Search-Scripts
There exist the following scripts located in the Moloc
bin-directory at the Basel site. (obtain the directory path by issuing the
command 'which mx'). They access the following databases:
- rosrch: Rosis available structures, partitioned on 8 files.
- nosrch: Rosis NOT available structures, partitioned on 8 files.
- acdsrch: ACD available chemicals, partitioned on 6 files.
- mdcmsrch: DAYLIGHT's medchem database, partitioned on 4 files.
Each of these databases exist in two different levels of coarseness.
The resolution parameter of the search scrips selects in the following way:
- 0 (low-resolution): selects the database that has been
generated with the command 'Mtprgn -lP7 -b3.5 -d10 -h1.3 -a1.3' (start Mtprgn
without arguments to obtain the significance of the arguments).
- 1 (high-resolution): selects the database that has
been generated with the command 'Mtprgn -lP7 -b2 -d7 -h1.3 -a1.3'
For consistency it is advisable to use the corresponding parameter values
for generation of the .tpr file of the query structure (Moloc option php/v,w,
or Mtprgn).
Of course, the program Mtprgn can also be used to build up user specific
databases.
Modifying Scripts
It is possible to modify any of the scripts to adjust to special needs such as
modified maximum number of pharmacophors, or modifies parameters for the
similarity calculation. To obtain the meaning of the parameters
start Mtprsml without arguments.
Running Scripts
The scripts are designed to run in parallel on all partitions of a database
(8 for the case of Rosis). Thus, do NOT run these scripts on a
single processor machine! It will be hopelessly overloaded. In such
circumstances parallel running must be avoided by removing the &-signs from
the script.
The scripts return a full similarity file ranked with respect to a substructure
type search (second number, column 3), a top-400 .tdt file to interface
DAYLIGHT software, and a top-400 .mab file of 3-d structures.
Remarks
CPU-times for similarity calculations with topological pharmacophors grow
very quickly with increasing number of agons. The default value of 8 yields
reasonable times (few minutes). If the search pharmacophor is small, the
limit may be increased (leading to larger pharmacophors of the base to be
included). This makes usually only sense, if the substructure type surch is
of interest.
In general, with rather large lead structures, a more coarse grained
description must be chosen to achieve reasonable search times. On the other
hand, the fine grained description is suited to look for more detailed analogs
of smaller structures or fragments.