From eb1a37142515018a96b0eda1cd11429eaa4dc728 Mon Sep 17 00:00:00 2001
From: cfrainay <clement.frainay@inrae.fr>
Date: Thu, 3 Feb 2022 18:12:59 +0100
Subject: [PATCH 1/7] [Toolbox] create subnetwork extraction from reaction
graph
---
.../ExtractSubReactionNetwork.java | 142 ++++++++++++++++++
1 file changed, 142 insertions(+)
create mode 100644 met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
diff --git a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
new file mode 100644
index 000000000..9faa1976c
--- /dev/null
+++ b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
@@ -0,0 +1,142 @@
+package fr.inrae.toulouse.metexplore.met4j_toolbox.networkAnalysis;
+
+import fr.inrae.toulouse.metexplore.met4j_core.biodata.BioMetabolite;
+import fr.inrae.toulouse.metexplore.met4j_core.biodata.BioReaction;
+import fr.inrae.toulouse.metexplore.met4j_core.biodata.BioNetwork;
+import fr.inrae.toulouse.metexplore.met4j_core.biodata.collection.BioCollection;
+import fr.inrae.toulouse.metexplore.met4j_graph.computation.connect.KShortestPath;
+import fr.inrae.toulouse.metexplore.met4j_graph.computation.connect.ShortestPath;
+import fr.inrae.toulouse.metexplore.met4j_graph.computation.connect.SteinerTreeApprox;
+import fr.inrae.toulouse.metexplore.met4j_graph.computation.connect.weighting.DefaultWeightPolicy;
+import fr.inrae.toulouse.metexplore.met4j_graph.computation.connect.weighting.DegreeWeightPolicy;
+import fr.inrae.toulouse.metexplore.met4j_graph.computation.connect.weighting.WeightsFromFile;
+import fr.inrae.toulouse.metexplore.met4j_graph.core.BioPath;
+import fr.inrae.toulouse.metexplore.met4j_graph.core.GraphFactory;
+import fr.inrae.toulouse.metexplore.met4j_graph.core.WeightingPolicy;
+import fr.inrae.toulouse.metexplore.met4j_graph.core.compound.CompoundGraph;
+import fr.inrae.toulouse.metexplore.met4j_graph.core.reaction.CompoundEdge;
+import fr.inrae.toulouse.metexplore.met4j_graph.core.reaction.ReactionGraph;
+import fr.inrae.toulouse.metexplore.met4j_graph.core.compound.ReactionEdge;
+import fr.inrae.toulouse.metexplore.met4j_graph.io.Bionetwork2BioGraph;
+import fr.inrae.toulouse.metexplore.met4j_graph.io.ExportGraph;
+import fr.inrae.toulouse.metexplore.met4j_graph.io.NodeMapping;
+import fr.inrae.toulouse.metexplore.met4j_io.jsbml.reader.JsbmlReader;
+import fr.inrae.toulouse.metexplore.met4j_io.jsbml.reader.Met4jSbmlReaderException;
+import fr.inrae.toulouse.metexplore.met4j_mapping.Mapper;
+import fr.inrae.toulouse.metexplore.met4j_toolbox.generic.AbstractMet4jApplication;
+import org.kohsuke.args4j.Option;
+
+import java.io.IOException;
+import java.util.HashSet;
+import java.util.List;
+
+public class ExtractSubReactionNetwork extends AbstractMet4jApplication {
+
+ @Option(name = "-i", usage = "input SBML file", required = true)
+ public String inputPath = null;
+ @Option(name = "-s", usage = "input sources txt file", required = true)
+ public String sourcePath = null;
+ @Option(name = "-t", usage = "input targets txt file", required = true)
+ public String targetPath = null;
+ @Option(name = "-o", usage = "output gml file", required = true)
+ public String outputPath = null;
+
+
+ @Option(name = "-sc", aliases = {"--side"}, usage = "an optional file containing list of side compounds to ignore")
+ public String sideCompoundFile = null;
+
+ @Option(name = "-cw", aliases = {"--customWeights"}, usage = "an optional file containing weights for compound pairs", forbids = {"-dw", "-sw"})
+ public String weightFile = null;
+
+ //@Option(name = "-u", aliases = {"--undirected"}, usage = "Ignore reaction direction")
+ //public Boolean undirected = false;
+
+ @Option(name = "-k", usage = "Extract k-shortest paths", forbids = {"-st"})
+ public int k = 1;
+ @Option(name = "-st", aliases = {"--steinertree"}, usage = "Extract Steiner Tree", forbids = {"-k"})
+ public boolean st = false;
+
+
