Use PhenotypeHighlight for heat map column header highlighting

Author: mikekucera

EnrichmentMapPlugin/src/main/java/org/baderlab/csplugins/enrichmentmap/model/CompressedClass.java

@@ -3,14 +3,16 @@
 import java.util.ArrayList;
 import java.util.LinkedHashSet;
 import java.util.List;
+import java.util.Map;
 import java.util.Optional;
 
-import org.baderlab.csplugins.enrichmentmap.model.Phenotype.Type;
+import org.baderlab.csplugins.enrichmentmap.model.PhenotypeHighlight.Highlight;
 
 public class CompressedClass implements ExpressionData {
 
+	private final boolean commonButDiffPheno;
 	private final ExpressionCache expressionCache;
-	private List<Phenotype> headers = new ArrayList<>();
+	private List<PhenotypeHighlight> headers = new ArrayList<>();
 
 	/**
 	 * If the expressions are the same for multiple data sets then we will only get one data set passed in
@@ -22,52 +24,62 @@ public CompressedClass(EnrichmentMap map, List<EMDataSet> datasets, ExpressionCa
 		// Special case. There is more than one dataset in the map, and they have the same expressions,
 		// but the pheontypes are different. We want to show the phenotypes for the selected data sets
 		// but we don't want to repeat all the expression data.
-		if(ExpressionData.commonExpressionsButDifferentPhenotypes(map)) {
+		commonButDiffPheno = ExpressionData.commonExpressionsButDifferentPhenotypes(map);
+		
+		if(commonButDiffPheno) {
 			// we only get passed one dataset because expressions are the same
 			var dataset = datasets.get(0); 
+			var allDatasets = map.getDataSetList();
 			var classPhenos = ExpressionData.getPhenotypesFromClassFile(dataset);
 			var highlightPhenos = ExpressionData.getPhenotypesToHighlight(map.getDataSetList()); // get the phenotypes from all the data sets
-			addHeaders(dataset, classPhenos, highlightPhenos);
+			addHeaders(allDatasets, classPhenos, highlightPhenos);
 
 		} else {
 			for(var dataset : datasets) {
 				var classPhenos = ExpressionData.getPhenotypesFromClassFile(dataset);
 				var highlightPhenos = ExpressionData.getPhenotypesToHighlight(dataset);
-				addHeaders(dataset, classPhenos, highlightPhenos);
+				addHeaders(List.of(dataset), classPhenos, highlightPhenos);
 			}
 		}
 	}
 
+	@Override
+	public boolean commonButDiffPheno() {
+		return commonButDiffPheno;
+	}
 
-	private void addHeaders(EMDataSet mainDataSet, LinkedHashSet<String> classPhenos, List<Phenotype> highlightPhenos) {
+	private void addHeaders(List<EMDataSet> datasets, LinkedHashSet<String> classPhenos, Map<String,PhenotypeHighlight> highlightPhenos) {
 		// move highlighted phenotypes to the front
-		for(Phenotype pheno : highlightPhenos) {
+		var phenos = new ArrayList<>(highlightPhenos.values());
+		
+		for(PhenotypeHighlight pheno : phenos) {
 			if(classPhenos.contains(pheno.getName())) {
 				headers.add(pheno);
 			}
 		}
-		for(Phenotype pheno : highlightPhenos) {
+		for(PhenotypeHighlight pheno : phenos) {
 			classPhenos.remove(pheno.getName());
 		}
 
 		for(String phenoName : classPhenos) {
-			headers.add(new Phenotype(mainDataSet, phenoName, Type.OTHER));
+			headers.add(new PhenotypeHighlight(datasets, phenoName, Highlight.NONE));
 		}
 	}
 
 
 	@Override
 	public EMDataSet getDataSet(int idx) {
-		return headers.get(idx).getDataSet();
+		// all of the data sets should have the same expression values, so returning any one of them is fine.
+		return getHighlight(idx).getDatasets().iterator().next();
 	}
-
+	
 	@Override
 	public String getName(int idx) {
-		return getPhenotype(idx).getName();
+		return getHighlight(idx).getName();
 	}
 
 	@Override
-	public Phenotype getPhenotype(int idx) {
+	public PhenotypeHighlight getHighlight(int idx) {
 		return headers.get(idx);
 	}
 

EnrichmentMapPlugin/src/main/java/org/baderlab/csplugins/enrichmentmap/model/CompressedDataSet.java

@@ -3,6 +3,8 @@
 import java.util.List;
 import java.util.Optional;
 
