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

Example 1 with GeneSet

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

the class LegacySessionLoader method loadSession.

/**
	 * Restore Enrichment maps
	 *
	 * @param pStateFileList - list of files associated with thie session
	 */
@SuppressWarnings("unchecked")
public void loadSession(CySession session) {
    Map<Long, EnrichmentMapParameters> paramsMap = new HashMap<>();
    Map<Long, EnrichmentMap> enrichmentMapMap = new HashMap<>();
    List<File> fileList = session.getAppFileListMap().get(CyActivator.APP_NAME);
    try {
        //go through the prop files first to create the correct objects to be able to add other files to.
        for (File prop_file : fileList) {
            if (prop_file.getName().contains(".props")) {
                InputStream reader = streamUtil.getInputStream(prop_file.getAbsolutePath());
                String fullText = new Scanner(reader, "UTF-8").useDelimiter("\\A").next();
                //Given the file with all the parameters create a new parameter
                EnrichmentMapParameters params = enrichmentMapParametersFactory.create(fullText);
                EnrichmentMap em = new EnrichmentMap(params.getCreationParameters(), serviceRegistrar);
                //get the network name
                String param_name = em.getName();
                //TODO:distinguish between GSEA and EM saved sessions
                String props_name = (prop_file.getName().split("\\."))[0];
                String networkName = param_name;
                //related to bug ticket #49
                if (!props_name.equalsIgnoreCase(param_name))
                    networkName = props_name;
                //after associated the properties with the network
                //initialized each Dataset that we have files for
                HashMap<String, DataSetFiles> files = params.getFiles();
                for (Iterator<String> j = params.getFiles().keySet().iterator(); j.hasNext(); ) {
                    String current_dataset = j.next();
                    Method method = EnrichmentMapParameters.stringToMethod(params.getMethod());
                    em.createDataSet(current_dataset, method, files.get(current_dataset));
                }
                CyNetwork network = getNetworkByName(networkName);
                Long suid = network.getSUID();
                em.setNetworkID(suid);
                paramsMap.put(suid, params);
                enrichmentMapMap.put(suid, em);
            }
        }
        // go through the rest of the files
        for (File propFile : fileList) {
            FileNameParts parts = ParseFileName(propFile);
            if (parts == null || propFile.getName().contains(".props"))
                continue;
            CyNetwork net = getNetworkByName(parts.name);
            EnrichmentMap em = net == null ? null : enrichmentMapMap.get(net.getSUID());
            EnrichmentMapParameters params = paramsMap.get(net.getSUID());
            Method method = EnrichmentMapParameters.stringToMethod(params.getMethod());
            if (em == null) {
                System.out.println("network for file" + propFile.getName() + " does not exist.");
            } else if ((!propFile.getName().contains(".props")) && (!propFile.getName().contains(".expression1.txt")) && (!propFile.getName().contains(".expression2.txt"))) {
                HashMap<String, String> props = params.getProps();
                //if this a dataset specific file make sure there is a dataset object for it
                if (!(parts.dataset == null) && em.getDataSet(parts.dataset) == null && !parts.dataset.equalsIgnoreCase("signature"))
                    em.createDataSet(parts.dataset, method, params.getFiles().get(parts.dataset));
                if (parts.type == null)
                    System.out.println("Sorry, unable to determine the type of the file: " + propFile.getName());
                //read the file
                InputStream reader = streamUtil.getInputStream(propFile.getAbsolutePath());
                String fullText = new Scanner(reader, "UTF-8").useDelimiter("\\A").next();
                //if the file is empty then skip it
                if (fullText == null || fullText.equalsIgnoreCase(""))
                    continue;
                if (propFile.getName().contains(".gmt")) {
                    HashMap<String, GeneSet> gsMap = (HashMap<String, GeneSet>) params.repopulateHashmap(fullText, 1);
                    if (propFile.getName().contains(".signature.gmt")) {
                        // TODO Find a better way to serialize EMSignatureDataSet
                        String sdsName = propFile.getName().replace(".signature.gmt", "");
                        sdsName = NamingUtil.getUniqueName(sdsName, em.getSignatureDataSets().keySet());
                        EMSignatureDataSet sigDataSet = new EMSignatureDataSet(sdsName);
                        em.addSignatureDataSet(sigDataSet);
                        SetOfGeneSets sigGeneSets = sigDataSet.getGeneSetsOfInterest();