+ public void run() throws IOException, Met4jSbmlReaderException {
+ //import network
+ JsbmlReader reader = new JsbmlReader(this.inputPath, false);
+ BioNetwork network = reader.read();
+
+ //Graph processing: import side compounds
+ System.err.println("importing side compounds...");
+ Mapper<BioMetabolite> mapper = new Mapper<>(network,BioNetwork::getMetabolitesView).skipIfNotFound();
+ BioCollection<BioMetabolite> sideCpds = mapper.map(sideCompoundFile);
+ if(mapper.getNumberOfSkippedEntries()>0) System.err.println(mapper.getNumberOfSkippedEntries() + " side compounds not found in network.");
+ System.err.println(sideCpds.size() + " side compounds ignored during graph build.");
+
+ //get sources and targets
+ System.err.println("extracting sources and targets");
+ Mapper<BioReaction> rmapper = new Mapper<>(network,BioNetwork::getReactionsView).skipIfNotFound();
+ HashSet<BioReaction> sources = new HashSet<>(rmapper.map(sourcePath));
+ if(rmapper.getNumberOfSkippedEntries()>0) System.err.println(rmapper.getNumberOfSkippedEntries() + " source not found in network.");
+ HashSet<BioReaction> targets = new HashSet<>(rmapper.map(targetPath));
+ if(rmapper.getNumberOfSkippedEntries()>0) System.err.println(rmapper.getNumberOfSkippedEntries() + " target not found in network.");
+
+ //Create compound graph
+ Bionetwork2BioGraph builder = new Bionetwork2BioGraph(network);
+ ReactionGraph graph = builder.getReactionGraph(sideCpds);
+
+ //Graph processing: set weights [optional]
+ WeightingPolicy<BioReaction, CompoundEdge, ReactionGraph> wp = new DefaultWeightPolicy<>();
+ if (weightFile != null) {
+ wp = new WeightsFromFile(weightFile, true);
+ }
+ wp.setWeight(graph);
+
+ //extract sub-network
+ GraphFactory<BioReaction, CompoundEdge, ReactionGraph> factory = new GraphFactory<>() {
+ @Override
+ public ReactionGraph createGraph() {
+ return new ReactionGraph();
+ }
+ };
+ ReactionGraph subnet;
+ if (st) {
+ SteinerTreeApprox<BioReaction, CompoundEdge, ReactionGraph> stComp = new SteinerTreeApprox<>(graph);
+ List<CompoundEdge> stEdges = stComp.getSteinerTreeList(sources, targets, (weightFile != null));
+ subnet = factory.createGraphFromEdgeList(stEdges);
+ } else if (k > 1) {
+ KShortestPath<BioReaction, CompoundEdge, ReactionGraph> kspComp = new KShortestPath<>(graph);
+ List<BioPath<BioReaction, CompoundEdge>> kspPath = kspComp.getKShortestPathsUnionList(sources, targets, k);
+ subnet = factory.createGraphFromPathList(kspPath);
+ } else {
+ ShortestPath<BioReaction, CompoundEdge, ReactionGraph> spComp = new ShortestPath<>(graph);
+ List<BioPath<BioReaction, CompoundEdge>> spPath = spComp.getShortestPathsUnionList(sources, targets);
+ subnet = factory.createGraphFromPathList(spPath);
+ }
+
+ //export sub-network
+ ExportGraph.toGmlWithAttributes(subnet, outputPath);
+
+ }
+
+ public static void main(String[] args) throws IOException, Met4jSbmlReaderException {
+ ExtractSubReactionNetwork app = new ExtractSubReactionNetwork();
+ app.parseArguments(args);
+ app.run();
+ }
+
+ @Override
+ public String getLabel() {
+ return this.getClass().getSimpleName();
+ }
+
+ @Override
+ public String getLongDescription() {
+ return this.getShortDescription() + "\n" +
+ "The subnetwork correspond to part of the network that connects reactions from the first list to reactions from the second list.\n" +
+ "Sources and targets list can have elements in common. The connecting part can be defined as the union of shortest or k-shortest paths between sources and targets, " +
+ "or the Steiner tree connecting them. Contrary to Compound graph, reaction graph often lacks weighting policy for edge relevance. In order to ensure appropriate " +
+ "network density, a list of side compounds to ignore for linking reactions must be provided";
+ }
+
+ @Override
+ public String getShortDescription() {