+import org.baderlab.csplugins.enrichmentmap.model.PhenotypeHighlight.Highlight;
+
 public class CompressedDataSet implements ExpressionData {
 
 	private final ExpressionCache expressionCache;
@@ -29,10 +31,10 @@ public double getValue(int geneID, int idx, Compress compress, Transform transfo
 			return Float.NaN;
 
 		switch (compress) {
-			case DATASET_MEDIAN:	return GeneExpression.median(expression.get());
-			case DATASET_MAX:	return GeneExpression.max(expression.get());
-			case DATASET_MIN:	return GeneExpression.min(expression.get());
-			default:				return Float.NaN;
+			case DATASET_MEDIAN: return GeneExpression.median(expression.get());
+			case DATASET_MAX:	 return GeneExpression.max(expression.get());
+			case DATASET_MIN:	 return GeneExpression.min(expression.get());
+			default:			 return Float.NaN;
 		}
 	}
 
@@ -46,4 +48,16 @@ public String getName(int idx) {
 	public int getSize() {
 		return datasets.size();
 	}
+
+	@Override
+	public PhenotypeHighlight getHighlight(int col) {
+		var dataSet = getDataSet(col);
+		var name = getName(col);
+		return new PhenotypeHighlight(dataSet, name, Highlight.NONE);
+	}
+	
+	@Override
+	public boolean commonButDiffPheno() {
+		return false;
+	}
 }

EnrichmentMapPlugin/src/main/java/org/baderlab/csplugins/enrichmentmap/model/ExpressionData.java

@@ -1,67 +1,48 @@
 package org.baderlab.csplugins.enrichmentmap.model;
 
-import java.util.ArrayList;
 import java.util.Iterator;
+import java.util.LinkedHashMap;
 import java.util.LinkedHashSet;
 import java.util.List;
+import java.util.Map;
+
+import org.baderlab.csplugins.enrichmentmap.model.PhenotypeHighlight.Highlight;
 
-import org.baderlab.csplugins.enrichmentmap.model.Phenotype.Type;
 
 /**
- * Common interface for different levels of compression.
+ * Common interface for ways of representing (compressing) expression data for use in the HeatMap.
  */
 public interface ExpressionData {
 
-	EMDataSet getDataSet(int col);
-
-	double getValue(int geneID, int col, Compress compress, Transform transform);
-
-	String getName(int col);
-
-	public default Phenotype getPhenotype(int col) {
-		return null;
-	};
-
+	/**
+	 * number of columns
+	 */
 	int getSize();
 
-	
-	
 	/**
-	 * Special case. There is more than one dataset in the map, and they have the same expressions,
-	 * but the pheontypes are different. We want to show the phenotypes for the selected data sets
-	 * but we don't want to repeat all the expression data.
+	 * column name
 	 */
-	static boolean commonExpressionsButDifferentPhenotypes(EnrichmentMap map) {
-		return map != null && map.isCommonExpressionValues() && !phenotypesAreCommon(map);
-	}
+	String getName(int col); 
 
+	/**
+	 * Returns expression value for given gene and column.
+	 */
+	double getValue(int geneID, int col, Compress compress, Transform transform);
+
+	/**
+	 * Returns a data set that contains the expression value. Note, if multiple data
+	 * sets have the same expression values, only one will be returned.
+	 */
+	EMDataSet getDataSet(int col);
 
 	/**
-	 * These are the phenotypes entered into the "Phenotypes" fields in the creation dialog.
-	 * They need to be highlighted in the table.
+	 * Info on how to highlight a column based on the classes and phenotypes.
+	 * Should not return null.
 	 */
-	static List<Phenotype> getPhenotypesToHighlight(List<EMDataSet> datasets) {
-		List<Phenotype> phenos = new ArrayList<>();
-		
-		for(var dataset : datasets) {
-			var enrichments = dataset.getEnrichments();
-			
-			String pheno1 = enrichments.getPhenotype1();
-			if(pheno1 != null) {
-				phenos.add(new Phenotype(dataset, pheno1, Type.POSITIVE));
-			}
-			String pheno2 = enrichments.getPhenotype2();
-			if(pheno2 != null) {
-				phenos.add(new Phenotype(dataset, pheno2, Type.NEGATIVE));
-			}
-		}
-		
-		return phenos;
-	}
+	PhenotypeHighlight getHighlight(int col);
+
 
-	static List<Phenotype> getPhenotypesToHighlight(EMDataSet dataset) {
-		return getPhenotypesToHighlight(List.of(dataset));
-	}
+	boolean commonButDiffPheno();
 
 
 	/**
@@ -84,7 +65,50 @@ static LinkedHashSet<String> getPhenotypesFromClassFile(EMDataSet dataset) {
 	}
 
