/** Invoked by the framework when this viewer is selected. */
public void onSelected(IFramework framework) {
geneClusterCB.removeAllItems();
for (int i = 0; i < framework.getClusterRepository(0).size(); i++) {
if (framework.getClusterRepository(0).getCluster(i + 1) == null) break;
Cluster cluster = framework.getClusterRepository(0).getCluster(i + 1);
geneClusterCB.addItem(
"Cluster #: " + cluster.getSerialNumber() + ", " + cluster.getClusterLabel());
}
}
public AlgorithmData getScriptParameters(IFramework framework) {
this.experiment = framework.getData().getExperiment();
exptNamesVector = new Vector<String>();
int number_of_samples = experiment.getNumberOfSamples();
// int number_of_genes = experiment.getNumberOfGenes();
for (int i = 0; i < number_of_samples; i++) {
exptNamesVector.add(framework.getData().getFullSampleName(experiment.getSampleIndex(i)));
}
DEGseqInitBox DEGseqDialog =
new DEGseqInitBox(
(JFrame) framework.getFrame(), true, exptNamesVector, framework.getClusterRepository(1)
// framework.getClusterRepository(0)
);
DEGseqDialog.setVisible(true);
if (!DEGseqDialog.isOkPressed()) return null;
if (DEGseqDialog.getTestDesign() == DEGseqInitBox.TWO_CLASS) {
if (DEGseqDialog.getSelectionDesign() == DEGseqInitBox.CLUSTER_SELECTION) {
groupAssignments = DEGseqDialog.getClusterTwoClassAssignments();
}
if (DEGseqDialog.getSelectionDesign() == DEGseqInitBox.BUTTON_SELECTION) {
groupAssignments = DEGseqDialog.getTwoClassAssignments();
}
}
AlgorithmData data = new AlgorithmData();
// alg name
data.addParam("name", "DESeq");
// alg type
data.addParam("alg-type", "data-visualization");
// output class
data.addParam("output-class", "partition-output");
// output nodes
String[] outputNodes = new String[2];
outputNodes[0] = "Significant Genes";
outputNodes[1] = "Non-significant Genes";
data.addStringArray("output-nodes", outputNodes);
return data;
}
public DefaultMutableTreeNode executeScript(
IFramework framework, AlgorithmData algData, Experiment experiment)
throws AlgorithmException {
Listener listener = new Listener();
this.experiment = experiment;
this.data = framework.getData();
// this.timeAssignments = algData.getIntArray("time_assignments");
this.groupAssignments = algData.getIntArray("condition_assignments");
exptNamesVector = new Vector<String>();
int number_of_samples = experiment.getNumberOfSamples();
for (int i = 0; i < number_of_samples; i++) {
exptNamesVector.add(this.data.getFullSampleName(i));
}
try {
algData.addMatrix("experiment", experiment.getMatrix());
algorithm = framework.getAlgorithmFactory().getAlgorithm("MINET");
algorithm.addAlgorithmListener(listener);
this.progress = new Progress(framework.getFrame(), "Running MINET Analysis...", listener);
this.progress.show();
long start = System.currentTimeMillis();
AlgorithmData result = algorithm.execute(algData);
long time = System.currentTimeMillis() - start;
AlgorithmParameters params = algData.getParams();
GeneralInfo info = new GeneralInfo();
return createResultTree(info);
} finally {
if (algorithm != null) {
algorithm.removeAlgorithmListener(listener);
}
if (progress != null) {
progress.dispose();
}
}
}
/**
* Constructs a <code>Box/Bar Chart Viewer</code> for insertion into ClusterTable
*
* @param experiment the data of an experiment.