+ return "Create a subnetwork from a GSMN in SBML format, and two files containing lists of reactions of interests ids, one per row, plus one file of the same format containing side compounds ids.";
+ }
+}
--
GitLab
From 8055b9d1c6b9f09d6d56030b00592691d9e3ba6b Mon Sep 17 00:00:00 2001
From: cfrainay <clement.frainay@inrae.fr>
Date: Fri, 4 Feb 2022 11:43:41 +0100
Subject: [PATCH 2/7] fix cli ption
---
.../networkAnalysis/ExtractSubReactionNetwork.java | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
index 9faa1976c..46892348b 100644
--- a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
+++ b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
@@ -45,12 +45,9 @@ public class ExtractSubReactionNetwork extends AbstractMet4jApplication {
@Option(name = "-sc", aliases = {"--side"}, usage = "an optional file containing list of side compounds to ignore")
public String sideCompoundFile = null;
- @Option(name = "-cw", aliases = {"--customWeights"}, usage = "an optional file containing weights for compound pairs", forbids = {"-dw", "-sw"})
+ @Option(name = "-cw", aliases = {"--customWeights"}, usage = "an optional file containing weights for reactions pairs")
public String weightFile = null;
- //@Option(name = "-u", aliases = {"--undirected"}, usage = "Ignore reaction direction")
- //public Boolean undirected = false;
-
@Option(name = "-k", usage = "Extract k-shortest paths", forbids = {"-st"})
public int k = 1;
@Option(name = "-st", aliases = {"--steinertree"}, usage = "Extract Steiner Tree", forbids = {"-k"})
--
GitLab
From 5ed81cc165c81ccb2e3c3f06c0afd41814245960 Mon Sep 17 00:00:00 2001
From: Ludovic Cottret <ludovic.cottret@inra.fr>
Date: Tue, 8 Feb 2022 09:02:56 +0100
Subject: [PATCH 3/7] grammar/spelling/typo fix
---
.../networkAnalysis/ExtractSubReactionNetwork.java | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
index 46892348b..8325de502 100644
--- a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
+++ b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
@@ -126,7 +126,7 @@ public class ExtractSubReactionNetwork extends AbstractMet4jApplication {
@Override
public String getLongDescription() {
return this.getShortDescription() + "\n" +
- "The subnetwork correspond to part of the network that connects reactions from the first list to reactions from the second list.\n" +
+ "The subnetwork corresponds to part of the network that connects reactions from the first list to reactions from the second list.\n" +
"Sources and targets list can have elements in common. The connecting part can be defined as the union of shortest or k-shortest paths between sources and targets, " +
"or the Steiner tree connecting them. Contrary to Compound graph, reaction graph often lacks weighting policy for edge relevance. In order to ensure appropriate " +
"network density, a list of side compounds to ignore for linking reactions must be provided";
--
GitLab
From 50b421abebc63ff4776c469d2ab1100a7eb7ef8a Mon Sep 17 00:00:00 2001
From: Ludovic Cottret <ludovic.cottret@inra.fr>
Date: Tue, 8 Feb 2022 09:03:20 +0100
Subject: [PATCH 4/7] grammar/spelling/typo fix
---
.../networkAnalysis/ExtractSubReactionNetwork.java | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
index 8325de502..0fa763913 100644
--- a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
+++ b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
@@ -128,7 +128,7 @@ public class ExtractSubReactionNetwork extends AbstractMet4jApplication {
return this.getShortDescription() + "\n" +
"The subnetwork corresponds to part of the network that connects reactions from the first list to reactions from the second list.\n" +
"Sources and targets list can have elements in common. The connecting part can be defined as the union of shortest or k-shortest paths between sources and targets, " +
- "or the Steiner tree connecting them. Contrary to Compound graph, reaction graph often lacks weighting policy for edge relevance. In order to ensure appropriate " +