 
-	static boolean phenotypesAreCommon(EnrichmentMap map) {
+	/**
+	 * These are the phenotypes entered into the "Phenotypes" fields in the creation dialog.
+	 * They need to be highlighted in the table.
+	 */
+	static Map<String,PhenotypeHighlight> getPhenotypesToHighlight(List<EMDataSet> datasets) {
+		Map<String,PhenotypeHighlight> phenoMap = new LinkedHashMap<>();
+		
+		for (var dataset : datasets) {
+			var enrichments = dataset.getEnrichments();
+
+			String phenoNamePos = enrichments.getPhenotype1();
+			if(phenoNamePos != null) {
+				var phenoHighlight = new PhenotypeHighlight(dataset, phenoNamePos, Highlight.POSITIVE);
+				var existingPheno = phenoMap.get(phenoNamePos);
+				phenoMap.put(phenoNamePos, phenoHighlight.merge(existingPheno));
+			}
+			
+			String phenoNameNeg = enrichments.getPhenotype2();
+			if (phenoNameNeg != null) {
+				var phenoHighlight = new PhenotypeHighlight(dataset, phenoNameNeg, Highlight.NEGATIVE);
+				var existingPheno = phenoMap.get(phenoNameNeg);
+				phenoMap.put(phenoNameNeg, phenoHighlight.merge(existingPheno));
+			}
+		}
+		
+		return phenoMap;
+	}
+	
+	static Map<String,PhenotypeHighlight> getPhenotypesToHighlight(EMDataSet dataset) {
+		return getPhenotypesToHighlight(List.of(dataset));
+	}
+	
+	
+	
+	/**
+	 * Special case. There is more than one dataset in the map, and they have the same expressions,
+	 * but the pheontypes are different. We want to show the phenotypes for the selected data sets
+	 * but we don't want to repeat all the expression data.
+	 */
+	static boolean commonExpressionsButDifferentPhenotypes(EnrichmentMap map) {
+		return map != null && map.isCommonExpressionValues() && !phenotypesAreCommon(map);
+	}
+	
+	private static boolean phenotypesAreCommon(EnrichmentMap map) {
 		Iterator<EMDataSet> iter = map.getDataSets().values().iterator();
 		SetOfEnrichmentResults r = iter.next().getEnrichments();
 		String p1 = r.getPhenotype1();

EnrichmentMapPlugin/src/main/java/org/baderlab/csplugins/enrichmentmap/model/Phenotype.java

@@ -0,0 +1,73 @@
+package org.baderlab.csplugins.enrichmentmap.model;
+
+import java.util.LinkedHashMap;
+import java.util.LinkedHashSet;
+import java.util.List;
+import java.util.Objects;
+import java.util.Set;
+
+public class PhenotypeHighlight {
+	
+	public static enum Highlight {
+		POSITIVE,
+		NEGATIVE,
+		NONE
+	}
+	
+	private final LinkedHashMap<EMDataSet, Highlight> highlights;
+	private final String name; // A bit redundant, ExpressionData has the same information
+	
+	
+	public PhenotypeHighlight(EMDataSet dataset, String name, Highlight highlight) {
+		this.highlights = new LinkedHashMap<>();
+		this.highlights.put(Objects.requireNonNull(dataset), Objects.requireNonNull(highlight));
+		this.name = Objects.requireNonNull(name);
+	}
+	
+	public PhenotypeHighlight(List<EMDataSet> datasets, String name, Highlight highlight) {
+		this.highlights = new LinkedHashMap<>();
+		for(var ds : datasets) {
+			this.highlights.put(ds, highlight);
+		}
+		this.name = Objects.requireNonNull(name);
+	}
+	
+	private PhenotypeHighlight(LinkedHashMap<EMDataSet, Highlight> highlights, String name) {
+		this.highlights = highlights;
+		this.name = name;
+	}
+	
+	public Set<EMDataSet> getDatasets() {
+		return new LinkedHashSet<EMDataSet>(highlights.keySet());
+	}
+
+	public String getName() {
+		return name;
+	}
+
+	public Highlight getHighlight(EMDataSet dataset) {
+		return highlights.get(dataset);
+	}
+	
+	
+	public PhenotypeHighlight merge(PhenotypeHighlight that) {
+		if(that == null)
+			return this;
+		
+		if(name.equals(that.name)) {
+			var highlights = new LinkedHashMap<EMDataSet, Highlight>();
+			highlights.putAll(this.highlights);
+			highlights.putAll(that.highlights);
+			return new PhenotypeHighlight(highlights, name);
+		}
+
+		throw new IllegalArgumentException("Phenotype names do not match: " + name + " != " + that.name);
+	}
+
+	@Override
+	public String toString() {
+		return "PhenotypeHighlight[name=" + name + ", highlights=" + highlights + "]";
+	}
+	
+	
+}