                        gsMap.forEach((k, v) -> sigGeneSets.addGeneSet(k, v));
                    } else if (propFile.getName().contains(".set2.gmt")) {
                        // account for legacy session files
                        if (em.getAllGeneSets().containsKey(LegacySupport.DATASET2)) {
                            SetOfGeneSets gs = new SetOfGeneSets(LegacySupport.DATASET2, props);
                            gs.setGeneSets(gsMap);
                        }
                    } else {
                        SetOfGeneSets gs = new SetOfGeneSets(parts.dataset, props);
                        gs.setGeneSets(gsMap);
                        em.getDataSets().get(parts.dataset).setSetOfGeneSets(gs);
                    }
                }
                if (propFile.getName().contains(".genes.txt")) {
                    HashMap<String, Integer> genes = params.repopulateHashmap(fullText, 2);
                    genes.forEach(em::addGene);
                    //ticket #188 - unable to open session files that have empty enrichment maps.
                    if (genes != null && !genes.isEmpty())
                        // Ticket #107 : restore also gene count (needed to determine the next free hash in case we do PostAnalysis with a restored session)
                        em.setNumberOfGenes(Math.max(em.getNumberOfGenes(), Collections.max(genes.values()) + 1));
                }
                if (propFile.getName().contains(".hashkey2genes.txt")) {
                    HashMap<Integer, String> hashkey2gene = params.repopulateHashmap(fullText, 5);
                    //ticket #188 - unable to open session files that have empty enrichment maps.
                    if (hashkey2gene != null && !hashkey2gene.isEmpty())
                        // Ticket #107 : restore also gene count (needed to determine the next free hash in case we do PostAnalysis with a restored session)
                        em.setNumberOfGenes(Math.max(em.getNumberOfGenes(), Collections.max(hashkey2gene.keySet()) + 1));
                }
                if ((parts.type != null && (parts.type.equalsIgnoreCase("ENR") || (parts.type.equalsIgnoreCase("SubENR")))) || propFile.getName().contains(".ENR1.txt") || propFile.getName().contains(".SubENR1.txt")) {
                    SetOfEnrichmentResults enrichments;
                    int temp = 1;
                    //check to see if this dataset has enrichment results already
                    if (parts.dataset != null && em.getDataSet(parts.dataset).getEnrichments() != null) {
                        enrichments = em.getDataSet(parts.dataset).getEnrichments();
                    } else if (parts.dataset == null) {
                        enrichments = em.getDataSet(LegacySupport.DATASET1).getEnrichments();
                    /*enrichments = new SetOfEnrichmentResults(EnrichmentMap.DATASET1,props);
                			em.getDataset(EnrichmentMap.DATASET1).setEnrichments(enrichments);*/
                    } else {
                        enrichments = new SetOfEnrichmentResults(parts.dataset, props);
                        em.getDataSet(parts.dataset).setEnrichments(enrichments);
                    }
                    if (parts.type.equalsIgnoreCase("ENR") || propFile.getName().contains(".ENR1.txt")) {
                        if (params.getMethod().equalsIgnoreCase(EnrichmentMapParameters.method_GSEA))
                            enrichments.setEnrichments(params.repopulateHashmap(fullText, 3));
                        else
                            enrichments.setEnrichments(params.repopulateHashmap(fullText, 4));
                    }
                }
                //it would only happen for sessions saved with version 0.8
                if (propFile.getName().contains(".RANKS1.txt") || propFile.getName().contains(".RANKS1Genes.txt")) {
                    Ranking new_ranking;
                    //Check to see if there is already GSEARanking
                    if (em.getDataSet(LegacySupport.DATASET1).getExpressionSets().getAllRanksNames().contains(Ranking.GSEARanking)) {
                        new_ranking = em.getDataSet(LegacySupport.DATASET1).getExpressionSets().getRanksByName(Ranking.GSEARanking);
                    } else {
                        new_ranking = new Ranking();
                        em.getDataSet(LegacySupport.DATASET1).getExpressionSets().addRanks(Ranking.GSEARanking, new_ranking);
                    }
                    if (propFile.getName().contains(".RANKS1.txt")) {
                        Map<Integer, Rank> ranks = (Map<Integer, Rank>) params.repopulateHashmap(fullText, 7);
                        ranks.forEach(new_ranking::addRank);
                    }
                //						if(prop_file.getName().contains(".RANKS1Genes.txt"))
                //							new_ranking.setRank2gene(em.getParams().repopulateHashmap(fullText,7));
                //						if(prop_file.getName().contains(".RANKS1.txt"))
                //							new_ranking.setRanking(em.getParams().repopulateHashmap(fullText,6));
                }
                if (propFile.getName().contains(".RANKS.txt")) {
