Examples with EMDataSet org.baderlab.csplugins.enrichmentmap.model.EMDataSet used on opensource projects

Example 1 with EMDataSet

use of org.baderlab.csplugins.enrichmentmap.model.EMDataSet in project EnrichmentMapApp by BaderLab.

the class FilterGenesetsByDatasetGenes method filterGenesets.

/*
	 * Filter all the genesets by the dataset genes. If there are multiple sets
	 * of genesets make sure to filter by the specific dataset genes.
	 */
public void filterGenesets(TaskMonitor taskMonitor) {
    Map<String, EMDataSet> datasets = map.getDataSets();
    for (String k : datasets.keySet()) {
        taskMonitor.setStatusMessage("Filtering Data Set: " + k);
        EMDataSet current_set = datasets.get(k);
        //only filter the genesets if dataset genes are not null or empty
        Set<Integer> datasetGenes = current_set.getDataSetGenes();
        if (datasetGenes != null && !datasetGenes.isEmpty()) {
            current_set.getGeneSetsOfInterest().filterGeneSets(datasetGenes);
        } else {
            System.out.println("Dataset Genes is empty, because expression and ranks not provided: " + current_set.getName());
        }
    }
    // of the genes in the expression file are in the specified genesets.
    if (!anyGenesLeftAfterFiltering(datasets.values()))
        throw new IllegalThreadStateException("No genes in the expression file are found in the GMT file ");
    // if there are multiple datasets check to see if they have the same set of genes
    if (datasetsAreDistinct(datasets.values())) {
        map.setDistinctExpressionSets(true);
    }
}

Example 2 with EMDataSet

use of org.baderlab.csplugins.enrichmentmap.model.EMDataSet in project EnrichmentMapApp by BaderLab.

the class FilterGenesetsByDatasetGenes method datasetsAreDistinct.

private static boolean datasetsAreDistinct(Collection<EMDataSet> datasets) {
    Set<Set<Integer>> uniqueGeneSets = new HashSet<>();
    for (EMDataSet dataset : datasets) {
        Set<Integer> genes = dataset.getDataSetGenes();
        uniqueGeneSets.add(genes);
        if (uniqueGeneSets.size() > 1) {
            return true;
        }
    }
    return false;
}

Example 3 with EMDataSet

use of org.baderlab.csplugins.enrichmentmap.model.EMDataSet in project EnrichmentMapApp by BaderLab.

the class InitializeGenesetsOfInterestTask method initializeSets.

/**
	 * filter the genesets, restricting them to only those passing the user
	 * specified thresholds.
	 * 
	 * @return true if successful and false otherwise.
	 */
public boolean initializeSets(TaskMonitor tm) {
    if (tm == null)
        tm = new NullTaskMonitor();
    DiscreteTaskMonitor taskMonitor = new DiscreteTaskMonitor(tm, map.getDataSetCount());
    //create subset of genesets that contains only the genesets of interest with pvalue and qbalue less than values specified by the user.
    //Go through each Dataset populating the Gene set of interest in each dataset object
    Map<String, EMDataSet> datasets = map.getDataSets();
    // count how many experiments (DataSets) contain the geneset
    Optional<Integer> minExperiments = map.getParams().getMinExperiments();
    Map<String, Integer> occurrences = minExperiments.isPresent() ? new HashMap<>() : null;
    for (String datasetName : datasets.keySet()) {
        taskMonitor.inc();
        EMDataSet dataset = datasets.get(datasetName);
        // all these maps use the geneset name as key
        Map<String, EnrichmentResult> enrichmentResults = dataset.getEnrichments().getEnrichments();
        Map<String, GeneSet> genesets = dataset.getSetOfGeneSets().getGeneSets();
        Map<String, GeneSet> genesetsOfInterest = dataset.getGeneSetsOfInterest().getGeneSets();
        // If there are no genesets associated with this dataset then get the complete set assumption being that the gmt file applies to all datasets.
        if (genesets == null || genesets.isEmpty()) {
            genesets = map.getAllGeneSets();
        }
        //if there are no enrichment Results then do nothing
        if (enrichmentResults == null || enrichmentResults.isEmpty()) {
            return false;
        }
        //iterate through the GSEA Results to figure out which genesets we want to use
        for (String genesetName : enrichmentResults.keySet()) {
            EnrichmentResult result = enrichmentResults.get(genesetName);
            // update rank at max for leading edge calculation
            if (dataset.getMethod() == Method.GSEA) {
                Ranking ranks = dataset.getExpressionSets().getRanksByName(datasetName);
                updateRankAtMax((GSEAResult) result, ranks);
            }
            if (result.geneSetOfInterest(map.getParams())) {
                GeneSet geneset = genesets.get(genesetName);
                if (geneset != null) {
                    // while we are checking, update the size of the genesets based on post filtered data
                    result.setGsSize(geneset.getGenes().size());
                    if (occurrences != null) {
                        occurrences.merge(genesetName, 1, (v, d) -> v + 1);
                    }
                    genesetsOfInterest.put(genesetName, geneset);
                } else if (throwIfMissing) {
                    throw new IllegalThreadStateException("The Geneset: " + genesetName + " is not found in the GMT file.");
                }
            }
        }
    }
    // Remove gene-sets that don't pass the minimum occurrence cutoff
    if (occurrences != null) {
        for (EMDataSet dataset : datasets.values()) {
            Map<String, GeneSet> genesetsOfInterest = dataset.getGeneSetsOfInterest().getGeneSets();
            genesetsOfInterest.keySet().removeIf(geneset -> occurrences.getOrDefault(geneset, 0) < minExperiments.get());
        }
    }
    return true;
}

Example 4 with EMDataSet

use of org.baderlab.csplugins.enrichmentmap.model.EMDataSet in project EnrichmentMapApp by BaderLab.

the class CreateEMNetworkTask method createEdges.