+ "or the Steiner tree connecting them. Contrary to compound graph, reaction graph often lacks weighting policy for edge relevance. In order to ensure appropriate " +
"network density, a list of side compounds to ignore for linking reactions must be provided";
}
--
GitLab
From c79859d2ea0357301a9346ad023b583d5b36de92 Mon Sep 17 00:00:00 2001
From: cfrainay <clement.frainay@inrae.fr>
Date: Mon, 21 Feb 2022 16:37:51 +0100
Subject: [PATCH 5/7] fix typo
---
.../networkAnalysis/ExtractSubReactionNetwork.java | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
index 0fa763913..c6f7b4840 100644
--- a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
+++ b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
@@ -74,7 +74,7 @@ public class ExtractSubReactionNetwork extends AbstractMet4jApplication {
HashSet<BioReaction> targets = new HashSet<>(rmapper.map(targetPath));
if(rmapper.getNumberOfSkippedEntries()>0) System.err.println(rmapper.getNumberOfSkippedEntries() + " target not found in network.");
- //Create compound graph
+ //Create reaction graph
Bionetwork2BioGraph builder = new Bionetwork2BioGraph(network);
ReactionGraph graph = builder.getReactionGraph(sideCpds);
--
GitLab
From 15e4f7bf5a2f2fa59b710e65d140cce37bab1af9 Mon Sep 17 00:00:00 2001
From: cfrainay <clement.frainay@inrae.fr>
Date: Tue, 22 Feb 2022 13:59:33 +0100
Subject: [PATCH 6/7] makes side compound definition mendatory
avoid latency during reaction graph build and traversal due to over-density
---
.../networkAnalysis/ExtractSubReactionNetwork.java | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
index c6f7b4840..d5e04ede5 100644
--- a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
+++ b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
@@ -40,10 +40,9 @@ public class ExtractSubReactionNetwork extends AbstractMet4jApplication {
public String targetPath = null;
@Option(name = "-o", usage = "output gml file", required = true)
public String outputPath = null;
-
-
- @Option(name = "-sc", aliases = {"--side"}, usage = "an optional file containing list of side compounds to ignore")
+ @Option(name = "-sc", aliases = {"--side"}, usage = "a file containing list of side compounds to ignore", required = true)
public String sideCompoundFile = null;
+
@Option(name = "-cw", aliases = {"--customWeights"}, usage = "an optional file containing weights for reactions pairs")
public String weightFile = null;
--
GitLab
From a8bd7fac4fe452338fbe544793c39075c6a5cba0 Mon Sep 17 00:00:00 2001
From: cfrainay <clement.frainay@inrae.fr>
Date: Tue, 22 Feb 2022 16:52:38 +0100
Subject: [PATCH 7/7] Improve doc, mention weight file
---
.../networkAnalysis/ExtractSubReactionNetwork.java | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
index d5e04ede5..64fab85fd 100644
--- a/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
+++ b/met4j-toolbox/src/main/java/fr/inrae/toulouse/metexplore/met4j_toolbox/networkAnalysis/ExtractSubReactionNetwork.java
@@ -128,7 +128,7 @@ public class ExtractSubReactionNetwork extends AbstractMet4jApplication {
"The subnetwork corresponds to part of the network that connects reactions from the first list to reactions from the second list.\n" +
"Sources and targets list can have elements in common. The connecting part can be defined as the union of shortest or k-shortest paths between sources and targets, " +
"or the Steiner tree connecting them. Contrary to compound graph, reaction graph often lacks weighting policy for edge relevance. In order to ensure appropriate " +
- "network density, a list of side compounds to ignore for linking reactions must be provided";
+ "network density, a list of side compounds to ignore for linking reactions must be provided. An optional edge weight file, if available, can also be used.";
}
@Override
--
GitLab