Search in sources :

Example 11 with GeoDomainObjectGenerator

use of ubic.gemma.core.loader.expression.geo.GeoDomainObjectGenerator in project Gemma by PavlidisLab.

the class AbstractArrayDesignProcessingTest method setupArrayDesign.

@Before
public void setupArrayDesign() {
    ad = arrayDesignService.findByShortName(AbstractArrayDesignProcessingTest.ACCESSION);
    if (ad == null) {
        // first load small twoc-color
        GeoService geoService = this.getBean(GeoService.class);
        geoService.setGeoDomainObjectGenerator(new GeoDomainObjectGenerator());
        try {
            @SuppressWarnings("unchecked") final Collection<ArrayDesign> ads = (Collection<ArrayDesign>) geoService.fetchAndLoad(AbstractArrayDesignProcessingTest.ACCESSION, true, true, false, true, true);
            ad = ads.iterator().next();
        } catch (Exception e) {
            if (e.getCause() instanceof FileNotFoundException) {
                log.warn("problem with initializing array design for test: " + e.getCause().getMessage());
                return;
            }
            throw e;
        }
    }
    ad = arrayDesignService.thaw(ad);
}
Also used : GeoDomainObjectGenerator(ubic.gemma.core.loader.expression.geo.GeoDomainObjectGenerator) ArrayDesign(ubic.gemma.model.expression.arrayDesign.ArrayDesign) GeoService(ubic.gemma.core.loader.expression.geo.service.GeoService) FileNotFoundException(java.io.FileNotFoundException) Collection(java.util.Collection) FileNotFoundException(java.io.FileNotFoundException) Before(org.junit.Before)

Example 12 with GeoDomainObjectGenerator

use of ubic.gemma.core.loader.expression.geo.GeoDomainObjectGenerator in project Gemma by PavlidisLab.

the class DiffExTest method testCountData.