                    if (parts.ranks_name == null) {
                        //we need to get the name of this set of rankings
                        // network_name.ranking_name.ranks.txt --> split by "." and get 2
                        String[] file_name_tokens = (propFile.getName()).split("\\.");
                        if ((file_name_tokens.length == 4) && (file_name_tokens[1].equals("Dataset 1 Ranking") || file_name_tokens[1].equals("Dataset 2 Ranking")) || (propFile.getName().contains(Ranking.GSEARanking)))
                            parts.ranks_name = Ranking.GSEARanking;
                        else //this is an extra rank file for backwards compatability.  Ignore it.
                        if ((file_name_tokens.length == 4) && (file_name_tokens[1].equals("Dataset 1") || file_name_tokens[1].equals("Dataset 2")) && file_name_tokens[2].equals("RANKS"))
                            continue;
                        else
                            //file name is not structured properly --> default to file name
                            parts.ranks_name = propFile.getName();
                    }
                    Ranking new_ranking = new Ranking();
                    Map<Integer, Rank> ranks = (Map<Integer, Rank>) params.repopulateHashmap(fullText, 6);
                    ranks.forEach(new_ranking::addRank);
                    if (parts.dataset != null)
                        em.getDataSet(parts.dataset).getExpressionSets().addRanks(parts.ranks_name, new_ranking);
                    else
                        em.getDataSet(LegacySupport.DATASET1).getExpressionSets().addRanks(parts.ranks_name, new_ranking);
                }
                //Deal with legacy issues                    
                if (params.isTwoDatasets()) {
                    //make sure there is a Dataset2
                    if (!em.getDataSets().containsKey(LegacySupport.DATASET2))
                        em.createDataSet(LegacySupport.DATASET2, method, new DataSetFiles());
                    if (propFile.getName().contains(".ENR2.txt") || propFile.getName().contains(".SubENR2.txt")) {
                        SetOfEnrichmentResults enrichments;
                        //check to see if this dataset has enrichment results already
                        if (em.getDataSet(LegacySupport.DATASET2).getEnrichments() != null) {
                            enrichments = em.getDataSet(LegacySupport.DATASET2).getEnrichments();
                        } else {
                            enrichments = new SetOfEnrichmentResults(LegacySupport.DATASET2, props);
                            em.getDataSet(LegacySupport.DATASET2).setEnrichments(enrichments);
                        }
                        if (propFile.getName().contains(".ENR2.txt")) {
                            if (params.getMethod().equalsIgnoreCase(EnrichmentMapParameters.method_GSEA))
                                enrichments.setEnrichments(params.repopulateHashmap(fullText, 3));
                            else
                                enrichments.setEnrichments(params.repopulateHashmap(fullText, 4));
                        }
                    }
                    //it would only happen for sessions saved with version 0.8
                    if (propFile.getName().contains(".RANKS2.txt") || propFile.getName().contains(".RANKS2Genes.txt")) {
                        Ranking new_ranking;
                        // Check to see if there is already GSEARanking
                        if (em.getDataSet(LegacySupport.DATASET2).getExpressionSets().getAllRanksNames().contains(Ranking.GSEARanking)) {
                            new_ranking = em.getDataSet(LegacySupport.DATASET2).getExpressionSets().getRanksByName(Ranking.GSEARanking);
                        } else {
                            new_ranking = new Ranking();
                            em.getDataSet(LegacySupport.DATASET2).getExpressionSets().addRanks(Ranking.GSEARanking, new_ranking);
                        }
                        if (propFile.getName().contains(".RANKS2.txt")) {
                            Map<Integer, Rank> ranks = (Map<Integer, Rank>) params.repopulateHashmap(fullText, 6);
                            ranks.forEach(new_ranking::addRank);
                        }
                    }
                }
            }
        }
        //info from the parameters
        for (int i = 0; i < fileList.size(); i++) {
            File prop_file = fileList.get(i);
            FileNameParts parts_exp = ParseFileName(prop_file);
            //unrecognized file
            if ((parts_exp == null) || (parts_exp.name == null))
                continue;
            CyNetwork net = getNetworkByName(parts_exp.name);
            EnrichmentMap map = net == null ? null : enrichmentMapMap.get(net.getSUID());
            EnrichmentMapParameters params = paramsMap.get(net.getSUID());
            Map<String, String> props = params.getProps();
            if (parts_exp.type != null && parts_exp.type.equalsIgnoreCase("expression")) {
                if (map.getDataSets().containsKey(parts_exp.dataset)) {
