net.maizegenetics.analysis.association.WeightedMLMPlugin Maven / Gradle / Ivy
package net.maizegenetics.analysis.association;
import java.awt.BorderLayout;
import java.awt.Container;
import java.awt.Dimension;
import java.awt.Frame;
import java.awt.GridBagConstraints;
import java.awt.GridBagLayout;
import java.awt.Insets;
import java.awt.event.ActionEvent;
import java.awt.event.ActionListener;
import java.net.URL;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import javax.swing.BorderFactory;
import javax.swing.Box;
import javax.swing.ButtonGroup;
import javax.swing.ImageIcon;
import javax.swing.JButton;
import javax.swing.JDialog;
import javax.swing.JOptionPane;
import javax.swing.JPanel;
import javax.swing.JRadioButton;
import javax.swing.JTextField;
import net.maizegenetics.dna.snp.GenotypeTable;
import net.maizegenetics.gui.ReportDestinationDialog;
import net.maizegenetics.phenotype.GenotypePhenotype;
import net.maizegenetics.phenotype.Phenotype;
import net.maizegenetics.plugindef.DataSet;
import net.maizegenetics.plugindef.Datum;
import net.maizegenetics.taxa.distance.DistanceMatrix;
import org.apache.logging.log4j.LogManager;
import org.apache.logging.log4j.Logger;
public class WeightedMLMPlugin extends MLMPlugin{
public boolean isUseP3D() {
return useP3D;
}
public void setUseP3D(boolean useP3D) {
this.useP3D = useP3D;
}
public boolean isUseGenotype() {
return useGenotype;
}
public void setUseGenotype(boolean useGenotype) {
this.useGenotype = useGenotype;
}
public boolean isUseRefProb() {
return useRefProb;
}
public void setUseRefProb(boolean useRefProb) {
this.useRefProb = useRefProb;
}
private static final Logger myLogger = LogManager.getLogger(WeightedMLMPlugin.class);
protected DistanceMatrix kinshipMatrix;
protected boolean analyzeByColumn;
protected boolean useP3D = true;
protected CompressionType compressionType = CompressionType.None;
protected double compression = 1;
protected List weightList;
private boolean writeOutputToFile = false;
private String outputName = null;
private boolean filterOutput = false;
private double maxp = 1;
private boolean useGenotype = true;
private boolean useRefProb = false;
private boolean useAlleleProb = false;
public enum CompressionType {
Optimum, Custom, None
};
public WeightedMLMPlugin(Frame parentFrame, boolean isInteractive) {
super(parentFrame, isInteractive);
}
@Override
public DataSet processData(DataSet input) {
try {
java.util.List alignInList = input.getDataOfType(GenotypePhenotype.class);
if (alignInList.size() == 0) {
String message = "Invalid selection. Please select a GenotypePhenotype file";
if (isInteractive()) {
JOptionPane.showMessageDialog(getParentFrame(), message);
} else {
myLogger.error("performFunction: " + message);
}
return null;
//alignInList = input.getDataOfType(Phenotype.class);
}
this.weightList = input.getDataOfType(Phenotype.class);
java.util.List kinshipList = input.getDataOfType(DistanceMatrix.class);
if (alignInList.size() != 1) {
String message = "Invalid selection. Please select one dataset with marker and trait data.";
if (isInteractive()) {
JOptionPane.showMessageDialog(getParentFrame(), message);
} else {
myLogger.error("performFunction: " + message);
}
return null;
}
if (kinshipList.size() != 1) {
String message = "Please select exactly one kinship matrix.";
if (isInteractive()) {
JOptionPane.showMessageDialog(getParentFrame(), message);
} else {
myLogger.error("performFunction: " + message);
}
return null;
}
if(weightList.size()!= 1 && weightList.size()!=0) {
String message = "Please select exactly one weight matrix.";
if (isInteractive()) {
JOptionPane.showMessageDialog(getParentFrame(), message);
} else {
myLogger.error("performFunction: " + message);
}
return null;
}
if(weightList.size()!=0&&isUseP3D()==false) {
String message = "If You are using a Weight Matrix, you must use P3D.";
if (isInteractive()) {
JOptionPane.showMessageDialog(getParentFrame(), message);
} else {
myLogger.error("performFunction: " + message);
}
return null;
}
kinshipMatrix = (DistanceMatrix) kinshipList.get(0).getData();
Iterator itr = alignInList.iterator();
if (isInteractive()) {
GenotypePhenotype gp = (GenotypePhenotype) alignInList.get(0).getData();
WeightedMLMOptionDialog theOptions = new WeightedMLMOptionDialog(getParentFrame(), hasDataTypes(gp),weightList.size());
