Fine tuned server-side behavior of some parameters.

Added a sample analysis folder.
Added new executable for converting SIF files to json.

Author: ozgunbabur

src/main/java/org/panda/causalpath/network/GraphWriter.java

@@ -4,6 +4,7 @@
 import org.panda.causalpath.analyzer.NSCForNonCorr;
 import org.panda.causalpath.analyzer.NetworkSignificanceCalculator;
 import org.panda.causalpath.data.*;
+import org.panda.utility.ArrayUtil;
 import org.panda.utility.CollectionUtil;
 import org.panda.utility.FileUtil;
 import org.panda.utility.ValToColor;
@@ -477,18 +478,7 @@ else if (data instanceof ActivityData)
 
 			for (String sym : data.getGeneSymbols())
 			{
-				if (!geneMap.containsKey(sym))
-				{
-					HashMap<String, Object> node = new HashMap<>();
-					geneMap.put(sym, node);
-					nodes.add(node);
-					Map<String, Object> d = new HashMap<>();
-					node.put("data", d);
-					d.put("id", sym);
-					d.put("text", sym);
-					List<Map> sites = new ArrayList<>();
-					d.put("sites", sites);
-				}
+				initJsonNode(geneMap, nodes, sym);
 
 				Map node = geneMap.get(sym);
 
@@ -541,6 +531,134 @@ else if (data instanceof ActivityData)
 		writer.close();
 	}
 
+	/**
+	 * This method takes in a SIF graph defined by two files (.sif and .format), and generates a corresponding .json
+	 * file that the webserver can display.
+	 *
+	 * @param sifFileanme SIF filename
+	 * @param formatFilename Format filename
+	 * @param outJasonFilename JASON filename to produce
+	 */
+	public static void convertSIFToJSON(String sifFileanme, String formatFilename, String outJasonFilename) throws IOException
+	{
+		Map<String, Object> map = new HashMap<>();
+		List<Map> nodes = new ArrayList<>();
+		List<Map> edges = new ArrayList<>();
+		map.put("nodes", nodes);
+		map.put("edges", edges);
+
+		Map<String, Map> nodeMap = new HashMap<>();
+		Set<String> relMem = new HashSet<>();
+
+		Files.lines(Paths.get(sifFileanme)).map(l -> l.split("\t")).forEach(t ->
+		{
+			if (t.length > 2)
+			{
+				String key = t[0] + "\t" + t[1] + "\t" + t[2];
+				if (relMem.contains(key)) return;
+				else relMem.add(key);
+
+				Map<String, Object> edge = new HashMap<>();
+				edges.add(edge);
+				Map<String, Object> dMap = new HashMap<>();
+				edge.put("data", dMap);
+				dMap.put("source", t[0]);
+				dMap.put("target", t[2]);
+				dMap.put("edgeType", t[1]);
+				if (t.length > 4 && !t[4].trim().isEmpty())
+				{
+					dMap.put("tooltipText", t[2] + "-" + CollectionUtil.merge(Arrays.asList(t[4].split(";")), "-"));
+				}
+
+				if (t.length > 3 && !t[3].trim().isEmpty())
+				{
+					List<String> medList = Arrays.asList(t[3].split(";| "));
+					if (!medList.isEmpty())
+					{
+						dMap.put("pcLinks", medList);
+					}
+				}
+				initJsonNode(nodeMap, nodes, t[2]);
+			}
+
+			if (t.length > 0 && !t[0].isEmpty()) initJsonNode(nodeMap, nodes, t[0]);
+		});
+
+		Map<String, String> defaultColors = new HashMap<>();
+		String defBGCKey = "node BG color";
+		String defBorCKey = "node border color";
+
+		Files.lines(Paths.get(formatFilename)).map(l -> l.split("\t")).filter(t -> t.length > 3).forEach(t ->
+		{
+			if (t[1].equals("all-nodes"))
+			{
+				if (t[2].equals("color")) defaultColors.put(defBGCKey, jasonizeColor(t[3]));
+				else if (t[2].equals("bordercolor")) defaultColors.put(defBorCKey, jasonizeColor(t[3]));
+			}
+
+			if (t[0].equals("node"))
+			{
+				String name = t[1];
+				Map node = nodeMap.get(name);
+				if (node != null)
+				{
+					if (!node.containsKey("css")) node.put("css", new HashMap<>());
+
+					switch (t[2])
+					{
+						case "rppasite":
+							String[] x = t[3].split("\\|");
+							Map site = new HashMap();
+							((List) ((Map) node.get("data")).get("sites")).add(site);
+							site.put("siteText", x[1]);
+							site.put("siteInfo", x[0] + (x.length > 4 ? (" " + x[4]) : ""));
+							site.put("siteBackgroundColor", jasonizeColor(x[2]));
+							site.put("siteBorderColor", jasonizeColor(x[3]));
+							break;
+						case "color":
+							((Map) node.get("css")).put("backgroundColor", jasonizeColor(t[3]));
+							break;
+						case "bordercolor":
+							((Map) node.get("css")).put("borderColor", jasonizeColor(t[3]));
+							break;
+						case "borderwidth":
+							((Map) node.get("css")).put("borderWidth", t[3] + "px");
+							break;
+						case "tooltip":
+							((Map) node.get("data")).put("tooltipText", t[3]);
+							break;
+					}
+				}
+			}
+		});
+
+		BufferedWriter writer = Files.newBufferedWriter(Paths.get(outJasonFilename));
+		JsonUtils.writePrettyPrint(writer, map);
+		writer.close();
+	}
+
+	private static void initJsonNode(Map<String, Map> nodeMap, List<Map> nodes, String name)
+	{
+		if (!nodeMap.containsKey(name))
+		{
+			HashMap<String, Object> node = new HashMap<>();
+			nodeMap.put(name, node);
+			nodes.add(node);
+			Map<String, Object> d = new HashMap<>();
+			node.put("data", d);
+			d.put("id", name);
+			d.put("text", name);
+			List<Map> sites = new ArrayList<>();
+			d.put("sites", sites);
+		}
+	}
+
+	private static String jasonizeColor(String c)
+	{
+		String[] t = c.split(" ");
+		return "rgb(" + ArrayUtil.getString(",", t[0], t[1], t[2]) + ")";
+	}
+
 	private Set<ExperimentData> getExperimentDataToDraw()
 	{
 		if (experimentDataToDraw != null) return experimentDataToDraw;

