More information is available at http://faculty.cs.tamu.edu/shsze/graphmatch. INSTALLATION 1. Type ./install to install graphmatch. 2. Either move the executable file graphmatch to a directory on the search path or add the current directory to the search path. INPUT The following files are needed: 1. A file that specifies the query graph. The line v1 v2 v3 v4 defines edges (v1,v2), (v1,v3) and (v1,v4). An undirected edge (v1,v2) should be included twice, once in v1 ... v2 ... and once in v2 ... v1 ... Example: Hsc82p Bud6p Ste11p Bud6p Hsc82p Ste11p Ste11p Hsc82p Bud6p More examples are in ProteinInteractionNetwork/query and MetabolicNetwork/query. 2. A file that specifies the input graph in which related graphs are sought. The line v1 v2 v3 v4 defines edges (v1,v2), (v1,v3) and (v1,v4). An undirected edge (v1,v2) should be included twice, once in v1 ... v2 ... and once in v2 ... v1 ... Example: ifc-2 dyb-1 mig-15 ttx-1 dpy-14 dyb-1 ifc-2 dyp-14 mig-15 ttx-1 mig-15 A few pre-computed graphs from DIP (http://dip.doe-mbi.ucla.edu) and EcoCyc (http://ecocyc.org) representing are in ProteinInteractionNetwork/input and MetabolicNetwork/input. 3. A file specifying correspondences between vertices in the query graph and the input graph. Each correspondence is specified as v1 v2 corr where v1 is a vertex in the query graph, v2 is a vertex in the input graph and corr is the similarity score. Example: Hsc82p ifc-2 3e-04 Bud6p ttx-1 2e-03 More examples are in ProteinInteractionNetwork/corr and MetabolicNetwork/corr. USAGE graphmatch -q=query -i=input -c=corr -g=1 -l=1 -p=0.2 -v=7.0 -n=10 -o=result Command line parameters: -q= "file name containing query graph" -i= "file name containing input graph" -c= "file name containing correspondences between vertices" -g= "maximum number of indels between two matches that are adjacent in the query graph" -l= "type of match score" 0 -- use similarity score as match score 1 -- use negative logarithm of similarity score as match score -p= "indel penalty" -v= "non-associated vertex penalty" -n= "number of output graphs" -o= "output file name" The running time of graphmatch is sensitive to parameter settings. OUTPUT Each result is shown as a graph alignment in which each pair of matching vertices is shown on a line. For each adjacent match, corresponding edges are also shown on a line, with an edge that is in the query graph on the left and a path that is in the input graph on the right which can include indels between the matching vertices. It is possible that some results are repeated since a graph alignment may be represented more than once. Example: Result01: score=1511.95 matches: Hsc82p daf-21 Ste11p mig-15 Spa2p zyg-1 Mkk2p sek-1 Mkk1p sek-1 Fus3p mpk-1 Dig2p vab-3 Dig1p GNB:17543358 Kss1p mpk-1 Ste7p mig-15 Mpt5p PIR:T15820 edges: Hsc82p -- Ste11p daf-21 -- uaf-1 -- mig-15 Ste11p -- Spa2p mig-15 -- pal-1 -- zyg-1 Ste11p -- Fus3p mig-15 -- csn-5 -- mpk-1 Ste11p -- Kss1p mig-15 -- csn-5 -- mpk-1 Spa2p -- Mkk2p zyg-1 -- pal-1 -- sek-1 Spa2p -- Mkk1p zyg-1 -- pal-1 -- sek-1 Spa2p -- Ste7p zyg-1 -- pal-1 -- mig-15 Fus3p -- Dig2p mpk-1 -- PIR:T25092 -- vab-3 Fus3p -- Dig1p mpk-1 -- GNB:17543358 Fus3p -- Ste7p mpk-1 -- csn-5 -- mig-15 Fus3p -- Mpt5p mpk-1 -- GNB:32566959 -- PIR:T15820 Dig2p -- Dig1p vab-3 -- PIR:T15820 -- GNB:17543358 Dig2p -- Kss1p vab-3 -- PIR:T25092 -- mpk-1 Dig1p -- Kss1p GNB:17543358 -- mpk-1 Kss1p -- Ste7p mpk-1 -- csn-5 -- mig-15 Kss1p -- Mpt5p mpk-1 -- GNB:32566959 -- PIR:T15820 More examples are in ProteinInteractionNetwork/result and MetabolicNetwork/result.