*/
public BoxChartViewer(IFramework fm, Cluster[] clusterArray) {
inAnalysis = false;
framework = fm;
this.experiment = framework.getData().getExperiment();
setClusters(clusterArray);
populateGeneListComboBoxes();
generateSettingsComponents();
createComponentLayout();
init();
}
private void populateGeneListComboBoxes() {
geneClusterCB = new JComboBox();
geneClusterCB.removeAllItems();
for (int i = 0; i < framework.getClusterRepository(0).size(); i++) {
if (framework.getClusterRepository(0).getCluster(i + 1) == null) break;
Cluster cluster = framework.getClusterRepository(0).getCluster(i + 1);
geneClusterCB.addItem(
"Cluster #: " + cluster.getSerialNumber() + ", " + cluster.getClusterLabel());
}
geneClusterCB.setVisible(false);
geneClusterCB.addActionListener(
new ActionListener() {
public void actionPerformed(ActionEvent e) {
repaint();
}
});
geneAnnotationCB = new JComboBox(framework.getData().getAllFilledAnnotationFields());
geneAnnotationCB.addActionListener(
new ActionListener() {
public void actionPerformed(ActionEvent e) {
if (chartTypeCB.getSelectedIndex() > 1) // this method only applies for single gene
return;
geneCB.removeAllItems();
int selected = Math.max(geneAnnotationCB.getSelectedIndex(), 0);
int numGenes =
framework
.getData()
.getAnnotationList(framework.getData().getAllFilledAnnotationFields()[selected])
.length;
StringExt[] stringext = new StringExt[numGenes];
String[] geneNames =
framework
.getData()
.getAnnotationList(
framework.getData().getAllFilledAnnotationFields()[selected]);
for (int i = 0; i < numGenes; i++) stringext[i] = new StringExt(geneNames[i]);
DefaultComboBoxModel cbm = new DefaultComboBoxModel(stringext);
geneCB.setModel(cbm);
}
});
geneCB =
new JComboBox(
framework
.getData()
.getAnnotationList(framework.getData().getAllFilledAnnotationFields()[0]));
geneCB.setMaximumSize(new Dimension(150, geneCB.getHeight()));
geneCB.addActionListener(
new ActionListener() {
public void actionPerformed(ActionEvent e) {
repaint();
}
});
}
/**
* Constructs a <code>Box/Bar Chart Viewer</code> for use in analyses
*
* @param experiment the data of an experiment.
*/
public BoxChartViewer(
IFramework fm,
int[][] sampleArray,
int[] geneArray,
String geneClusterLabel,
String[] sampleClusterNames) {
this.sampleClusterNames = sampleClusterNames;
inAnalysis = true;
framework = fm;
this.geneClusterLabel = geneClusterLabel;
this.experiment = framework.getData().getExperiment();
setClusters(sampleArray);
geneIndices = geneArray;
populateGeneListComboBoxes();
generateSettingsComponents();
createComponentLayout();
init();
}
/**
* This method should return a tree with calculation results or null, if analysis start was
* canceled.
*
* @param framework the reference to <code>IFramework</code> implementation, which is used to
* obtain an initial analysis data and parameters.
* @throws AlgorithmException if calculation was failed.
* @throws AbortException if calculation was canceled.
* @see IFramework
*/
public DefaultMutableTreeNode execute(IFramework framework) throws AlgorithmException {
if (sysMsg("R 2.11.x", "DEGSeq") != JOptionPane.OK_OPTION) return null;
this.data = framework.getData();
exptNamesVector = new Vector<String>();
for (int i = 0; i < this.data.getFeaturesCount(); i++) {
exptNamesVector.add(framework.getData().getFullSampleName(i));
}
DEGseqInitBox DEGseqDialog =
new DEGseqInitBox(
(JFrame) framework.getFrame(),
true,
exptNamesVector,
framework.getClusterRepository(1));
DEGseqDialog.setVisible(true);
if (!DEGseqDialog.isOkPressed()) return null;
infMethod = DEGseqDialog.getMethodName();
sigMethod = DEGseqDialog.getCutOffField();
sigCutOff = DEGseqDialog.getPValue();
dataDesign = DEGseqDialog.getTestDesign();
if (DEGseqDialog.getTestDesign() == DEGseqInitBox.TWO_CLASS) {
groupAssignments = DEGseqDialog.getTwoClassAssignments();
