Perhaps Lachlan should have tried reading our Chem. Comm. before
sounding off about it. Introducing stupid terminology ('occult
methods'), albeit as a joke, doesn't help. Our Chem Comm
describes a SIMULATED ANNEALING approach. This was stated in
my posting. There was no mention of GA / AI, as neither was used.
As for this comment
> another obscure occult method of little relevance and little use
Do you consider three previously unknown complex structures to be
of no relevance ? Oddly enough, the referees disagreed with you.
As for availability of software, CCSL has been available free, as
source code, for many years. The structure solution software is
new, constantly evolving and not yet ready for distribution. Does
that mean that we are barred from publishing results derived from
it, or disclosing some key ideas that consitute the "solid science"
that Lachlan talks of ?
As for Armel's comments on Nxyz vs. Nhkl, it's a case of "horses
for courses". If we don't believe that a benzene ring consists
of C-C bonds of 1.4A with internal angles of 120 degrees then we
are indeed in trouble by building that information into our adjustable
model. However, we do believe it and that allows us to reduce the
Nxyz problem to an Nparameter problem (internal + external degrees
of freedom). It works, just look at the literature. The examples
in chem comm are highly flexible, which places them in a different league
of complexity from the rigid molecule (camphor) and one-torsion
(fluorescein) problems Armel mentions. Note that I am *not* doing down
those structures or the methods employed to solve them, as I know
both data sets well, know how they were solved and who solved them !!
I'm simply pointing out (and I'm sure the authors of both the camphor
and fluorscein papers would agree) that the internal flexibility
of the examples shown in chem comm makes them more of a challenge
to solve by model building methods.
Perhaps one should question the validity of the entire body of
protein structural work if one does not believe in the concept
of decreased degrees of freedom......
For those of you who don't have access to Chem Comm, here are
the details of the compounds solved
Promazine HCl : (C17 H21 N2 S1)+ (Cl)- P21/c Deg. of freedom=13
Thiothixene : C23 H29 N3 O2 S2 P21 Deg. of freedom=9
Capsaicin : C18 H27 N1 O3 P21/c Deg. of freedom=16
They were solved from 271,216 and 379 reflections respectively. I'll
leave it up to you to figure out Nxyz for each molecule. BTW, all
H atoms were included in the calculations.
Finally, I know from the previous exchanges on this list that I have
no chance of winning over the protagonists, so in the absence of any
more "mis-information" postings, there's nothing left to say. I'm off to
burn some incense, dance over the HRPD sample tank and wait for the
next eclipse, at which point I will solve my next structure.
K
----------------------------------
Dr. Kenneth Shankland,ISIS Facility,Rutherford Appleton Laboratory
Chilton,Didcot,Oxon OX11 0QX,United Kingdom
Tel : +44 1235 44 6381
Fax : +44 1235 44 5720
E-Mail: k.shankland@rl.ac.uk
Date: 22-Apr-98
Time: 14:25:01
----------------------------------
-- Dr. Kenneth Shankland,Building R3,ISIS Facility,Rutherford Appleton LaboratoryChilton,Didcot,Oxon OX11 0QX,United KingdomTel : +44 1235 44 6381 Fax : +44 1235 44 5720Email : k.shankland@rl.ac.uk