/**
* @throws InvalidColumnIndexException
* @throws IOException
*/
public void executeLimma() {
Limma limma = null;
if (classValues.size() > 2) {
limma = new Limma(babelomicsHomePath + "/bin/diffexp/limma_multiclasses.r");
} else if (classValues.size() == 2) {
limma = new Limma(babelomicsHomePath + "/bin/diffexp/limma_twoclasses.r");
} else if (classValues.size() == 1) {
limma = new Limma(babelomicsHomePath + "/bin/diffexp/limma_oneclass.r");
} else {
abort(
"testmismatched_executelimma_classcomparison",
"test " + test + " not supported for " + classValues.size() + "-class test",
"test " + test + " not supported for " + classValues.size() + "-class test",
"test " + test + " not supported for " + classValues.size() + "-class test");
}
// System.out.println("dataset = " + dataset.toString());
System.out.println("class name = " + className);
limma.setInputFilename(dataset.getDatasetFile().getAbsolutePath());
limma.setClasses(dataset.getVariables().getByName(className).getValues());
limma.setContrast(classValues);
try {
Dataset subDataset = dataset.getSubDataset(className, classValues);
// apply test and multiple test correction according
//
TestResultList<LimmaTestResult> res = limma.compute();
DiffExpressionUtils.multipleTestCorrection(res, correction);
// create output file
//
int[] columnOrder =
ListUtils.order(subDataset.getVariables().getByName(className).getValues());
int[] rowOrder = ListUtils.order(ArrayUtils.toList(res.getStatistics()), true);
DataFrame dataFrame = new DataFrame(subDataset.getFeatureNames().size(), 0);
dataFrame.addColumn(
"statistic",
ListUtils.toStringList(
ListUtils.ordered(ArrayUtils.toList(res.getStatistics()), rowOrder)));
dataFrame.addColumn(
"p-value",
ListUtils.toStringList(ListUtils.ordered(ArrayUtils.toList(res.getPValues()), rowOrder)));
dataFrame.addColumn(
"adj. p-value",
ListUtils.toStringList(
ListUtils.ordered(ArrayUtils.toList(res.getAdjPValues()), rowOrder)));
dataFrame.setRowNames(ListUtils.ordered(subDataset.getFeatureNames(), rowOrder));
FeatureData featureData = new FeatureData(dataFrame);
File file = new File(outdir + "/" + test + ".txt");
featureData.save(file);
if (file.exists()) {
result.addOutputItem(
new Item(
test + "file",
file.getName(),
"Limma output file",
TYPE.FILE,
new ArrayList<String>(2),
new HashMap<String, String>(2),
"Limma output files"));
}
// getting significative genes
//
DiffExpressionUtils.addSignificativeResults(
subDataset,
test,
"statistic",
res.getStatistics(),
"adj. p-value",
res.getAdjPValues(),
"p-value",
res.getPValues(),
null,
null,
null,
null,
className,
columnOrder,
pValue,
maxDisplay,
this);
DiffExpressionUtils.createFatiScanRedirection(dataFrame, test, "statistic", result, outdir);
} catch (Exception e) {
e.printStackTrace();
abort(
"exception_executelimma_classcomparison",
"error running limma",
"error running limma: " + e.toString(),
"error running limma: " + e.toString());
}
}
private void setFoldChangeResults(
Dataset subDataset, String test, String testLabel, double[] res, String className)
throws InvalidIndexException, IOException {
// int[] columnOrder =
// ListUtils.order(subDataset.getVariables().getByName(className).getValues());
int[] rowOrder = ListUtils.order(ArrayUtils.toList(res), true);
DataFrame dataFrame = new DataFrame(subDataset.getFeatureNames().size(), 0);