if (DEGseqDialog.getSelectionDesign() == DEGseqInitBox.CLUSTER_SELECTION) {
groupAssignments = DEGseqDialog.getClusterTwoClassAssignments();
}
if (DEGseqDialog.getSelectionDesign() == DEGseqInitBox.BUTTON_SELECTION) {
groupAssignments = DEGseqDialog.getTwoClassAssignments();
}
}
// count # of samples used in analysis
int samplesUsed = 0;
for (int i = 0; i < groupAssignments.length; i++) {
if (groupAssignments[i] != 0) samplesUsed++;
}
// get samples indices used
int sampleIndices[] = new int[samplesUsed];
for (int i = 0, ii = 0; i < groupAssignments.length; i++) {
if (groupAssignments[i] != 0) sampleIndices[ii++] = i;
}
// set up the group struct for algo
int[] twoClassGrps = new int[samplesUsed];
for (int i = 0, ii = 0; i < groupAssignments.length; i++) {
if (groupAssignments[i] != 0) twoClassGrps[ii++] = groupAssignments[i];
}
// Test
int grp1 = 0;
int grp2 = 0;
String grp1ColIndStr = "";
String grp2ColIndStr = "";
for (int i = 0; i < twoClassGrps.length; i++) {
if (twoClassGrps[i] == 1) {
grp1++;
grp1ColIndStr += String.valueOf(i + 2) + ","; // 1 for 0 based index, 1 for UID col in file
}
if (twoClassGrps[i] == 2) {
grp2++;
grp2ColIndStr += String.valueOf(i + 2) + ",";
}
}
// remove last ","
grp1ColIndStr = grp1ColIndStr.substring(0, grp1ColIndStr.lastIndexOf(","));
grp2ColIndStr = grp2ColIndStr.substring(0, grp2ColIndStr.lastIndexOf(","));
System.out.println("groupAssignments: " + Arrays.toString(groupAssignments));
System.out.println("twoClassGrps: " + Arrays.toString(twoClassGrps));
System.out.println(
samplesUsed
+ " out of "
+ groupAssignments.length
+ " used. Sample indices: "
+ Arrays.toString(sampleIndices));
System.out.println("grp1ColInd: " + grp1ColIndStr);
System.out.println("grp2ColInd: " + grp2ColIndStr);
this.experiment = framework.getData().getExperiment();
// int number_of_samples = this.data.getFeaturesCount();//experiment.getNumberOfSamples();
// int [] columnIndices = experiment.getColumnIndicesCopy();
sampleLabels = new ArrayList<String>();
geneLabels = new ArrayList<String>();
for (int i = 0; i < samplesUsed; i++) {
sampleLabels.add(framework.getData().getFullSampleName(sampleIndices[i])); // Raktim
}
// Raktim Use probe index as the gene labels in R
for (int i = 0; i < this.data.getFeaturesSize(); /*experiment.getNumberOfGenes();*/ i++) {
// geneLabels.add(framework.getData().getElementAnnotation(i,
// AnnotationFieldConstants.PROBE_ID)[0]); //Raktim
geneLabels.add(String.valueOf(i));
}
// Make Count Matrix based on data
int numGenes = this.data.getFeaturesSize();
System.out.println(
"data.getFeaturesCount(): "
+ this.data.getFeaturesCount()
+ " data.getFeaturesSize(): "
+ this.data.getFeaturesSize());
ArrayList<IRNASeqSlide> temp = (ArrayList<IRNASeqSlide>) data.getFeaturesList();
// TEst Code from EH
RNASeqChipAnnotation chipAnnotation = (RNASeqChipAnnotation) data.getChipAnnotation();
System.out.println("Library size, first sample: " + temp.get(0).getLibrarySize());
System.out.println("Library size, last sample: " + temp.get(temp.size() - 1).getLibrarySize());
System.out.println("\n");
System.out.println("Read length: " + chipAnnotation.getReadLength());
System.out.println("\n");
System.out.println("first count value for first slide: " + temp.get(0).getCount(0));
System.out.println(
"last count value for first slide: " + temp.get(0).getCount(temp.get(0).getSize() - 1));
System.out.println("\n");
System.out.println(
"first count value for last slide: " + temp.get(temp.size() - 1).getCount(0));
System.out.println(
"last count value for last slide: "
+ temp.get(temp.size() - 1).getCount(temp.get(0).getSize() - 1));
System.out.println("\n");
System.out.println(
"transcript length 0,0: "
+ ((RNASeqElement) (temp.get(0).getSlideDataElement(0))).getTranscriptLength());
System.out.println(
"classcode for 0,0: "
+ ((RNASeqElement) (temp.get(0).getSlideDataElement(0))).getClasscode());
