# This package was developed by Adam Sadovsky and Gal Chechik
# (C) 2005-2008
# Can be used for academic purposes. 
# 
# For other uses and licencing please contact gal@ai.stanford.edu

Installation: 

	tar -xvzf findmotifs.tgz
	make clean all


To run:

	findmotifs.pl --help 


Implementation details
----------------------
A motif = common ancestor (dist <= A anc threshold)
          no common descendant (up to dist <= A desc threshold)
V motif = common descendant (dist <= V desc threshold)
          no common ancestor (up to dist <= V anc threshold)
P motif = common descendant (dist <= P desc threshold)
          and common ancestor (dist <= P anc threshold)
C motif = cycle of length <= C threshold

* motifs are relations, so order matters. e.g. if A-B is a V12, 
  B-A is a V21
* when searching for a common descendant/ancestor, we only find
  the common desc/anc of least distance.. Consequence: if two genes
  are in both a V11 and a V12, only the V11 will be listed.
* since genes can catalyze multiple reactions, it is possible for
  a gene to be in a motif with itself (these aren't filtered out)
* Weak a motif is a superset of a-and, and weak v motif is a 
  superset of v-and (b/c any genes in a v-and/a-and share a common
  descendant/ancestor)
* P motifs contain all T motifs (b/c any genes in T motif have a
  common ancestor and a common descendant)
* Output with _cmpd contains the nodes (compounds) between each 
  genes. This is currently implemented only for T-chains and for 
  V/A motifs
* Currently, v-and and a-and (and v-nec and a-nec) only work for
  depth=1
* Cycles (C motifs) work for weak and strong, but for weak, there
  will be repeat cycles of the form A-B-C, B-C-A, C-A-B
* compound output seems to work but hasn't really been tested yet
* for t-chains, we prevent a chain from passing through the same 
  node or gene more than once. In weak mode, this means not using 
  any of the first gene's start-nodes again in the chain (and not
  reusing any other nodes). It also means that none of the last 
  gene's end-nodes can have been used before. (Note: this is a 
  complicated definition in weak mode) 
* if the input file is produced in windows, run dos2unix to 
  convert the endline chars