dataFrame.addColumn(test, ListUtils.ordered(ArrayUtils.toStringList(res), rowOrder));
dataFrame.setRowNames(ListUtils.ordered(subDataset.getFeatureNames(), rowOrder));
FeatureData featureData = new FeatureData(dataFrame);
File file = new File(outdir + "/" + test + "_foldchange.txt");
IOUtils.write(file, dataFrame.toString(true, true));
/** Get significative terms, babelomics 5 * */
List<String> featureNames = new ArrayList<String>();
List<String> featureValues = new ArrayList<String>();
List<String> featureNamesUp = new ArrayList<String>();
List<String> featureNamesDown = new ArrayList<String>();
for (String rowName : dataFrame.getRowNames()) {
List<String> row = dataFrame.getRow(rowName);
double stats = Double.parseDouble(row.get(0));
if (Math.abs(stats) >= foldChangeValue) {
featureNames.add(rowName);
featureValues.add(row.get(0));
if (stats >= 0) featureNamesUp.add(rowName);
else featureNamesDown.add(rowName);
}
}
dataFrame = new DataFrame(featureNames.size(), 0);
dataFrame.addColumn(test, featureValues);
dataFrame.setRowNames(featureNames);
List<Integer> sigRowIndexes = new ArrayList<Integer>();
for (String feat : featureNames) {
int idx = 0;
for (String featSub : subDataset.getFeatureNames()) {
if (feat.equalsIgnoreCase(featSub)) sigRowIndexes.add(idx);
idx++;
}
}
DoubleMatrix doubleMatrix =
new DoubleMatrix(dataFrame.getRowDimension(), subDataset.getColumnDimension());
for (int i = 0; i < sigRowIndexes.size(); i++) {
doubleMatrix.setRow(i, subDataset.getDoubleMatrix().getRow(sigRowIndexes.get(i)));
}
File fileXX = new File(outdir + "/" + test + "_foldchange_significative_dataset.txt");
Dataset sigDataset = new Dataset(subDataset.getSampleNames(), featureNames, doubleMatrix);
sigDataset.save(fileXX);
File fileAux = new File(outdir + "/" + test + "_foldchange_significative_table.txt");
IOUtils.write(fileAux, dataFrame.toString(true, true));
fileAux = new File(outdir + "/" + test + "_foldchange_significative_table_up.txt");
IOUtils.write(fileAux, featureNamesUp);
fileAux = new File(outdir + "/" + test + "_foldchange_significative_table_down.txt");
IOUtils.write(fileAux, featureNamesDown);
// fileAux = new File(outdir + "/" + test + "_foldchange_significative_dataset.txt");
// IOUtils.write(fileAux, featureNamesDown);
//
// List<String> featuresUp = new ArrayList<String>();
// List<String> featuresDown = new ArrayList<String>();
// for (String rowName : dataFrame.getRowNames()) {
// List<String> row = dataFrame.getRow(rowName);
// double stats = Double.parseDouble(row.get(0));
// if(stats>=foldChangeValue){
//
//
//// featuresUp.add(rowName);
//// else
//// featuresDown.add(rowName);
// }
// }
// file = new File(tool.getOutdir() + "/" + test + "_significative_table_up.txt");
// IOUtils.write(file, featuresUp);
// file = new File(tool.getOutdir() + "/" + test + "_significative_table_down.txt");
// IOUtils.write(file, featuresDown);
// featureData.save(file);
if (file.exists()) {
result.addOutputItem(
new Item(
test + "_foldchange",
file.getName(),
testLabel + " fold-change output file",
TYPE.FILE,
new ArrayList<String>(),
new HashMap<String, String>(),
testLabel + " fold-change.Output files"));
String json =
"{\\\"paramfilename\\\": \\\"input_params.txt\\\", \\\"testfilename\\\": \\\""
+ file.getName()
+ "\\\"}";
result.addOutputItem(
new Item(
"diff_expr_" + StringUtils.randomString(8),
json,
"Significative results",
TYPE.FILE,