// End EH Test Code
// My test code
// System.out.println("transcript length 0,0: " +
// ((RNASeqElement)(temp.get(0).getSlideDataElement(9))).getTranscriptLength());
// System.out.println("classcode for 0,0: " +
// ((RNASeqElement)(temp.get(0).getSlideDataElement(9))).getClasscode());
// End my test code
/** ArrayList<IRNASeqSlide> temp = (ArrayList<IRNASeqSlide>)data.getFeaturesList(); */
int[][] countMatrix = new int[numGenes][samplesUsed];
int[] libSize = new int[samplesUsed];
for (int row = 0; row < numGenes; row++) {
for (int col = 0; col < sampleIndices.length; col++) {
// get count value
countMatrix[row][col] = temp.get(sampleIndices[col]).getCount(row);
}
// get transcript len for each gene
// transcriptLen[row] =
// ((RNASeqElement)(temp.get(row).getSlideDataElement(row))).getTranscriptLength();
// transcriptLen[row] =
// ((RNASeqElement)(temp.get(0).getSlideDataElement(row))).getTranscriptLength();
}
// get lib size for each sample used
for (int col = 0; col < sampleIndices.length; col++) {
// get count value
libSize[col] = temp.get(sampleIndices[col]).getLibrarySize();
}
// System.out.println("CountMatrix: " + Arrays.deepToString(countMatrix));
// System.out.println("transcriptLen: " + Arrays.toString(transcriptLen));
// System.out.println("libSize: " + Arrays.toString(libSize));
Listener listener = new Listener();
try {
algorithm = framework.getAlgorithmFactory().getAlgorithm("DEGSEQ");
algorithm.addAlgorithmListener(listener);
this.progress = new Progress(framework.getFrame(), "Running DEGseq ...", listener);
this.progress.show();
AlgorithmData data = new AlgorithmData();
data.addIntMatrix("experiment", countMatrix);
data.addParam("dataDesign", String.valueOf(dataDesign));
data.addIntArray("group_assignments", twoClassGrps);
data.addParam("grp1ColIndStr", grp1ColIndStr);
data.addParam("grp2ColIndStr", grp2ColIndStr);
data.addIntArray("libSize", libSize);
data.addParam("numGenes", String.valueOf(numGenes));
data.addParam("numExps", String.valueOf(samplesUsed));
data.addParam("grp1", String.valueOf(grp1));
data.addParam("grp2", String.valueOf(grp2));
data.addParam("infMethod", infMethod);
data.addStringArray("geneLabels", geneLabels.toArray(new String[geneLabels.size()]));
data.addStringArray("sampleLabels", sampleLabels.toArray(new String[sampleLabels.size()]));
// run algorithm
long start = System.currentTimeMillis();
AlgorithmData result = algorithm.execute(data);
long time = System.currentTimeMillis() - start;
// getting the results
resultMatrix = result.getMatrix("result");
resultRowNames = result.getStringArray("rownames");
// System.out.println(Arrays.toString(resultRowNames));
// Process results
createHeaderNames();
createAuxData();
createResultClusters();
GeneralInfo info = new GeneralInfo();
info.time = time;
System.out.println("Creating Viewers for DEGSeq...");
return createResultTree(info);
} finally {
if (algorithm != null) {
algorithm.removeAlgorithmListener(listener);
}
if (progress != null) {
progress.dispose();
}
}
}
public void paint(Graphics g) {
super.paint(g);
if (clusterCount == 0) return;
boolean ismultigene = !isSingleGene;
Graphics2D g2 = (Graphics2D) g;
multigeneCount = 1;
Cluster cluster = null;
if (ismultigene) {
try {
if (!inAnalysis) {
int begin = geneClusterCB.getSelectedItem().toString().indexOf(":");
int end = geneClusterCB.getSelectedItem().toString().indexOf(",");
sn = geneClusterCB.getSelectedItem().toString().substring(begin + 2, end);
cluster = framework.getClusterRepository(0).getCluster(Integer.parseInt(sn));
geneIndices = cluster.getIndices();
}
} catch (Exception e) {
e.printStackTrace();
System.out.println("Gene Cluster Problem");
return;
}
if (geneIndices == null || geneIndices.length < 1) return;
if (aggregateGeneCluster) multigeneCount = 1;
else multigeneCount = geneIndices.length;
}
