose?


To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest

More information about the rc5 mailing list