                    EMDataSet ds = map.getDataSet(parts_exp.dataset);
                    ds.getDataSetFiles().setExpressionFileName(prop_file.getAbsolutePath());
                    ds.getExpressionSets().setFilename(prop_file.getAbsolutePath());
                    ExpressionFileReaderTask expressionFile1 = new ExpressionFileReaderTask(ds);
                    GeneExpressionMatrix matrix = expressionFile1.parse();
                    matrix.restoreProps(parts_exp.dataset, props);
                }
            }
            //Deal with legacy session files.
            if (prop_file.getName().contains("expression1.txt")) {
                EMDataSet ds1 = map.getDataSet(LegacySupport.DATASET1);
                ds1.getDataSetFiles().setExpressionFileName(prop_file.getAbsolutePath());
                ds1.getExpressionSets().setFilename(prop_file.getAbsolutePath());
                ExpressionFileReaderTask expressionFile1 = new ExpressionFileReaderTask(ds1);
                expressionFile1.parse();
            }
            if (prop_file.getName().contains("expression2.txt")) {
                EMDataSet ds2 = map.getDataSet(LegacySupport.DATASET2);
                ds2.getDataSetFiles().setExpressionFileName(prop_file.getAbsolutePath());
                ds2.getExpressionSets().setFilename(prop_file.getAbsolutePath());
                ExpressionFileReaderTask expressionFile2 = new ExpressionFileReaderTask(ds2);
                expressionFile2.parse();
                //are dealing with two distinct expression files.
                if (map.getDataSet(LegacySupport.DATASET2) != null && map.getDataSet(LegacySupport.DATASET2).getGeneSetsOfInterest() != null && !map.getDataSet(LegacySupport.DATASET2).getGeneSetsOfInterest().getGeneSets().isEmpty()) {
                    map.setDistinctExpressionSets(true);
                    map.getDataSet(LegacySupport.DATASET1).setDataSetGenes(new HashSet<Integer>((Set<Integer>) map.getDataSet(LegacySupport.DATASET1).getExpressionSets().getGeneIds()));
                    map.getDataSet(LegacySupport.DATASET2).setDataSetGenes(new HashSet<Integer>((Set<Integer>) map.getDataSet(LegacySupport.DATASET2).getExpressionSets().getGeneIds()));
                }
            }
        }
        //iterate over the networks
        for (Iterator<Long> j = enrichmentMapMap.keySet().iterator(); j.hasNext(); ) {
            Long id = j.next();
            EnrichmentMap map = enrichmentMapMap.get(id);
            //only initialize objects if there is a map for this network
            if (map != null) {
                if (map.getDataSets().size() > 1) {
                    Set<Integer> dataset1_genes = map.getDataSets().get(LegacySupport.DATASET1).getDataSetGenes();
                    Set<Integer> dataset2_genes = map.getDataSets().get(LegacySupport.DATASET2).getDataSetGenes();
                    if (!dataset1_genes.equals(dataset2_genes))
                        map.setDistinctExpressionSets(true);
                }
                //initialize the Genesets (makes sure the leading edge is set correctly)
                //Initialize the set of genesets and GSEA results that we want to compute over
                InitializeGenesetsOfInterestTask genesets_init = new InitializeGenesetsOfInterestTask(map);
                // MKTODO really?
                genesets_init.setThrowIfMissing(false);
                genesets_init.initializeSets(null);
            //					//for each map compute the similarity matrix, (easier than storing it) compute the geneset similarities
            //					ComputeSimilarityTask similarities = new ComputeSimilarityTask(map, ComputeSimilarityTask.ENRICHMENT);
            //					Map<String, GenesetSimilarity> similarity_results = similarities.computeGenesetSimilarities(null);
            //					map.setGenesetSimilarity(similarity_results);
            //
            //					// also compute geneset similarities between Enrichment- and Signature Genesets (if any)
            //					if (! map.getSignatureGenesets().isEmpty()){
            //						ComputeSimilarityTask sigSimilarities = new ComputeSimilarityTask(map, ComputeSimilarityTask.SIGNATURE);
            //						Map<String, GenesetSimilarity> sig_similarity_results = sigSimilarities.computeGenesetSimilarities(null);
            //						map.getGenesetSimilarity().putAll(sig_similarity_results);
            //					}
            }
        //end of if(map != null)
        }
        for (Iterator<Long> j = enrichmentMapMap.keySet().iterator(); j.hasNext(); ) {
            Long id = j.next();
            CyNetwork currentNetwork = cyNetworkManager.getNetwork(id);
            EnrichmentMap map = enrichmentMapMap.get(id);
            map.setLegacy(true);
            emManager.registerEnrichmentMap(map);
            if (!j.hasNext()) {
                //set the last network to be the one viewed and initialize the parameters panel
                cyApplicationManager.setCurrentNetwork(currentNetwork);
            }
        }
    } catch (Exception ee) {
        ee.printStackTrace();
    }
}

