From d9484afdc19b9529a8121646de63e991a33e3656 Mon Sep 17 00:00:00 2001
From: Ali Gholami Rudi <ali@rudi.ir>
Date: Wed, 17 Oct 2012 19:36:04 +0330
Subject: [PATCH] README: document the new file format and mkrangi.sh

---
 README | 118 +++++++++++++++++++++++++++++++++++------------------------------
 1 file changed, 63 insertions(+), 55 deletions(-)
 rewrite README (72%)

diff --git a/README b/README
dissimilarity index 72%
index ea294b9..e500b4b 100644
--- a/README
+++ b/README
@@ -1,55 +1,63 @@
-RANGI
-=====
-
-RANGI finds colored motifs in list-colored graphs.  You'll need a POSIX
-environment to compile this program.  To compile it, simply run make.
-rangi is the main program:
-
-  $ make
-    # compiles rangi, conv, and genrq
-  $ ./rangi -h
-    # shows available options
-  $ ./rangi <test.dot
-    # finds motifs of the largest size in test.dot
-  $ ./rangi -s 10 <test.dot
-    # finds motifs of size 10 in the graph specified in test.dot
-  $ ./rangi -s 10+ <test.dot
-    # finds largest motifs in test.dot starting from size 10
-  $ ./rangi -s 10- <test.dot
-    # finds largest motifs in test.dot that have at most 10 vertices
-  $ ./rangi -s 10 -n 4 <test.dot
-    # finds motifs of size 10 using 4 threads
-
-Input graphs can be in two formats: either the graphviz graph format, with
-Betzler et al.'s extensions, like test0.dot, or our custom graph format,
-as explained in the comments of readgraph() in rangi.c, like test1.rangi.
-
-For generating a graph from the data sets included in
-http://www.cs.tau.ac.il/~bnet/torque.html, download the archives
-for any of the three mentioned species.  Then invoke:
-
-  $ ./conv -h
-    # shows available options
-  $ unzip DM_predef_torque.zip
-    # unzip any of the three datasets
-  $ ./conv -g -q query_interactions PROT1 PROT2 ... <input_network >graph.dot
-    # generate list-colored graph with the given queries in graphviz format
-
-The -q option of conv expects the query interaction file, which lists the
-BLAST score of each protein in the query with the proteins in the input
-network (we have obtained these from Betzler et al.).  After options,
-the name of the proteins in the query should be provided.
-
-The graphs we generated using conv for testing rangi can be obtained from:
-http://ce.sharif.edu/~gholamirudi/rangi.graphs.tar.gz
-
-The genrq program generates random queries from a network.  These queries
-can be passed to conv program to create an input graph with random
-queries from the same network.  Note that conv removes the vertices in
-the query from the network before generating the input graph, if queries
-are selected from the same network and no -q option is provided.
-
-  $ ./genrq -h
-    # shows available options
-  $ ./genrq -n 10 -s 11 <network
-    # generates 10 queries of size 11 from the given network
+RANGI
+=====
+
+RANGI finds colored motifs in list-colored graphs.  You'll need a POSIX
+environment to compile this program.
+
+To compile it, simply run make.  rangi is the main program:
+
+  $ make
+    # compiles rangi, conv, and genrq
+  $ ./rangi -h
+    # shows available options
+  $ ./rangi <test.dot
+    # finds motifs of the largest size in test.dot
+  $ ./rangi -s 10 <test.dot
+    # finds motifs of size 10 in the graph specified in test.dot
+  $ ./rangi -s 10+ <test.dot
+    # finds largest motifs in test.dot starting from size 10
+  $ ./rangi -s 10- <test.dot
+    # finds largest motifs in test.dot that have at most 10 vertices
+  $ ./rangi -s 10 -n 4 <test.dot
+    # finds motifs of size 10 using 4 threads
+
+Input graphs should be in RANGI graph format.
+
+RANGI INPUT FORMAT
+==================
+
+RANGI expects the following format for the input list-colored graph:
+
+* The first number in the file indicates the number of nodes.
+* Next, each of the nodes are described: the name of each node is
+  followed by its list of colors terminated with -1, all separated
+  with spaces.
+* The edges of the graph are described next: each edge is defined
+  as a triple of the name of its endpoints and its weight.  The
+  list of the edges are read until end of file is reached.
+
+For instance the input file containing:
+  4
+  v1 1 -1
+  v2 2 -1
+  v3 3 -1
+  v4 4 -1
+  v1 v2 100
+  v2 v3 200
+  v3 v4 100
+Describes a graph with 4 nodes (v1 with color 1, v2 with color 2, v3 with
+color 3 and v4 with color 4) with three edges of weight 100, 200 and 100.
+
+CREATING INPUT FILES
+====================
+
+The included mkrangi.sh script creates RANGI input files for
+protein interaction networks.  You'll need the blast program
+(http://www.ncbi.nih.gov/BLAST/); pre-compiled versions of blast are
+available in many fameous Linux distributions as blast2.
+
+The biological files (network interactions, FASTA and
+complexes files) can be obtained from GraMoFoNe's website
+(http://igm.univ-mlv.fr/AlgoB/gramofone/).  I've already created RANGI
+input files for the networks in that page using blast2-2.2.24, which
+can be obtained from: http://ce.sharif.edu/~gholamirudi/rangi.dat.tar.gz
-- 
2.11.4.GIT