This page contains instructions for running the central programs used in the article Subnetwork hierarchies of biochemical pathways by Petter Holme, Mikael Huss, and Hawoong Jeong. These programs are developed under Debian GNU/Linux (but should be easy to adapt to other environments). Short descriptions of the three programs:
hi - This is the main program that deconstructs the biochemical network of an organism. It can either be run interactively (writing commands at a prompt or with a input file, in 'batchmode'). The syntax is the same for both cases: 'cell 1' means that the whole-cell network should be calculated for organism no. 1 (of the file 'orgnames'). 'meta 2' means that the metabolic network for organism no. 2 should be calculated. '-' starts the program. This input file should be located in a directory 'inp/' in the same directory as the executable. If the program is started by 'hi x' on the command line it reads the input file 'inp/x'. The output of the program are binary files in the directory 'data/' containing a matrix where the rows are the number of iterations of the algorithm, and columns are the cluster-index of a specific substance-node.A gzipped tar-file with the source code, makefile, and example input files can be obtained here (21kbytes). To get started type as follows:draw - This program draws the hierarchy trees in the Xfig format (input is a binary file produced by 'hi').
p2p - This program prints out a subnetwork in the Pajek format. (For explicit drawing and further analysis.)
tar -xzf hi.tgz (expands into a new directory 'hi/')Network-files for the other WIT organisms can be obtained here. Complete substance names can be obtained from WIT.
cd hi
make (you may have to edit the makefile first)
hi ex_hi (runs an example of hi, it takes ~10 secs on a 700MHz PIII)
ls data/* | draw (draws a hierarchy trees from hi's data-file)
p2p ex_p2p_meta (makes a Pajek file for a subnetwork in the output from 'draw')
p2p ex_p2p_cell (another example file for p2p)
Petter
Holme
April, 2002