if (theOptions.runClicked) {
useP3D = theOptions.useP3D();
compressionType = theOptions.getCompressionType();
compression = theOptions.getCompressionLevel();
theOptions.dispose();
// give the user the option of sending the output to a file
ReportDestinationDialog rdd = new ReportDestinationDialog();
rdd.setLocationRelativeTo(getParentFrame());
rdd.setVisible(true);
if (!rdd.isOkayChecked()) {
return null;
}
writeOutputToFile = rdd.wasUseFileChecked();
if (writeOutputToFile) {
outputName = rdd.getOutputFileName();
}
filterOutput = rdd.wasRestrictOutputChecked();
if (filterOutput) {
maxp = rdd.getMaxP();
}
} else {
theOptions.dispose();
return null;
}
} else { //non-interactive stuff
}
List myResults = new ArrayList();
while (itr.hasNext()) {
Datum current = itr.next();
CompressedMLMusingDoubleMatrix theAnalysis;
GenotypeTable myGenotype = ((GenotypePhenotype) current.getData()).genotypeTable();
useGenotype = myGenotype.hasGenotype();
if (!useGenotype) useRefProb = myGenotype.hasReferenceProbablity();
if (useP3D) {
if (compressionType.equals(CompressionType.Optimum)) {
if(weightList.size()==0) {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix,true, true, Double.NaN);
}
else {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, weightList.get(0),true, true, Double.NaN);
}
} else if (compressionType.equals(CompressionType.Custom)) {
if(weightList.size()==0) {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, true, true, compression);
}
else {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, weightList.get(0),true, true, compression);
}
} else {
if(weightList.size()==0) {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, false, true, Double.NaN);
}
else {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, weightList.get(0),false, true, Double.NaN);
}
}
theAnalysis.useGenotypeCalls(useGenotype);
theAnalysis.useReferenceProbability(useRefProb);
theAnalysis.useAlleleProbabilities(useAlleleProb);
} else {
if (compressionType.equals(CompressionType.Optimum)) {
if(weightList.size()==0) {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, true, false, Double.NaN);
}
else {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, weightList.get(0),true, false, Double.NaN);
}
} else if (compressionType.equals(CompressionType.Custom)) {
if(weightList.size()==0) {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, true, false, compression);
}
else {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, weightList.get(0),true, false, compression);
}
} else {
if(weightList.size()==0) {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, false, false, Double.NaN);
}
else {
theAnalysis = new CompressedMLMusingDoubleMatrix(this, current, kinshipMatrix, weightList.get(0),false, false, Double.NaN);
}
}
theAnalysis.useGenotypeCalls(useGenotype);
theAnalysis.useReferenceProbability(useRefProb);
theAnalysis.useAlleleProbabilities(useAlleleProb);
}
myResults.addAll(theAnalysis.solve());
}
if (myResults.size() > 0) {
return new DataSet(myResults, this);
}
else return null;
} finally {
fireProgress(100);
}
}
private boolean[] hasDataTypes(GenotypePhenotype gp) {
boolean[] hasTypes = new boolean[]{false, false, false};
if (gp.genotypeTable().hasGenotype()) hasTypes[0] = true;
if (gp.genotypeTable().hasReference()) hasTypes[1] = true;
if (gp.genotypeTable().hasAlleleProbabilities()) hasTypes[2] = true;
return hasTypes;
}
public ImageIcon getIcon() {
URL imageURL = WeightedMLMPlugin.class.getResource("/net/maizegenetics/analysis/images/Mix.gif");
if (imageURL == null) {
return null;
} else {
return new ImageIcon(imageURL);
}
}
public String getButtonName() {
return "WeightedMLM";
}
public String getToolTipText() {
return "Association analysis using mixed model";
}
//a few stub functions to avoid producing errors in existing pipeline code
public void setAnalyzeByColumn(boolean analyzeByColumn) {
this.analyzeByColumn = analyzeByColumn;
}
public void setMaximumNumOfIteration(int max) {/*does nothing*/
}
public void setFinalIterMarker(boolean myFinalIterMarker) {/*does nothing*/
}
public void addFactors(int[] factors) {/*does nothing*/
}
public void setColumnTypes(String[] types) {/*does nothing*/
}
public void addFactors(String[] names) {/*does nothing*/
}