int chartTop = chartTopHeight;
lowerOutliers = new float[multigeneCount][][];
upperOutliers = new float[multigeneCount][][];
expressionSDs = new float[multigeneCount][];
expressionAverages = new float[multigeneCount][];
expressionMedian = new float[multigeneCount][];
expressionQ1 = new float[multigeneCount][];
expressionQ3 = new float[multigeneCount][];
expressionIQR = new float[multigeneCount][];
expressionLowerWhisker = new float[multigeneCount][];
expressionUpperWhisker = new float[multigeneCount][];
chartTitles = new String[multigeneCount];
int maxString = 0;
for (int i = 0; i < clusterNames.length; i++) {
maxString = Math.max(maxString, g.getFontMetrics().stringWidth(clusterNames[i]));
}
chartGap = maxString + 50;
// for only drawing the visible area- dramatic speed improvements
Rectangle bounds = g.getClipBounds();
int indexBegin = (int) Math.max(-3 + bounds.y / (chartHeight + chartGap), 0f);
int indexEnd =
(int) Math.min(multigeneCount, 3 + (bounds.y + bounds.height) / (chartHeight + chartGap));
for (int multigeneIndex = indexBegin; multigeneIndex < indexEnd; multigeneIndex++) {
if (ismultigene) {
if (aggregateGeneCluster) {
getChartData(geneIndices, multigeneIndex);
if (inAnalysis) chartTitles[multigeneIndex] = geneClusterLabel;
else chartTitles[multigeneIndex] = cluster.getClusterLabel();
} else {
getChartData(geneIndices[multigeneIndex], multigeneIndex);
chartTitles[multigeneIndex] =
framework.getData()
.getElementAnnotation(
geneIndices[multigeneIndex], framework.getData().getFieldNames()[0])[0];
}
} else {
getChartData(geneCB.getSelectedIndex(), multigeneIndex);
chartTitles[multigeneIndex] = geneCB.getSelectedItem().toString();
}
float maxHeight = 0;
float minHeight = Float.POSITIVE_INFINITY;
for (int i = 0; i < clusterCount; i++) {
if (this.isBoxPlot) {
for (int j = 0; j < this.upperOutliers[multigeneIndex][i].length; j++) {
maxHeight = Math.max(upperOutliers[multigeneIndex][i][j], maxHeight);
}
maxHeight = Math.max(expressionUpperWhisker[multigeneIndex][i], maxHeight);
} else {
maxHeight =
Math.max(
expressionAverages[multigeneIndex][i] + this.expressionSDs[multigeneIndex][i],
maxHeight);
}
}
for (int i = 0; i < clusterCount; i++) {
if (this.isBoxPlot) {
for (int j = 0; j < this.lowerOutliers[multigeneIndex][i].length; j++) {
minHeight = Math.min(lowerOutliers[multigeneIndex][i][j], minHeight);
}
minHeight = Math.min(expressionLowerWhisker[multigeneIndex][i], minHeight);
} else minHeight = 0f; // Math.min(expressionAverages[i], minHeight);
}
maxHeight++;
float topgridline;
float bottomgridline;
if (this.autoScale) {
topgridline = (int) maxHeight + (int) maxHeight % 2;
bottomgridline = (int) minHeight - (int) minHeight % 2 - 2;
if (!isBoxPlot) bottomgridline = 0f;
fixedTop = topgridline;
fixedBottom = bottomgridline;
} else {
topgridline = fixedTop;
bottomgridline = fixedBottom;
if (!isBoxPlot) bottomgridline = 0f;
}
float gridrange = topgridline - bottomgridline;
// draw everything
// grid
g2.setColor(Color.black);
g2.drawLine(
chartleft - 5,
chartHeight + chartTop,
chartleft + clusterCount * chartIncrement,
chartHeight + chartTop); // horizontal line bottom
g2.drawLine(chartleft, chartTop, chartleft, chartHeight + chartTop + 5); // vertical line left
g2.setColor(Color.gray);
g2.drawLine(chartleft - 7, chartTop, chartleft, chartTop); // horizontal line ticktop
g2.drawLine(
chartleft - 5,
chartTop + chartHeight / 4,
chartleft,
chartTop + chartHeight / 4); // horizontal line tick75%
g2.drawLine(
chartleft - 7,
chartTop + chartHeight / 2,
chartleft,
chartTop + chartHeight / 2); // horizontal line tickMid
g2.drawLine(
chartleft - 5,
chartTop + 3 * chartHeight / 4,
chartleft,
chartTop + 3 * chartHeight / 4); // horizontal line tick25%
if (showGridlines) {
g2.setColor(Color.gray);
g2.drawLine(
chartleft,
chartTop,
chartleft + chartIncrement * clusterCount,