src/main/java/org/panda/causalpath/run/CausalPath.java

@@ -1035,9 +1035,8 @@ enum ValueTransformation
 			"detection, using the threshold-for-data-significance.", false),
 
 		MAX("max", "The value with maximum absolute is used for the analysis. There should only be one group of " +
-			"values (marked with value-column), the values have to be distributed around " +
-			"zero, and a threshold value should be provided for significance detection, using the " +
-			"threshold-for-data-significance.", false),
+			"values (marked with value-column), the values have to be distributed around zero, and a threshold value " +
+			"should be provided for significance detection, using the threshold-for-data-significance.", false),
 
 		DIFFERENCE_OF_MEANS("difference-of-means", "There should be control and test values, whose difference would " +
 			"be used for significance detection. The threshold for significance (threshold-for-data-significance) " +
@@ -1052,13 +1051,13 @@ enum ValueTransformation
 		SIGNIFICANT_CHANGE_OF_MEAN("significant-change-of-mean", "There should be sufficient amount of control and " +
 			"test values to detect the significance of change with a t-test. Technically there should be more than 3" +
 			" controls and 3 tests, practically, they should be much more to provide statistical power. The " +
-			"threshold-for-data-significance should be used for a p-value threshold, or " +
-			"alternatively, fdr-threshold-for-data-significance should be used for " +
-			"controlling significance at the false discovery rate level.", true),
+			"threshold-for-data-significance should be used for a p-value threshold, or alternatively, " +
+			"fdr-threshold-for-data-significance should be used for controlling significance at the false discovery " +
+			"rate level.", true),
 
 		SIGNIFICANT_CHANGE_OF_MEAN_PAIRED("significant-change-of-mean-paired", "There should be sufficient amount of " +
 			"control and test values to detect the significance of change with a paired t-test. Technically there " +
-			"should be more than 3 controls and 3 tests, practically, they should be much more to provide statistical" +
+			"should be at least 3 controls and 3 tests, practically, they should be much more to provide statistical" +
 			" power. The order of control and test value columns indicate the pairing. First control column in the " +
 			"parameters file is paired with first test column, second is paired with second, etc. The " +
 			"threshold-for-data-significance should be used for a p-value threshold, or alternatively, " +
@@ -1068,11 +1067,18 @@ enum ValueTransformation
 		SIGNED_P_VALUES("signed-p-values", "If the dataset has its own calculation of p-values desired to be used " +
 			"directly, then for each comparison, there must be a column in the dataset that has these p-values " +
 			"multiplied with the sign of the change. For instance a value -0.001 means downregulation with a p-value " +
-			"of 0.001. If an FDR control is desired, then the p-values should not be adjusted.", false),
+			"of 0.001. Don't use 0 and -0 values in the data file with this option. Java cannot distinguish between " +
+			"the two. Instead, convert 0 values to very small but nonzero values, such as 1e-10 or -1e-10. " +
+			"If an FDR control is desired, there are two ways to have it. First way is to use unadjusted p-values in " +
+			"the data file and then use the fdr-threshold-for-data-significance parameter to set the desired FDR " +
+			"level. Second way is to use adjusted p-values in the data file and use the " +
+			"threshold-for-data-significance parameter to set the FDR level. Don't mix these two ways. " +
+			"If fdr-threshold-for-data-significance is used over already-adjusted p-values, then the code will apply " +
+			"the Benjamini-Hochberg procedure over those already-adjusted p-values.", false),
 