/**
	 * Note, we expect that GenesetSimilarity object that don't pass the cutoff have already been filtered out.
	 * @param network
	 * @param nodes
	 */
private void createEdges(CyNetwork network, Map<String, CyNode> nodes) {
    Map<SimilarityKey, GenesetSimilarity> similarities = supplier.get();
    for (SimilarityKey key : similarities.keySet()) {
        GenesetSimilarity similarity = similarities.get(key);
        CyNode node1 = nodes.get(similarity.getGeneset1Name());
        CyNode node2 = nodes.get(similarity.getGeneset2Name());
        CyEdge edge = network.addEdge(node1, node2, false);
        String datasetName = key.getName();
        if (datasetName != null) {
            EMDataSet dataset = map.getDataSet(datasetName);
            if (dataset != null) {
                dataset.addEdgeSuid(edge.getSUID());
            }
        }
        List<String> overlapGenes = similarity.getOverlappingGenes().stream().map(map::getGeneFromHashKey).collect(Collectors.toList());
        String edgeName = key.toString();
        CyRow row = network.getRow(edge);
        row.set(CyNetwork.NAME, edgeName);
        row.set(CyEdge.INTERACTION, similarity.getInteractionType());
        Columns.EDGE_SIMILARITY_COEFF.set(row, prefix, null, similarity.getSimilarityCoeffecient());
        Columns.EDGE_OVERLAP_SIZE.set(row, prefix, null, similarity.getSizeOfOverlap());
        Columns.EDGE_OVERLAP_GENES.set(row, prefix, null, overlapGenes);
        if (key.isCompound()) {
            Columns.EDGE_DATASET.set(row, prefix, null, Columns.EDGE_DATASET_VALUE_COMPOUND);
        } else {
            Columns.EDGE_DATASET.set(row, prefix, null, similarity.getDataSetName());
        }
    }
}

Example 5 with EMDataSet

use of org.baderlab.csplugins.enrichmentmap.model.EMDataSet in project EnrichmentMapApp by BaderLab.

the class CreateEnrichmentMapTaskFactory method createTaskIterator.

@Override
public TaskIterator createTaskIterator() {
    TaskIterator tasks = new TaskIterator();
    if (dataSets.isEmpty())
        return tasks;
    tasks.append(new TitleTask("Building EnrichmentMap"));
    EnrichmentMap map = new EnrichmentMap(params, serviceRegistrar);
    for (DataSetParameters dataSetParameters : dataSets) {
        String datasetName = dataSetParameters.getName();
        Method method = dataSetParameters.getMethod();
        DataSetFiles files = dataSetParameters.getFiles();
        EMDataSet dataset = map.createDataSet(datasetName, method, files);
        // Load GMT File
        if (!Strings.isNullOrEmpty(dataset.getSetOfGeneSets().getFilename())) {
            tasks.append(new GMTFileReaderTask(dataset));
        }
        // Load the enrichments 
        tasks.append(new DetermineEnrichmentResultFileReader(dataset).getParsers());
        // If there is no expression file then create a dummy file to associate with this dataset so we can still use the expression viewer (heat map)
        if (Strings.isNullOrEmpty(dataset.getDataSetFiles().getExpressionFileName())) {
            tasks.append(new CreateDummyExpressionTask(dataset));
        } else {
            tasks.append(new ExpressionFileReaderTask(dataset));
        }
        // Load ranks if present
        String ranksName = dataset.getMethod() == Method.GSEA ? Ranking.GSEARanking : datasetName;
        if (dataset.getExpressionSets().getRanksByName(ranksName) != null) {
            String filename = dataset.getExpressionSets().getRanksByName(ranksName).getFilename();
            tasks.append(new RanksFileReaderTask(filename, dataset, ranksName, false));
        }
        if (!Strings.isNullOrEmpty(dataset.getDataSetFiles().getClassFile())) {
            tasks.append(new ClassFileReaderTask(dataset));
        }
    }
    // NOTE: First filter out genesets that don't pass the thresholds, 
    // Then filter the remaining genesets of interest to only contain genes from the expression file.
    // Filter out genesets that don't pass the p-value and q-value thresholds
    InitializeGenesetsOfInterestTask genesetsTask = new InitializeGenesetsOfInterestTask(map);
    //		genesetsTask.setThrowIfMissing(false); // TEMPORARY
    tasks.append(genesetsTask);
    // Trim the genesets to only contain the genes that are in the data file.
    tasks.append(new FilterGenesetsByDatasetGenes(map));
    // Link the ComputeSimilarityTask to the MasterMapNetworkTask by a "pipe"
    Baton<Map<SimilarityKey, GenesetSimilarity>> pipe = new Baton<>();
    // Compute the geneset similarities
    tasks.append(new ComputeSimilarityTaskParallel(map, pipe.consumer()));
    // Create the network
    tasks.append(createEMNetworkTaskFactory.create(map, pipe.supplier()));
    // Create style and layout
    if (!headless) {
        tasks.append(createEMViewTaskFactory.create(map));
    }
    return tasks;
}