chartTop); // horizontal line grid
g2.drawLine(
chartleft,
chartTop + chartHeight / 4,
chartleft + chartIncrement * clusterCount,
chartTop + chartHeight / 4); // horizontal line grid
g2.drawLine(
chartleft,
chartTop + chartHeight / 2,
chartleft + chartIncrement * clusterCount,
chartTop + chartHeight / 2); // horizontal line grid
g2.drawLine(
chartleft,
chartTop + 3 * chartHeight / 4,
chartleft + chartIncrement * clusterCount,
chartTop + 3 * chartHeight / 4); // horizontal line grid
}
g2.setColor(Color.black);
g2.drawString(
Float.toString(bottomgridline + gridrange),
chartleft - 40,
chartTop + 3); // vertical axis label
g2.drawString(
Float.toString(bottomgridline + gridrange * 3f / 4f),
chartleft - 40,
(chartTop + 3) + chartHeight / 4); // vertical axis label
g2.drawString(
Float.toString(bottomgridline + gridrange / 2f),
chartleft - 40,
(chartTop + 3) + chartHeight / 2); // vertical axis label
g2.drawString(
Float.toString(bottomgridline + gridrange / 4f),
chartleft - 40,
(chartTop + 3) + (3 * chartHeight / 4)); // vertical axis label
g2.drawString(
Float.toString(bottomgridline),
chartleft - 40,
(chartTop + 3) + (chartHeight)); // vertical axis label
Font font = g2.getFont();
g2.setFont(new Font("Helvetica", Font.BOLD, 16));
g2.drawString(
chartTitles[multigeneIndex],
chartleft + chartIncrement / 2,
chartTop - 15); // Gene/group Name label
g2.setFont(font);
g2.rotate(-Math.PI / 2);
g2.drawString(
"Expression", -((chartTop + 3) + 2 * chartHeight / 3), (chartleft - 60)); // y-axis Label
g2.rotate(Math.PI / 2);
for (int i = 0; i < this.clusterCount; i++) {
int xorigin = (int) ((.25f + (float) i) * chartIncrement) + chartleft;
int yorigin = chartHeight + chartTop;
int barWidth = chartIncrement / 2;
int barHeight =
(int) ((expressionAverages[multigeneIndex][i]) * ((float) chartHeight) / gridrange);
int errorBarHeight =
(int) (expressionSDs[multigeneIndex][i] * ((float) chartHeight) / gridrange);
int boxBottom =
(int)
((-bottomgridline + expressionQ1[multigeneIndex][i])
* ((float) chartHeight)
/ gridrange);
int boxTop =
(int)
((-bottomgridline + expressionQ3[multigeneIndex][i])
* ((float) chartHeight)
/ gridrange);
int boxMedian =
(int)
((-bottomgridline + expressionMedian[multigeneIndex][i])
* ((float) chartHeight)
/ gridrange);
int boxLowerWhisker =
(int)
((-bottomgridline + expressionLowerWhisker[multigeneIndex][i])
* ((float) chartHeight)
/ gridrange);
int boxUpperWhisker =
(int)
((-bottomgridline + expressionUpperWhisker[multigeneIndex][i])
* ((float) chartHeight)
/ gridrange);
g2.setColor(Color.gray);
g2.drawLine(
xorigin + 3 * barWidth / 2,
yorigin,
xorigin + 3 * barWidth / 2,
yorigin + 5); // vertical tickmark
g2.setColor(Color.black);
g2.rotate(-Math.PI / 2);
g2.drawString(
clusterNames[i],
-(yorigin + 5) - g.getFontMetrics().stringWidth(clusterNames[i]),
(xorigin + barWidth / 2)); // cluster names
g2.rotate(Math.PI / 2);
if (!isBoxPlot) { // bar graph
g2.setColor(clusterColors[i]);
g2.fillRect(xorigin, yorigin - barHeight, barWidth, barHeight); // box
g2.setColor(Color.black);
g2.drawRect(xorigin, yorigin - barHeight, barWidth, barHeight); // box outline
g2.drawLine(
xorigin + barWidth / 2,
yorigin - barHeight - errorBarHeight,
xorigin + barWidth / 2,
yorigin - barHeight); // vertical errorbar
g2.drawLine(
xorigin + barWidth / 4,
yorigin - barHeight - errorBarHeight,
xorigin + 3 * barWidth / 4,
yorigin - barHeight - errorBarHeight); // horizontal error bar
} else if (isBoxPlot) { // box plot
g2.setColor(clusterColors[i]);
g2.fillRect(xorigin, yorigin - boxTop, barWidth, boxTop - boxBottom); // box
g2.setColor(Color.black);
g2.drawRect(xorigin, yorigin - boxTop, barWidth, boxTop - boxBottom); // box outline
g2.drawLine(
xorigin, yorigin - boxMedian, xorigin + barWidth, yorigin - boxMedian); // median line
g2.drawLine(
xorigin + barWidth / 2,
yorigin - boxTop,
xorigin + barWidth / 2,