 		CORRELATION("correlation", "There should be one group of values (marked with value-column). There must be at " +
-			"least 3 value columns technically, but many more " +
-			"than that practically to have some statistical power for significant correlation. ", false);
+			"least 3 value columns technically, but many more than that practically to have some statistical power " +
+			"for significant correlation. ", false);
 
 		ValueTransformation(String name, String description, boolean twoGroupComparison)
 		{
@@ -1332,7 +1338,9 @@ enum Parameter
 		CALCULATE_NETWORK_SIGNIFICANCE((value, cp) -> cp.calculateNetworkSignificance = webServerMode ? false : Boolean.valueOf(value),
 			"Calculate network significance",
 			"Whether to calculate significances of the properties of the graph. When turned on, a p-value for network" +
-				" size, and also downstream activity enrichment p-values for each gene on the graph are calculated.",
+				" size, and also downstream activity enrichment p-values for each gene on the graph are calculated. " +
+				"This parameter is ignored by webserver due to resource limitations. To calculate network significance, " +
+				"please run CausalPath locally from its JAR file.",
 			new EntryType(Boolean.class), new Boolean[][]{{Boolean.FALSE}}, true, false, new Cond(Logical.NOT)),
 		PERMUTATIONS_FOR_SIGNIFICANCE((value, cp) -> cp.permutationCount = Integer.valueOf(value),
 			"Number of permutations for calculating network significance",
@@ -1410,7 +1418,7 @@ enum Parameter
 			new String[]{NetworkLoader.ResourceType.PC.name()},
 			new String[]{NetworkLoader.ResourceType.PhosphoNetworks.name()},
 			new String[]{NetworkLoader.ResourceType.IPTMNet.name()}},
-			true, true, null),
+			false, true, null),
 		RELATION_FILTER_TYPE((value, cp) ->
 		{
 			if (!cp.cs.hasGraphFilter())
@@ -1453,7 +1461,7 @@ enum Parameter
 			"Specifies the value where node colors reach most intense color. Has to be a positive value, and used " +
 				"symmetrically. In the case of value-transformation is significant-change-of-mean, the value is " +
 				"-log(p) with a sign associated to it.",
-			new EntryType(Double.class), new String[][]{{"1"}}, false, false,
+			new EntryType(Double.class), new String[][]{{"10"}}, false, false,
 			new Cond( Logical.NOT, new Cond(VALUE_TRANSFORMATION.getText(), ValueTransformation.CORRELATION.name))),
 		SHOW_ALL_GENES_WITH_PROTEOMIC_DATA((value, cp) -> cp.showAllGenesWithProteomicData = Boolean.valueOf(value),
 			"Show all genes with significant proteomic data",

src/main/java/org/panda/causalpath/run/JasonizeResultGraphsRecursively.java

@@ -0,0 +1,47 @@
+package org.panda.causalpath.run;
+
+import org.panda.causalpath.network.GraphWriter;
+
+import java.io.File;
+import java.io.IOException;
+import java.nio.file.Files;
+import java.nio.file.Paths;
+
+/**
+ * This class is used for converting SIF and format files to json files that can be uploaded to the webserver for
+ * visualization.
+ *
+ * Usage: java -jar causalpath.jar <in-dir> <sif-name> <out-dir> <json-name>
+ *
+ */
+public class JasonizeResultGraphsRecursively
+{
+	public static void main(String[] args) throws IOException
+	{
+		String inBase = new File(args[0]).getPath();
+		String sifNameWOExtension = args[1];
+		String outBase = new File(args[2]).getPath();
+		String jsonName = args.length > 3 ? args[3] : "causative.json";
+
+		generate(inBase, inBase, sifNameWOExtension, outBase, jsonName);
+	}
+
+	private static void generate(String inBase, String inDir, String sifName, String outBase, String jsonName) throws IOException
+	{
+		String sifPath = inDir + File.separator + sifName + ".sif";
+		String formatPath = inDir + File.separator + sifName + ".format";
+
+		if (Files.exists(Paths.get(sifPath)) && Files.exists(Paths.get(formatPath)))
+		{
+			String outDir = inDir.replace(inBase, outBase);
+			Files.createDirectories(Paths.get(outDir));
+
+			GraphWriter.convertSIFToJSON(sifPath, formatPath, outDir + File.separator + jsonName);
+		}
+
+		for (File sub : new File(inDir).listFiles())
+		{
+			if (sub.isDirectory()) generate(inBase, sub.getPath(), sifName, outBase, jsonName);
+		}
+	}
+}