CCL:QXP program for drug design (fwd)

[Eugene Leitl]

Anyone knows how QXP and similiar packages tax the messaging interconnect?

---------- Forwarded message ----------
Date: Wed, 20 Feb 2002 06:16:32 -0500
From: "STOUTEN, PIETER [R&D/0467]" <pieter.stouten at pharmacia.com>
To: 'Oliver Hucke' <Oliver.Hucke at physchem.uni-freiburg.de>
Cc: 'Computational Chemistry List' <chemistry at ccl.net>
Subject: CCL:QXP program for drug design

Hello Oliver,

>Do you have any experience with QXP,
>
Yes. In fact, we think it is so good that we recently concluded a global
agreement so that our research sites in St. Louis, Chicago, Kalamazoo (USA)
and Nerviano (Italy) have unlimited access and can run an unlimited number
of simultaneous QXP jobs. We are in fact in the process of purchasing two
Linux clusters with the specific aim to run QXP on them.

>especially with its ability to
>perform ligand docking calculations with flexible protein binding site
>and with the accuracy of predicted binding free energies?
>
Well, that's the holy grail, isn't it? In order to obtain correlation with
experiment, one often has to scale the various energy terms and even then
the results are not always satisfactory. Colin McMartin is about ready to
release a new scoring function that is based both on crystal structure data
and on inhibition data of congeneric sets of compounds. In terms of docking
quality (i.e., the ability to reproduce known binding modes), we find that
often QXP finds the correct mode (among, say, the best 10-20 poses), but
does not score it highest. Both docking and scoring quality also depend to a
large extent on the target under consideration. Another docking program you
might consider is MolSoft's ICM, whose scoring function has been optimized
such that "actives" and "inactives" are separated as much as possible.

>Could you point us to some papers describing QXP applications or other
>sources of information about the program?
>
I saw two back-to-back presentations at the 219th ACS meeting in San
Francisco (March 2000), one by Dominic Ryan of SKB and one by Venkatraman
Mohan of Isis Pharma. They both prominently featured application of QXP and
both speakers were quite positive about their experiences.

In terms of ease of use QXP/Flo surpasses anything I have seen. It is
clearly designed and developed for modellers by a modeller (who got
frustrated with generally available tools and decided to build his own).

Hope this helps,

Pieter

Pieter F.W. Stouten, Ph.D.
Head, Computational Sciences
Department of Chemistry
Pharmacia Italia, Nerviano (Mi), Italy
Phone:  +39-02-4838-5227 (secretary: 3962)
Fax:    +39-02-4838-3965
Mobile: +39-348-260-3380


-= This is automatically added to each message by mailing script =-
CHEMISTRY at ccl.net -- To Everybody  | CHEMISTRY-REQUEST at ccl.net -- To Admins
MAILSERV at ccl.net -- HELP CHEMISTRY or HELP SEARCH
CHEMISTRY-SEARCH at ccl.net -- archive search    |    Gopher: gopher.ccl.net 70
Ftp: ftp.ccl.net  |  WWW: http://www.ccl.net/chemistry/   | Jan: jkl at osc.edu