/**
 * Test differential expression analysis on RNA-seq data. See bug 3383. R code in voomtest.R
 */
@Test
public void testCountData() throws Exception {
    geoService.setGeoDomainObjectGenerator(new GeoDomainObjectGenerator());
    ExpressionExperiment ee = eeService.findByShortName("GSE29006");
    if (ee != null) {
        eeService.remove(ee);
    }
    assertTrue(eeService.findByShortName("GSE29006") == null);
    try {
        Collection<?> results = geoService.fetchAndLoad("GSE29006", false, false, false);
        ee = (ExpressionExperiment) results.iterator().next();
    } catch (AlreadyExistsInSystemException e) {
        throw new IllegalStateException("Need to remove this data set before test is run");
    }
    ee = eeService.thaw(ee);
    try (InputStream is = this.getClass().getResourceAsStream("/data/loader/expression/flatfileload/GSE29006_design.txt")) {
        assertNotNull(is);
        experimentalDesignImporter.importDesign(ee, is);
    }
    // Load the data from a text file.
    DoubleMatrixReader reader = new DoubleMatrixReader();
    ArrayDesign targetArrayDesign;
    try (InputStream countData = this.getClass().getResourceAsStream("/data/loader/expression/flatfileload/GSE29006_expression_count.test.txt")) {
        DoubleMatrix<String, String> countMatrix = reader.read(countData);
        Collection<ExperimentalFactor> experimentalFactors = ee.getExperimentalDesign().getExperimentalFactors();
        assertEquals(1, experimentalFactors.size());
        List<String> probeNames = countMatrix.getRowNames();
        assertEquals(199, probeNames.size());
        // we have to find the right generic platform to use.
        targetArrayDesign = this.getTestPersistentArrayDesign(probeNames, taxonService.findByCommonName("human"));
        targetArrayDesign = arrayDesignService.thaw(targetArrayDesign);
        // the experiment has 8 samples but the data has 4 columns so allow missing samples
        // GSM718707 GSM718708 GSM718709 GSM718710
        dataUpdater.addCountData(ee, targetArrayDesign, countMatrix, null, 36, true, true);
    }
    // make sure to do a thawRawAndProcessed() to get the addCountData() updates
    ee = eeService.thaw(ee);
    // verify rows and columns
    Collection<DoubleVectorValueObject> processedDataArrays = processedExpressionDataVectorService.getProcessedDataArrays(ee);
    assertEquals(199, processedDataArrays.size());
    for (DoubleVectorValueObject v : processedDataArrays) {
        assertEquals(4, v.getBioAssays().size());
    }
    // I confirmed that log2cpm is working same as voom here; not bothering to test directly.
    TestUtils.assertBAs(ee, targetArrayDesign, "GSM718709", 320383);
    // DE analysis without weights to assist comparison to R
    DifferentialExpressionAnalysisConfig config = new DifferentialExpressionAnalysisConfig();
    config.setUseWeights(false);
    config.setFactorsToInclude(ee.getExperimentalDesign().getExperimentalFactors());
    Collection<DifferentialExpressionAnalysis> analyses = analyzer.run(ee, config);
    assertNotNull(analyses);
    assertEquals(1, analyses.size());
    DifferentialExpressionAnalysis results = analyses.iterator().next();
    boolean found = false;
    ExpressionAnalysisResultSet resultSet = results.getResultSets().iterator().next();
    for (DifferentialExpressionAnalysisResult r : resultSet.getResults()) {
        if (r.getProbe().getName().equals("ENSG00000000938")) {
            found = true;
            ContrastResult contrast = r.getContrasts().iterator().next();
            assertEquals(0.007055717, r.getPvalue(), // R: 0.006190738; coeff = 2.2695215; t=12.650422; R with our weights: 0.009858270, 2.2317534; t=9.997007
            0.00001);
            // up to sign
            assertEquals(2.2300049, Math.abs(contrast.getCoefficient()), 0.001);
            break;
        }
    }
    assertTrue(found);
    // With weights
    config = new DifferentialExpressionAnalysisConfig();
    // <----
    config.setUseWeights(true);
    config.setFactorsToInclude(ee.getExperimentalDesign().getExperimentalFactors());
    analyses = analyzer.run(ee, config);
    results = analyses.iterator().next();
    resultSet = results.getResultSets().iterator().next();
    for (DifferentialExpressionAnalysisResult r : resultSet.getResults()) {
        if (r.getProbe().getName().equals("ENSG00000000938")) {
            assertEquals(1, r.getContrasts().size());
            ContrastResult contrast = r.getContrasts().iterator().next();
            // yes!
            assertEquals(2.232816, Math.abs(contrast.getCoefficient()), 0.001);
            assertEquals(0.000311, contrast.getPvalue(), 0.00001);
            assertEquals(56.66342, Math.abs(contrast.getTstat()), 0.001);
            assertEquals(0.007068, r.getPvalue(), 0.00001);
            break;
        }
    }
}
Also used : InputStream(java.io.InputStream) ArrayDesign(ubic.gemma.model.expression.arrayDesign.ArrayDesign) ExperimentalFactor(ubic.gemma.model.expression.experiment.ExperimentalFactor) DifferentialExpressionAnalysis(ubic.gemma.model.analysis.expression.diff.DifferentialExpressionAnalysis) DifferentialExpressionAnalysisResult(ubic.gemma.model.analysis.expression.diff.DifferentialExpressionAnalysisResult) ExpressionExperiment(ubic.gemma.model.expression.experiment.ExpressionExperiment) DoubleMatrixReader(ubic.basecode.io.reader.DoubleMatrixReader) ExpressionAnalysisResultSet(ubic.gemma.model.analysis.expression.diff.ExpressionAnalysisResultSet) GeoDomainObjectGenerator(ubic.gemma.core.loader.expression.geo.GeoDomainObjectGenerator) AlreadyExistsInSystemException(ubic.gemma.core.loader.util.AlreadyExistsInSystemException) DoubleVectorValueObject(ubic.gemma.model.expression.bioAssayData.DoubleVectorValueObject) ContrastResult(ubic.gemma.model.analysis.expression.diff.ContrastResult) AbstractGeoServiceTest(ubic.gemma.core.loader.expression.geo.AbstractGeoServiceTest) Test(org.junit.Test)

Aggregations

GeoDomainObjectGenerator (ubic.gemma.core.loader.expression.geo.GeoDomainObjectGenerator)12 Test (org.junit.Test)7 AbstractGeoServiceTest (ubic.gemma.core.loader.expression.geo.AbstractGeoServiceTest)7 ExpressionExperiment (ubic.gemma.model.expression.experiment.ExpressionExperiment)7 AlreadyExistsInSystemException (ubic.gemma.core.loader.util.AlreadyExistsInSystemException)6 ArrayDesign (ubic.gemma.model.expression.arrayDesign.ArrayDesign)6 InputStream (java.io.InputStream)5 DoubleMatrixReader (ubic.basecode.io.reader.DoubleMatrixReader)4 DoubleVectorValueObject (ubic.gemma.model.expression.bioAssayData.DoubleVectorValueObject)3 ExpressionDataDoubleMatrix (ubic.gemma.core.datastructure.matrix.ExpressionDataDoubleMatrix)2 GeoConverter (ubic.gemma.core.loader.expression.geo.GeoConverter)2 GeoService (ubic.gemma.core.loader.expression.geo.service.GeoService)2 BioAssay (ubic.gemma.model.expression.bioAssay.BioAssay)2 BufferedReader (java.io.BufferedReader)1 FileInputStream (java.io.FileInputStream)1 FileNotFoundException (java.io.FileNotFoundException)1 InputStreamReader (java.io.InputStreamReader)1 Collection (java.util.Collection)1 Before (org.junit.Before)1 PreprocessingException (ubic.gemma.core.analysis.preprocess.PreprocessingException)1