StringUtils.toList("DIFF_EXPRESSION_VIEWER"),
new HashMap<String, String>(),
testLabel + " fold-change. Significative results"));
DiffExpressionUtils.createFatiScanRedirection(
dataFrame, test, test, result, outdir, testLabel + " fold-change.");
}
/*
List<Double> orderedRes = ListUtils.ordered(ArrayUtils.toList(res), rowOrder);
int posValues = 0;
int negValues = 0;
for(int i=0 ; i<orderedRes.size() ; i++) {
if (Math.abs(orderedRes.get(i))>foldChangeValue) {
if (orderedRes.get(i)>0) {
posValues++;
} else {
negValues++;
}
}
}
if (posValues + negValues == 0) {
result.addOutputItem(new Item("no_sig_results", "No significative results (fold-change value = " + foldChangeValue + ")", "Significative results", TYPE.MESSAGE, new ArrayList<String>(), new HashMap<String, String>(2), testLabel + " fold-change.Significative results"));
return;
}
int halfDisplay = maxDisplay/2;
int posValuesToDisplay = posValues;
int negValuesToDisplay = negValues;
if (posValues + negValues > maxDisplay) {
if (Math.min(posValues, negValues)>halfDisplay) {
posValuesToDisplay = halfDisplay;
negValuesToDisplay = halfDisplay;
} else {
posValuesToDisplay = posValues>negValues ? (maxDisplay-negValues) : posValues;
negValuesToDisplay = negValues>posValues ? (maxDisplay-posValues) : negValues;
}
}
int nbToDisplay = posValuesToDisplay + negValuesToDisplay;
DoubleMatrix doubleMatrix = new DoubleMatrix(nbToDisplay, subDataset.getColumnDimension());
int rowIndex = 0;
int negLimit = rowOrder.length-negValuesToDisplay;
List<Integer> sigRowIndexes = new ArrayList<Integer>();
for(int i=0 ; i<rowOrder.length ; i++) {
if (i<posValuesToDisplay || i>=negLimit) {
doubleMatrix.setRow(rowIndex++, subDataset.getDoubleMatrix().getRow(rowOrder[i]));
//System.out.println(subDataset.getFeatureNames().get(sigOrder[i]));
sigRowIndexes.add(rowOrder[i]);
}
}
file = new File(outdir + "/" + test +"_fold_change_significative_dataset.txt");
Dataset sigDataset = new Dataset(subDataset.getSampleNames(), ListUtils.subList(subDataset.getFeatureNames(), ListUtils.toIntArray(sigRowIndexes)), doubleMatrix);
sigDataset.setVariables(subDataset.getVariables());
sigDataset.validate();
sigDataset.save(file);
if (file.exists()) {
String tags = "datamatrix,expression";
result.addOutputItem(new Item(test + "_sig_dataset", file.getName(), "Significative values dataset (fold-change value = " + foldChangeValue + ")", TYPE.DATA, StringUtils.toList(tags, ","), new HashMap<String, String>(2), testLabel + " fold-change.Significative results"));
File redirectionFile = new File(outdir + "/clustering.redirection");
DiffExpressionUtils.createClusteringRedirectionFile(redirectionFile, file);
if ( redirectionFile.exists() ) {
tags = "REDIRECTION(" + redirectionFile.getName() + ":Send to Clustering tool...)";
result.addOutputItem(new Item(test + "_sig_dataset", file.getName(), "Significative values dataset (fold-change value = " + foldChangeValue + ")", TYPE.FILE, StringUtils.toList(tags, ","), new HashMap<String, String>(2), testLabel + " fold-change.Significative results"));
}
}
rowOrder = ListUtils.order(ListUtils.subList(ArrayUtils.toList(res), ListUtils.toIntArray(sigRowIndexes)), true);
DataFrame sigDataFrame = new DataFrame(sigDataset.getFeatureNames().size(), 0);
sigDataFrame.addColumn(test, ListUtils.toStringList(ListUtils.ordered(ListUtils.subList(ArrayUtils.toList(res), ListUtils.toIntArray(sigRowIndexes)), rowOrder)));