Example 2 with GeneSet

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

the class ParseGREATEnrichmentResults method parseLines.

@Override
public void parseLines(List<String> lines, EMDataSet dataset, TaskMonitor taskMonitor) {
    if (taskMonitor == null)
        taskMonitor = new NullTaskMonitor();
    taskMonitor.setTitle("Parsing Enrichment Result file");
    boolean hasBackground = false;
    EMCreationParameters params = dataset.getMap().getParams();
    //Get the type of filter user specified on the GREAT results
    //If it is hyper use column 14 Hypergeometric p-value and 16 FDR for hyper
    //If it is binom use column 5 bionomial p-value and 7 FDR for binom
    //If they specify both use the highest p-value and q-value from the above columns
    GreatFilter filterType = dataset.getMap().getParams().getGreatFilter();
    Map<String, GeneSet> genesets = dataset.getSetOfGeneSets().getGeneSets();
    EnrichmentMap map = dataset.getMap();
    Map<String, EnrichmentResult> results = dataset.getEnrichments().getEnrichments();
    int currentProgress = 0;
    int maxValue = lines.size();
    taskMonitor.setStatusMessage("Parsing Great Results file - " + maxValue + " rows");
    //for great files there is an FDR
    dataset.getMap().getParams().setFDR(true);
    //skip the first l9 which just has the field names (start i=1)
    //check to see how many columns the data has
    //go through each line until we find the header line
    int k = 0;
    String line = lines.get(k);
    String[] tokens = line.split("\t");
    for (; k < lines.size(); k++) {
        line = lines.get(k);
        tokens = line.split("\t");
        int length = tokens.length;
        if ((length == 24) && tokens[3].equalsIgnoreCase("BinomRank")) {
            break;
        } else //There is not binom rank and no binomial data.
        if ((length == 20) && tokens[3].equalsIgnoreCase("Rank")) {
            hasBackground = true;
            break;
        }
    }
    //go through the rest of the lines
    for (int i = k + 1; i < lines.size(); i++) {
        line = lines.get(i);
        tokens = line.split("\t");
        //there are extra lines at the end of the file that should be ignored.
        if (!hasBackground && tokens.length != 24)
            continue;
        if (hasBackground && tokens.length != 20)
            continue;
        double pvalue = 1.0;
        double FDRqvalue = 1.0;
        GenericResult result;
        int gs_size = 0;
        double NES = 1.0;
        //details of export file
        //http://bejerano.stanford.edu/help/display/GREAT/Export
        //The second column of the file is the name of the geneset
        final String name = tokens[1].trim() + "-" + tokens[2].trim();
        //the first column of the file is the description
        final String description = tokens[2].trim();
        //when there are two different species it is possible that the gene set could
        //already exist in the set of genesets.  if it does exist then add the genes
        //in this set to the geneset
        ImmutableSet.Builder<Integer> builder = ImmutableSet.builder();
        if (genesets.containsKey(name))
            builder = builder.addAll(genesets.get(name).getGenes());
        String[] gene_tokens;
        if (!hasBackground)
            gene_tokens = tokens[23].split(",");
        else
            gene_tokens = tokens[18].split(",");
        //All subsequent fields in the list are the geneset associated with this geneset.
        for (int j = 0; j < gene_tokens.length; j++) {
            String gene = gene_tokens[j].toUpperCase();
            //if it is already in the hash then get its associated key and put it into the set of genes
            if (map.containsGene(gene)) {
                builder.add(map.getHashFromGene(gene));
            } else if (!gene.isEmpty()) {
                Integer hash = map.addGene(gene).get();
                builder.add(hash);
            }
        }
        //finished parsing that geneset
        //add the current geneset to the hashmap of genesets
        GeneSet gs = new GeneSet(name, description, builder.build());
        genesets.put(name, gs);
        //There are two tests run by GREAT, the binomial on regions and the hypergeometric based on genes
        //The first pass of results shows only those that are significant both
        //The user can then choose to use either or both together
        //
        //If it is hyper use column 14 Hypergeometric p-value and 16 FDR for hyper
        //If it is binom use column 5 bionomial p-value and 7 FDR for binom
        //If they specify both use the highest p-value and q-value from the above columns
        double hyper_pvalue = 1;
        double hyper_fdr = 1;
        double binom_pvalue = 1;
        double binom_fdr = 1;
        if (!hasBackground) {
            if (!tokens[4].equalsIgnoreCase(""))
                binom_pvalue = Double.parseDouble(tokens[4]);
            if (!tokens[6].equalsIgnoreCase(""))
                binom_fdr = Double.parseDouble(tokens[6]);
            if (!tokens[13].equalsIgnoreCase(""))
                hyper_pvalue = Double.parseDouble(tokens[13]);
            if (!tokens[15].equalsIgnoreCase(""))
                hyper_fdr = Double.parseDouble(tokens[15]);
        } else {
            if (!tokens[4].equalsIgnoreCase(""))
                hyper_pvalue = Double.parseDouble(tokens[4]);
            if (!tokens[6].equalsIgnoreCase(""))
                hyper_fdr = Double.parseDouble(tokens[6]);
        }
        if (filterType == GreatFilter.HYPER) {
            pvalue = hyper_pvalue;
            FDRqvalue = hyper_fdr;
        } else if (filterType == GreatFilter.BINOM) {
            pvalue = binom_pvalue;
            FDRqvalue = binom_fdr;
        } else if (filterType == GreatFilter.BOTH) {
            pvalue = Math.max(hyper_pvalue, binom_pvalue);
            FDRqvalue = Math.max(hyper_fdr, binom_fdr);
        } else if (filterType == GreatFilter.EITHER) {
            pvalue = Math.min(hyper_pvalue, binom_pvalue);
            FDRqvalue = Math.min(hyper_fdr, binom_fdr);
        } else {
            System.out.println("Invalid attribute setting for GREAT p-value specification");
        }
        //Keep track of minimum p-value to better calculate jslider
        if (pvalue < params.getPvalueMin())
            params.setPvalueMin(pvalue);
        if (FDRqvalue < params.getQvalueMin())
            params.setQvalueMin(FDRqvalue);
        //If this is a background set then it is in the 16th column
        if ((!hasBackground) && (!tokens[19].equalsIgnoreCase("")))
            gs_size = Integer.parseInt(tokens[19]);
        else if ((hasBackground) && (!tokens[15].equalsIgnoreCase("")))
            gs_size = Integer.parseInt(tokens[15]);
        result = new GenericResult(name, description, pvalue, gs_size, FDRqvalue);
        // Calculate Percentage.  This must be a value between 0..100.
        int percentComplete = (int) (((double) currentProgress / maxValue) * 100);
        taskMonitor.setProgress(percentComplete);
        currentProgress++;
        //check to see if the gene set has already been entered in the results
        //it is possible that one geneset will be in both phenotypes.
        //if it is already exists then we want to make sure the one retained is the result with the
        //lower p-value.
        //ticket #149
        GenericResult temp = (GenericResult) results.get(name);
        if (temp == null)
            results.put(name, result);
        else {
            if (result.getPvalue() < temp.getPvalue())
                results.put(name, result);
        }
    }
}

Example 3 with GeneSet

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

the class ParseGenericEnrichmentResults method parseLines.