yorigin - boxUpperWhisker); // vertical whisker upper
g2.drawLine(
xorigin + barWidth / 2,
yorigin - boxBottom,
xorigin + barWidth / 2,
yorigin - boxLowerWhisker); // vertical whisker lower
g2.drawLine(
xorigin + barWidth / 4,
yorigin - boxUpperWhisker,
xorigin + 3 * barWidth / 4,
yorigin - boxUpperWhisker); // horizontal whisker upper
g2.drawLine(
xorigin + barWidth / 4,
yorigin - boxLowerWhisker,
xorigin + 3 * barWidth / 4,
yorigin - boxLowerWhisker); // horizontal whisker lower
// draw outliers
int circleWidth = 6;
for (int j = 0; j < upperOutliers[multigeneIndex][i].length; j++) {
g2.drawOval(
xorigin + barWidth / 2 - circleWidth / 2,
yorigin
- (int)
((float) ((-bottomgridline + upperOutliers[multigeneIndex][i][j]))
* ((float) chartHeight)
/ gridrange)
- circleWidth / 2,
circleWidth,
circleWidth);
}
for (int j = 0; j < lowerOutliers[multigeneIndex][i].length; j++) {
g2.drawOval(
xorigin + barWidth / 2 - circleWidth / 2,
yorigin
- (int)
((float) ((-bottomgridline + lowerOutliers[multigeneIndex][i][j]))
* ((float) chartHeight)
/ gridrange)
- circleWidth / 2,
circleWidth,
circleWidth);
}
} else {
System.out.println("unknown chart type");
}
}
chartTop =
chartTopHeight
+ (multigeneIndex + 1) * (chartHeight + chartGap); // moves top down for next graph
}
jScrollBar.setUnitIncrement(chartHeight + chartGap);
int viewerHeight = chartTopHeight + (multigeneCount) * (chartHeight + chartGap);
this.setSize(new Dimension(this.getWidth(), viewerHeight));
this.setPreferredSize(new Dimension(this.getWidth(), viewerHeight));
}
/** Initialize the algorithm's parameters and execute it. */
public DefaultMutableTreeNode execute(IFramework framework) throws AlgorithmException {
this.framework = framework;
this.experiment = framework.getData().getExperiment();
this.iData = framework.getData();
Listener listener = new Listener();
try {
DAMInitDialog damInitDialog = new DAMInitDialog(framework.getFrame(), true);
if (damInitDialog.showModal() != JOptionPane.OK_OPTION) {
return null;
}
classifyGenes = damInitDialog.isEvaluateGenesSelected();
algorithmSelection = damInitDialog.getAssessmentSelection();
isPDA = damInitDialog.isPDASelected();
numberOfClasses = damInitDialog.getNumClasses();
kValue = damInitDialog.getKValue();
alpha = damInitDialog.getAlphaValue();
preSelectGenes = !(damInitDialog.getSkipGeneSelectionValue());
DAMClassificationEditor damClassEditor =
new DAMClassificationEditor(framework, classifyGenes, numberOfClasses);
damClassEditor.setVisible(true);
while (!damClassEditor.isNextPressed()) {
if (damClassEditor.isCancelPressed()) {
return null;
} else {
continue;
}
}
AlgorithmData data = new AlgorithmData();
boolean useGenes = damInitDialog.isEvaluateGenesSelected();
if (useGenes) {
mode = 1;
data.addParam("dam-mode", "1");
} else {
mode = 3;
data.addParam("dam-mode", "3");
}
classificationVector = damClassEditor.getClassification();
trainingIndices = new int[classificationVector[0].size()];
classes = new int[classificationVector[1].size()];
testIndices = new int[classificationVector[2].size()];
for (int i = 0; i < trainingIndices.length; i++) {
trainingIndices[i] = ((Integer) (classificationVector[0].get(i))).intValue();
classes[i] = ((Integer) (classificationVector[1].get(i))).intValue();
}
for (int i = 0; i < testIndices.length; i++) {
testIndices[i] = ((Integer) (classificationVector[2].get(i))).intValue();
}
algorithm = framework.getAlgorithmFactory().getAlgorithm("DAM");
algorithm.addAlgorithmListener(listener);
logger = new Logger(framework.getFrame(), "DAM Log Window", listener);
logger.show();
logger.append("Starting DAM calculation\n");
FloatMatrix Cov;
Experiment experiment = framework.getData().getExperiment();
if (classifyGenes) {
// Problem here: if the program has a gene cluster (maybe also if it
// has an experiment cluster) the matrix returned is null.