sigDataFrame.setRowNames(ListUtils.ordered(ListUtils.subList(subDataset.getFeatureNames(), ListUtils.toIntArray(sigRowIndexes)), rowOrder));
// adding table to results
//
file = new File(outdir + "/" + test + "fold_change_significative_table.txt");
IOUtils.write(file, sigDataFrame.toString(true, true));
if ( file.exists() ) {
result.addOutputItem(new Item(test + "fold_change_table", file.getName(), "Significative values table (fold-change value = " + foldChangeValue + ")", TYPE.FILE, StringUtils.toList("TABLE," + test.toUpperCase() + "_FOLD_CHANGE_TABLE", ","), new HashMap<String, String>(2), testLabel + " fold-change.Significative results"));
}
// adding heatmap to results
//
Canvas sigHeatmap = DiffExpressionUtils.generateHeatmap(sigDataset, className, columnOrder, rowOrder, testLabel, ListUtils.toDoubleArray(ListUtils.subList(ArrayUtils.toList(res), ListUtils.toIntArray(sigRowIndexes))), null, null);
if (sigHeatmap == null) {
printError("ioexception_execute_fold_change_classcomparison", "ERROR", "Error generating " + test + " fold-change heatmap image");
} else {
try {
File sigHeatmapFile = new File(outdir + "/" + test + "fold_change_heatmap_significative.png");
sigHeatmap.save(sigHeatmapFile.getAbsolutePath());
if (sigHeatmapFile.exists()) {
result.addOutputItem(new Item(test + "_fold_change_heatmap_significative", sigHeatmapFile.getName(), testLabel + " fold-change heatmap with significative values (fold-change value = " + foldChangeValue + ")", TYPE.IMAGE, new ArrayList<String>(2), new HashMap<String, String>(2), testLabel + " fold-change.Significative results"));
}
} catch (IOException e) {
printError("ioexception_execute_fold_change_classcomparison", "ERROR", "Error saving " + test + " fold-change heatmap image");
}
}
DiffExpressionUtils.createFatiGoRedirection(dataFrame.getRowNames(), dataFrame.getColumnAsDoubleArray(test), test, result, outdir, testLabel + " fold-change.");
DiffExpressionUtils.createFatiScanRedirection(sigDataFrame, test, test, result, outdir, testLabel + " fold-change.");
*/
}
public void executeAnova() {
DoubleMatrix matrix = null;
List<String> vars = new ArrayList<String>();
List<Integer> indices = new ArrayList<Integer>();
List<String> values = dataset.getVariables().getByName(className).getValues();
if (values.size() == classValues.size()) {
matrix = dataset.getDoubleMatrix();
vars = values;
} else {
for (int i = 0; i < values.size(); i++) {
if (classValues.contains(values.get(i))) {
indices.add(i);
vars.add(values.get(i));
}
}
matrix = dataset.getSubMatrixByColumns(ListUtils.toIntArray(indices));
}
try {
Dataset subDataset = dataset.getSubDataset(className, classValues);
// apply test and multiple test correction according
//
AnovaTest anova = new AnovaTest(matrix, vars);
TestResultList<AnovaTestResult> res = anova.compute();
DiffExpressionUtils.multipleTestCorrection(res, correction);
// create output file
//
int[] columnOrder =
ListUtils.order(subDataset.getVariables().getByName(className).getValues());
int[] rowOrder = ListUtils.order(ArrayUtils.toList(res.getStatistics()), true);
DataFrame dataFrame = new DataFrame(subDataset.getFeatureNames().size(), 0);
dataFrame.addColumn(
"statistic",
ListUtils.toStringList(
ListUtils.ordered(ArrayUtils.toList(res.getStatistics()), rowOrder)));
dataFrame.addColumn(
"p-value",