@Override
public void parseLines(List<String> lines, EMDataSet dataset, TaskMonitor taskMonitor) {
    if (taskMonitor == null)
        taskMonitor = new NullTaskMonitor();
    taskMonitor.setTitle("Parsing Generic Result file");
    //Get the current genesets so we can check that all the results are in the geneset list
    //and put the size of the genesets into the visual style
    Map<String, GeneSet> genesets = dataset.getSetOfGeneSets().getGeneSets();
    int currentProgress = 0;
    int maxValue = lines.size();
    taskMonitor.setStatusMessage("Parsing Generic Results file - " + maxValue + " rows");
    boolean FDR = false;
    //skip the first line which just has the field names (start i=1)
    //check to see how many columns the data has
    String line = lines.get(0);
    String[] tokens = line.split("\t");
    int length = tokens.length;
    EnrichmentMap map = dataset.getMap();
    SetOfEnrichmentResults enrichments = dataset.getEnrichments();
    Map<String, EnrichmentResult> results = enrichments.getEnrichments();
    String upPhenotype = enrichments.getPhenotype1();
    String downPhenotype = enrichments.getPhenotype2();
    //check to see if there are genesets.
    //if there are no genesets then populate the genesets from the generic file
    //can only do this if the 6th column has a list of genes for that geneset.
    boolean populate_gs = false;
    if (genesets == null || genesets.isEmpty())
        populate_gs = true;
    else
        //as this is the default for gprofiler use the Description in the visual style instead of the formatted name
        //but only if there is a gmt supplied.  If using just the generic output file there is not field for description
        dataset.getMap().getParams().setEMgmt(true);
    for (int i = 1; i < lines.size(); i++) {
        line = lines.get(i);
        tokens = line.split("\t");
        //update the length each time because some line might have missing values
        length = tokens.length;
        double pvalue = 1.0;
        double FDRqvalue = 1.0;
        GenericResult result;
        int gs_size = 0;
        double NES = 1.0;
        //The first column of the file is the name of the geneset
        final String name = tokens[0].toUpperCase().trim();
        final String description = tokens[1].toUpperCase();
        if (genesets.containsKey(name)) {
            gs_size = genesets.get(name).getGenes().size();
        }
        //The third column is the nominal p-value
        if (tokens[2] == null || tokens[2].equalsIgnoreCase("")) {
        //do nothing
        } else {
            pvalue = Double.parseDouble(tokens[2]);
        }
        if (length > 3) {
            //the fourth column is the FDR q-value
            if (tokens[3] == null || tokens[3].equalsIgnoreCase("")) {
            //do nothing
            } else {
                FDRqvalue = Double.parseDouble(tokens[3]);
                FDR = true;
            }
            // and if it is a number the only important part is the sign
            if (length > 4) {
                if (tokens[4] == null || tokens[4].equalsIgnoreCase("")) {
                } else {
                    //check to see if the string matches the specified phenotypes
                    if (tokens[4].equalsIgnoreCase(upPhenotype))
                        NES = 1.0;
                    else if (tokens[4].equalsIgnoreCase(downPhenotype))
                        NES = -1.0;
                    else //try and see if the user has specified the phenotype as a number
                    {
                        try {
                            NES = Double.parseDouble(tokens[4]);
                        } catch (NumberFormatException nfe) {
                            throw new IllegalThreadStateException(tokens[4] + " is not a valid phenotype.  Phenotype specified in generic enrichment results file must have the same phenotype as specified in advanced options or must be a positive or negative number.");
                        }
                    }
                }
                //its enrichment
                if (length > 5 && populate_gs) {
                    //get all the genes in the field
                    String[] gene_tokens = tokens[5].split(",");
                    ImmutableSet.Builder<Integer> builder = ImmutableSet.builder();
                    //All subsequent fields in the list are the geneset associated with this geneset.
                    for (String token : gene_tokens) {
                        String gene = token.trim().toUpperCase();
                        //if it is already in the hash then get its associated key and put it into the set of genes
                        if (map.containsGene(gene)) {
                            builder.add(map.getHashFromGene(gene));
                        } else if (!gene.isEmpty()) {
                            Integer hash = map.addGene(gene).get();
                            builder.add(hash);
                        }
                    }
                    GeneSet gs = new GeneSet(name, description, builder.build());
                    gs_size = gs.getGenes().size();
                    //put the new or filtered geneset back into the set.
                    genesets.put(name, gs);
                }
                //end of tokens>5
                result = new GenericResult(name, description, pvalue, gs_size, FDRqvalue, NES);
            } else
                //end of tokens>4
                result = new GenericResult(name, description, pvalue, gs_size, FDRqvalue);
        } else {
            result = new GenericResult(name, description, pvalue, gs_size);
        }
        // Calculate Percentage.  This must be a value between 0..100.
        int percentComplete = (int) (((double) currentProgress / maxValue) * 100);
        taskMonitor.setProgress(percentComplete);
        currentProgress++;
        //check to see if the gene set has already been entered in the results
        //it is possible that one geneset will be in both phenotypes.
        //if it is already exists then we want to make sure the one retained is the result with the
        //lower p-value.
        //ticket #149
        GenericResult temp = (GenericResult) results.get(name);
        if (temp == null)
            results.put(name, result);
        else {
            if (result.getPvalue() < temp.getPvalue())
                results.put(name, result);
        }
    }
    if (FDR)
        dataset.getMap().getParams().setFDR(FDR);
}