FloatMatrix temp;
temp = (experiment.getMatrix()).transpose();
// System.out.println("floatmatrix size: " + temp.m + ", " + temp.n);
data.addMatrix("experiment", temp);
} else {
data.addMatrix("experiment", experiment.getMatrix());
}
data.addParam("distance-factor", String.valueOf(1.0f));
IDistanceMenu menu = framework.getDistanceMenu();
data.addParam("distance-absolute", String.valueOf(menu.isAbsoluteDistance()));
data.addParam("algorithmSelection", String.valueOf(algorithmSelection));
data.addParam("isPDA", String.valueOf(isPDA));
data.addParam("preSelectGenes", String.valueOf(preSelectGenes));
data.addParam("numberOfClasses", String.valueOf(numberOfClasses));
data.addParam("kValue", String.valueOf(kValue));
data.addParam("alpha", String.valueOf(alpha));
data.addIntArray("trainingIndices", trainingIndices);
data.addIntArray("classes", classes);
data.addIntArray("testIndices", testIndices);
int function = menu.getDistanceFunction();
if (function == Algorithm.DEFAULT) {
function = Algorithm.COVARIANCE;
}
data.addParam("distance-function", String.valueOf(function));
data.addParam("dam-mode", String.valueOf(mode));
AlgorithmData result = null;
DefaultMutableTreeNode node = null;
long start = System.currentTimeMillis();
switch (mode) {
case 1: // Spots
data.addParam("distance-function", String.valueOf(function));
result = algorithm.execute(data);
matrixS = result.getMatrix("S");
matrix3D = result.getMatrix("matrix3D");
usedGeneIndices = result.getIntArray("usedGeneIndices");
unusedGeneIndices = result.getIntArray("unusedGeneIndices");
node = new DefaultMutableTreeNode("DAM - genes");
break;
case 3: // Experiments
result = algorithm.execute(data);
matrixS = result.getMatrix("S");
matrix3D = result.getMatrix("matrix3D");
usedGeneIndices = result.getIntArray("usedGeneIndices");
unusedGeneIndices = result.getIntArray("unusedGeneIndices");
/*
if (preSelectGenes) {
System.out.println("DAMGUI.java: usedGeneIndices size: " + usedGeneIndices.length);
for(int i=0; i< usedGeneIndices.length; i++) {
System.out.print(usedGeneIndices[i] + ", ");
}
System.out.println(" ");
System.out.println(" ");
System.out.println("DAMGUI.java: unusedGeneIndices size: " + unusedGeneIndices.length);
for(int i=0; i< unusedGeneIndices.length; i++) {
System.out.print(unusedGeneIndices[i] + ", ");
}
System.out.println(" ");
System.out.println(" ");
}
*/
node = new DefaultMutableTreeNode("DAM - samples");
break;
default:
break;
}
Cluster result_cluster = result.getCluster("cluster");
NodeList nodeList = result_cluster.getNodeList();
int k = numberOfClasses * 3;
this.clusters = new int[k][];
// System.out.println(" ");
for (int i = 0; i < k; i++) {
clusters[i] = nodeList.getNode(i).getFeaturesIndexes();
}
Cluster gene_cluster = result.getCluster("geneCluster");
nodeList = gene_cluster.getNodeList();
this.geneClusters = new int[2][];
// System.out.println(" ");
for (int i = 0; i < 2; i++) {
geneClusters[i] = nodeList.getNode(i).getFeaturesIndexes();
}
this.means = result.getMatrix("clusters_means");
this.variances = result.getMatrix("clusters_variances");
this.means_used = result.getMatrix("clusters_means_used");
this.variances_used = result.getMatrix("clusters_variances_used");
this.means_unused = result.getMatrix("clusters_means_unused");
this.variances_unused = result.getMatrix("clusters_variances_unused");
columns = new int[(experiment.getMatrix()).getColumnDimension()];
for (int i = 0; i < columns.length; i++) {
columns[i] = i;
}
rows = new int[(experiment.getMatrix()).getRowDimension()];
for (int i = 0; i < rows.length; i++) {
rows[i] = i;
}
if (classifyGenes) {
usedExperiment =
new Experiment(
(experiment.getMatrix()).getMatrix(rows, usedGeneIndices), usedGeneIndices, rows);
unusedExperiment =
new Experiment(
(experiment.getMatrix()).getMatrix(rows, unusedGeneIndices),
unusedGeneIndices,
rows);
} else {
usedExperiment =
new Experiment(
(experiment.getMatrix()).getMatrix(usedGeneIndices, columns),
columns,
usedGeneIndices);
unusedExperiment =
new Experiment(
(experiment.getMatrix()).getMatrix(unusedGeneIndices, columns),
columns,
unusedGeneIndices);
}
/*
System.out.println("DAMGUI.java - means: " + means.getRowDimension() + " X " + means.getColumnDimension());
System.out.println("DAMGUI.java - variances: " + variances.getRowDimension() + " X " + variances.getColumnDimension());
System.out.println("DAMGUI.java - matrix3D " + matrix3D.getRowDimension() + " X " + matrix3D.getColumnDimension());
for(int i=0; i< matrix3D.getRowDimension(); i++) {
for(int j=0; j< matrix3D.getColumnDimension(); j++) {
System.out.print(matrix3D.get(i, j) + ", ");
}
System.out.println(" ");
}
*/
logger.append("Creating the result viewers\n");
long time = System.currentTimeMillis() - start;
if (algorithmSelection == A3) // only classification
addClassificationResultNodes(
framework.getFrame(), node, time, menu.getFunctionName(function), experiment);
else
addValidationResultNodes(
framework.getFrame(), node, time, menu.getFunctionName(function), experiment);
return node;
} finally {
if (algorithm != null) {
algorithm.removeAlgorithmListener(listener);
}
if (logger != null) {
logger.dispose();
}
}
}