ListUtils.toStringList(ListUtils.ordered(ArrayUtils.toList(res.getPValues()), rowOrder)));
dataFrame.addColumn(
"adj. p-value",
ListUtils.toStringList(
ListUtils.ordered(ArrayUtils.toList(res.getAdjPValues()), rowOrder)));
dataFrame.setRowNames(ListUtils.ordered(subDataset.getFeatureNames(), rowOrder));
FeatureData featureData = new FeatureData(dataFrame);
File file = new File(outdir + "/" + test + ".txt");
featureData.save(file);
if (file.exists()) {
result.addOutputItem(
new Item(
test + "file",
file.getName(),
"Anova output file",
TYPE.FILE,
new ArrayList<String>(2),
new HashMap<String, String>(2),
"Anova output files"));
}
// getting significative genes
//
DiffExpressionUtils.addSignificativeResults(
subDataset,
test,
"statistic",
res.getStatistics(),
"adj. p-value",
res.getAdjPValues(),
"p-value",
res.getPValues(),
null,
null,
null,
null,
className,
columnOrder,
pValue,
maxDisplay,
this);
DiffExpressionUtils.createFatiScanRedirection(dataFrame, test, "statistic", result, outdir);
} catch (Exception e) {
e.printStackTrace();
abort(
"exception_executeanova_classcomparison",
"error running anova",
"error running anova: " + e.getMessage(),
"error running anova: " + e.getMessage());
}
}
public void executeT() {
int[] cols = dataset.getColumnIndexesByVariableValue(className, classValues.get(0));
DoubleMatrix sample1 = dataset.getSubMatrixByColumns(cols);
cols = dataset.getColumnIndexesByVariableValue(className, classValues.get(1));
DoubleMatrix sample2 = dataset.getSubMatrixByColumns(cols);
try {
Dataset subDataset = dataset.getSubDataset(className, classValues);
// apply test and multiple test correction according
//
TTest tTest = new TTest();
TestResultList<TTestResult> res = tTest.tTest(sample1, sample2);
DiffExpressionUtils.multipleTestCorrection(res, correction);
// create output file
//
int[] columnOrder =
ListUtils.order(subDataset.getVariables().getByName(className).getValues());
int[] rowOrder = ListUtils.order(ArrayUtils.toList(res.getStatistics()), true);
DataFrame dataFrame = new DataFrame(subDataset.getFeatureNames().size(), 0);
dataFrame.addColumn(
"statistic",
ListUtils.toStringList(
ListUtils.ordered(ArrayUtils.toList(res.getStatistics()), rowOrder)));
dataFrame.addColumn(
"p-value",
ListUtils.toStringList(ListUtils.ordered(ArrayUtils.toList(res.getPValues()), rowOrder)));
dataFrame.addColumn(
"adj. p-value",
ListUtils.toStringList(
ListUtils.ordered(ArrayUtils.toList(res.getAdjPValues()), rowOrder)));
dataFrame.setRowNames(ListUtils.ordered(subDataset.getFeatureNames(), rowOrder));
FeatureData featureData = new FeatureData(dataFrame);
File file = new File(outdir + "/t.txt");
featureData.save(file);
if (file.exists()) {
result.addOutputItem(
new Item(
"tfile",
file.getName(),
"T-test output file",
TYPE.FILE,
new ArrayList<String>(),
new HashMap<String, String>(),
"T-test output files"));
}
// getting significative genes
//
DiffExpressionUtils.addSignificativeResults(
subDataset,
test,
"statistic",
res.getStatistics(),
"adj. p-value",
res.getAdjPValues(),
"p-value",
res.getPValues(),
null,
null,
null,
null,
className,
columnOrder,
pValue,
maxDisplay,
this);
DiffExpressionUtils.createFatiScanRedirection(dataFrame, test, "statistic", result, outdir);
} catch (Exception e) {
e.printStackTrace();
abort("exception_executet_classcomparison", "ERROR", "Error running t-test", "");
}
}