Example 4 with GeneSet

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

the class HeatMapRanksTest method testLeadingEdge.

@Test
public void testLeadingEdge(EnrichmentMapManager emManager) throws Exception {
    final String geneSetName = "ENVELOPE%GO%GO:0031975";
    final int leadingEdgeSize = 170;
    // Sanity test
    EnrichmentMap map = emManager.getAllEnrichmentMaps().values().iterator().next();
    EMDataSet dataset = map.getDataSet(LegacySupport.DATASET1);
    GeneSet gs = dataset.getGeneSetsOfInterest().getGeneSets().get(geneSetName);
    assertNotNull(gs);
    // Run the ranking
    RankingOption rankingOption = new GSEALeadingEdgeRankingOption(dataset, geneSetName, Ranking.GSEARanking);
    Map<Integer, RankValue> ranks = rankingOption.computeRanking(gs.getGenes()).get().get();
    assertEquals(454, ranks.size());
    // Convert to useful collections
    Map<RankValue, Integer> rankToGeneId = HashBiMap.create(ranks).inverse();
    List<RankValue> sortedRanks = ranks.values().stream().sorted().collect(Collectors.toList());
    // Test leading edge
    for (int i = 0; i < sortedRanks.size(); i++) {
        RankValue v = sortedRanks.get(i);
        assertTrue(v.isSignificant() == i < leadingEdgeSize);
    }
    // Test genes are the same
    List<Integer> expectedGeneOrder = getGeneOrderFromFile(map, PATH + "gene_order_leading_edge.txt");
    List<Integer> actualGeneOrder = sortedRanks.stream().map(rankToGeneId::get).collect(Collectors.toList());
    assertEquals(expectedGeneOrder, actualGeneOrder);
}

Example 5 with GeneSet

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

the class LegacySessionLoadTest method test_1_LoadedLegacyData.

