KRISTOFFERSON@BIONET-20.ARPA (04/12/88)
From: <BAIROCH%CGECMU51.BITNET@CUNYVM.CUNY.EDU>
Brookhaven on-line on BIONET: a dissenting view.
It seems that most reply to the decision of BIONET to , at least temporarily,
not to provide the PDB data bank, on-line, has been negative. I must admit
I have a dissenting opinion.
The use of the PDB data bank data by molecular graphics software is generally
dependant on work stations with interactive, high quality, graphic display. By
the nature of the hardware on which BIONET is running I do not see how one
could run software package(s) to take advantage of that data bank on-line. Of
course one may answer me that having them on-line allows users to download the
structure(s) for which they are interested, but here the problem is the size
of an entry: a PDB entry with its atomic coordinates is generally as large
as 100 Kb and downloading those entries is going to take a long time. One
must also take in account that the majority of users which have molecular
graphic software also subscribe to get copy the PDB tapes (you generally do
not want to start downloading files 20 hours a week when you spent a relatively
high number of K dollars to buy hardware and software, especially when the PDB
tapes are not so expensive).
In addition one of the few use of the PDB data bank to users with no molecular
modelling software is to run programs to obtain secondary structure parameters
from the tertiary structure and to use this information for secondary structure
prediction of homologous proteins. In this case there is a solution that BIONET
can implement: Chris Sander at the EMBL maintain a data bank of such secondary
structure data as derived from the PDB data bank: the DSSP data bank [1], I
believe that it would not be too difficult to get the permission to have that
bank on-line and it does not take so much disk space.
Also, to give hope to the under-equipped researcher, which like myself , have
only access to PCs: Oxford University Press with the help of IBM-England and
the permission of Brookhaven is going to release a CD-ROM package which
include the PDB data bank and some software to access and display molecular
structures on the PS/2 family of systems [2].
Amos Bairoch
Dept. Med. Biochemistry
University of Geneva
----------------------------------------
[1] Kabsch W., Sander C.
Dictionary of protein secondary structure: pattern recognition of
hydrogen-bonded and geometrical features."
Biopolymers 22:2577-2637(1983).
[2] For information contact:
Mr. A. Hodgkin
Oxford Electronic Publishing
Oxford University Press
Walton Street
Oxford OX2 6DPKRISTOFFERSON@BIONET-20.ARPA (04/12/88)
From: David Kristofferson <Kristofferson@BIONET-20.ARPA> Amos is correct in pointing out that BIONET does not have the capabilities currently to utilize the on-line data in molecular modeling software. While we do not foresee this changing in the near future, the advent of high speed network connections may eventually change this situation. This would require the end users to have direct ARPANET/NSFNET access to BIONET and also require that these networks be enhanced to provide far better performance than they currently do. I do not think that all of these conditions are going to be met in the near term. I was not aware of the DSSP databank from Chris Sander and am grateful for this information. DSSP was one of the programs that we had for on-line use with Brookhaven, the other being ANNOT from Kanehisa's IDEAS package. We have considered simply putting the output of these programs on-line instead of the raw data, but this new information provides another possible path. The PS/2 software that AMOS describes sounds very interesting. My hunch is that this will not be an inexpensive package however and may require the higher end PS/2 models, yes? Sincerely, Dave Kristofferson BIONET Resource Manager kristofferson@bionet-20.arpa -------