@Test
@SessionFile("em_session_2.2.cys")
public void test_1_LoadedLegacyData() throws Exception {
    EnrichmentMap map = getEnrichmentMap();
    assertEquals("EM1_Enrichment Map", map.getName());
    CyNetwork network = networkManager.getNetwork(map.getNetworkID());
    assertNotNull(network);
    assertEquals(1, map.getDataSetCount());
    assertEquals(14067, map.getNumberOfGenes());
    assertEquals(14067, map.getAllGenes().size());
    // Number of edges: 3339 - that's how many geneset similarity objects there should be!!!
    CyTable edgeTable = network.getDefaultEdgeTable();
    assertEquals(3339, edgeTable.getRowCount());
    EMCreationParameters params = map.getParams();
    String prefix = params.getAttributePrefix();
    assertEquals("EM1_", prefix);
    assertEquals(0.5, params.getCombinedConstant(), 0.0);
    assertFalse(params.isEMgmt());
    assertEquals("Geneset_Overlap", params.getEnrichmentEdgeType());
    assertTrue(params.isFDR());
    assertEquals(GreatFilter.HYPER, params.getGreatFilter());
    assertEquals(0.005, params.getPvalue(), 0.0);
    assertEquals(1.0, params.getPvalueMin(), 0.0);
    assertEquals(0.1, params.getQvalue(), 0.0);
    assertEquals(1.0, params.getQvalueMin(), 0.0);
    assertEquals(0.5, params.getSimilarityCutoff(), 0.0);
    assertEquals(SimilarityMetric.OVERLAP, params.getSimilarityMetric());
    //		assertFalse(params.isDistinctExpressionSets());
    String geneset1 = "RESOLUTION OF SISTER CHROMATID COHESION%REACTOME%REACT_150425.2";
    String geneset2 = "CHROMOSOME, CENTROMERIC REGION%GO%GO:0000775";
    Collection<CyRow> rows = edgeTable.getMatchingRows(CyNetwork.NAME, geneset1 + " (Geneset_Overlap) " + geneset2);
    assertEquals(1, rows.size());
    CyRow row = rows.iterator().next();
    assertEquals("Geneset_Overlap", row.get(CyEdge.INTERACTION, String.class));
    assertEquals(0.6097560975609756, EMStyleBuilder.Columns.EDGE_SIMILARITY_COEFF.get(row, prefix), 0.0);
    EMDataSet dataset = map.getDataSet("Dataset 1");
    assertNotNull(dataset);
    assertSame(map, dataset.getMap());
    assertEquals(Method.GSEA, dataset.getMethod());
    assertEquals(12653, dataset.getDataSetGenes().size());
    assertEquals(389, dataset.getGeneSetsOfInterest().getGeneSets().size());
    //		assertEquals(17259, dataset.getSetofgenesets().getGenesets().size()); // MKTODO why? what is this used for
    assertEndsWith(dataset.getSetOfGeneSets().getFilename(), "Human_GO_AllPathways_no_GO_iea_April_15_2013_symbol.gmt");
    for (long suid : dataset.getNodeSuids()) {
        assertNotNull(network.getNode(suid));
    }
    GeneSet geneset = dataset.getGeneSetsOfInterest().getGeneSets().get("NCRNA PROCESSING%GO%GO:0034470");
    assertEquals(88, geneset.getGenes().size());
    assertEquals("NCRNA PROCESSING%GO%GO:0034470", geneset.getName());
    assertEquals("ncRNA processing", geneset.getDescription());
    assertEquals(Optional.of("GO"), geneset.getSource());
    SetOfEnrichmentResults enrichments = dataset.getEnrichments();
    assertEquals(4756, enrichments.getEnrichments().size());
    assertEndsWith(enrichments.getFilename1(), "gsea_report_for_ES12_1473194913081.xls");
    assertEndsWith(enrichments.getFilename2(), "gsea_report_for_NT12_1473194913081.xls");
    assertEquals("ES12", enrichments.getPhenotype1());
    assertEquals("NT12", enrichments.getPhenotype2());
    EnrichmentResult result = enrichments.getEnrichments().get("RIBONUCLEOSIDE TRIPHOSPHATE BIOSYNTHETIC PROCESS%GO%GO:0009201");
    assertTrue(result instanceof GSEAResult);
    GSEAResult gseaResult = (GSEAResult) result;
    assertEquals("RIBONUCLEOSIDE TRIPHOSPHATE BIOSYNTHETIC PROCESS%GO%GO:0009201", gseaResult.getName());
    assertEquals(0.42844063, gseaResult.getES(), 0.0);
    assertEquals(0.45225498, gseaResult.getFdrqvalue(), 0.0);
    assertEquals(1.0, gseaResult.getFwerqvalue(), 0.0);
    assertEquals(23, gseaResult.getGsSize());
    assertEquals(1.1938541, gseaResult.getNES(), 0.0);
    assertEquals(0.2457786, gseaResult.getPvalue(), 0.0);
    assertEquals(4689, gseaResult.getRankAtMax());
    assertEquals(Optional.of("GO"), gseaResult.getSource());
    GeneExpressionMatrix expressions = dataset.getExpressionSets();
    assertEquals(20326, expressions.getExpressionUniverse());
    assertEquals(3.686190609, expressions.getClosesttoZero(), 0.0);
    //		assertEndsWith(expressions.getFilename(), "MCF7_ExprMx_v2_names.gct");
    assertEquals(15380.42388, expressions.getMaxExpression(), 0.0);
    assertEquals(3.686190609, expressions.getMinExpression(), 0.0);
    assertEquals(20, expressions.getNumConditions());
    assertEquals(12653, expressions.getExpressionMatrix().size());
    assertEquals(12653, expressions.getExpressionMatrix_rowNormalized().size());
    GeneExpression expression = expressions.getExpressionMatrix().get(0);
    assertEquals("MOCOS", expression.getName());
    assertEquals("MOCOS (molybdenum cofactor sulfurase)", expression.getDescription());
    assertEquals(18, expression.getExpression().length);
    Ranking ranking = expressions.getRanks().get("GSEARanking");
    assertEquals(12653, ranking.getAllRanks().size());
    assertEquals(12653, ranking.getRanking().size());
    Rank rank = ranking.getRanking().get(0);
    assertEquals("MOCOS", rank.getName());
    assertEquals(1238, rank.getRank().intValue());
    assertEquals(0.54488367, rank.getScore(), 0.0);
    DataSetFiles files = dataset.getDataSetFiles();
    assertEndsWith(files.getClassFile(), "ES_NT.cls");
    assertEndsWith(files.getEnrichmentFileName1(), "gsea_report_for_ES12_1473194913081.xls");
    assertEndsWith(files.getEnrichmentFileName2(), "gsea_report_for_NT12_1473194913081.xls");
    //		assertEndsWith(files.getExpressionFileName(), "MCF7_ExprMx_v2_names.gct");
    assertEndsWith(files.getGMTFileName(), "Human_GO_AllPathways_no_GO_iea_April_15_2013_symbol.gmt");
    assertEndsWith(files.getGseaHtmlReportFile(), "estrogen_treatment_12hr_gsea_enrichment_results.Gsea.1473194913081/index.html");
    assertEndsWith(files.getRankedFile(), "ranked_gene_list_ES12_versus_NT12_1473194913081.xls");
    assertEquals("ES12", files.getPhenotype1());
    assertEquals("NT12", files.getPhenotype2());
}