public void updateProgress(int progress) {
if (progress < 0) {
progress = 0;
} else if (progress > 100) {
progress = 100;
}
fireProgress(progress);
}
public void setVarCompEst(String value) {
if (value.equalsIgnoreCase("P3D")) {
useP3D = true;
} else if (value.equalsIgnoreCase("EachMarker")) {
useP3D = false;
} else {
throw new IllegalArgumentException("MLMPlugin: setVarCompEst: don't know how to handle value: " + value);
}
}
public void setCompressionType(CompressionType type) {
compressionType = type;
}
public boolean isWriteOutputToFile() {
return writeOutputToFile;
}
public void setWriteOutputToFile(boolean writeOutputToFile) {
this.writeOutputToFile = writeOutputToFile;
}
public String getOutputName() {
return outputName;
}
public void setOutputName(String outputName) {
this.outputName = outputName;
this.writeOutputToFile = true;
}
public boolean isFilterOutput() {
return filterOutput;
}
public void setFilterOutput(boolean filterOutput) {
this.filterOutput = filterOutput;
}
public double getMaxp() {
return maxp;
}
public void setMaxp(double maxp) {
this.maxp = maxp;
this.filterOutput = true;
}
public double getCustomCompression() {
return compression;
}
public void setCustomCompression(double value) {
compression = value;
}
public void useGenotypeCalls() {
useGenotype = true;
useRefProb = false;
useAlleleProb = false;
}
public void useReferenceProbability() {
useGenotype = false;
useRefProb = true;
useAlleleProb = false;
}
public void useAlleleProbabilities() {
useGenotype = false;
useRefProb = false;
useAlleleProb = true;
}
@Override
public String getCitation() {
return "Shang Xue, Zachary Miller, Janu Verma, " +
" First Annual Tassel Hackathon";
}
}
class WeightedMLMOptionDialog extends JDialog implements ActionListener {
JRadioButton btnOptimum, btnCustom, btnNoCompression, btnEachMarker, btnP3D;
ButtonGroup bgCompress, bgVariance, bgType;
JTextField txtCustom;
boolean runClicked = false;
boolean useP3D = true;
WeightedMLMPlugin.CompressionType compressionType = WeightedMLMPlugin.CompressionType.None;
boolean useDiscrete = true;
boolean useRefprob = false;
boolean useAlleleprob = false;
int weightedSize;
WeightedMLMOptionDialog(Frame parentFrame, boolean[] hasTypes,int weightedSize) {
super(parentFrame, true);
this.weightedSize = weightedSize;
final Frame pframe = parentFrame;
setTitle("MLM Options");
setSize(new Dimension(350, 300));
setLocationRelativeTo(pframe);
Container theContentPane = getContentPane();
theContentPane.setLayout(new BorderLayout());
JPanel compressionPanel = new JPanel(new GridBagLayout());
compressionPanel.setBorder(BorderFactory.createTitledBorder("Compression Level"));
//the method radio buttons
btnOptimum = new JRadioButton("Optimum Level", false);
btnOptimum.setActionCommand("Optimum");
btnOptimum.addActionListener(this);
btnCustom = new JRadioButton("Custom Level:", false);
btnCustom.setActionCommand("Custom");
btnCustom.addActionListener(this);
btnNoCompression = new JRadioButton("No Compression", true);
btnNoCompression.setActionCommand("None");
btnNoCompression.addActionListener(this);
if(weightedSize==0) {
btnEachMarker = new JRadioButton("Re-estimate after each marker", false);
btnEachMarker.setActionCommand("Eachmarker");
btnEachMarker.addActionListener(this);
}
btnP3D = new JRadioButton("P3D (estimate once)", true);
btnP3D.setActionCommand("P3D");
btnP3D.addActionListener(this);
bgCompress = new ButtonGroup();
bgCompress.add(btnOptimum);
bgCompress.add(btnCustom);
bgCompress.add(btnNoCompression);
bgVariance = new ButtonGroup();
if(weightedSize==0) {
bgVariance.add(btnEachMarker);
}
bgVariance.add(btnP3D);
txtCustom = new JTextField(5);
Insets inset1 = new Insets(5, 15, 5, 5);
Insets inset2 = new Insets(5, 5, 5, 5);
GridBagConstraints gbc = new GridBagConstraints();
gbc.gridx = 0;
gbc.gridwidth = 2;
gbc.gridy = 0;
gbc.weightx = 0;
gbc.weighty = 0;
gbc.anchor = GridBagConstraints.WEST;
gbc.insets = inset1; //top, left, bottom, right
compressionPanel.add(btnOptimum, gbc);
gbc.gridy++;
gbc.gridwidth = 1;
compressionPanel.add(btnCustom, gbc);
gbc.gridx++;
gbc.insets = inset2;
compressionPanel.add(txtCustom, gbc);
gbc.gridy++;
gbc.gridx = 0;
gbc.gridwidth = 2;
gbc.insets = inset1;
compressionPanel.add(btnNoCompression, gbc);
theContentPane.add(compressionPanel, BorderLayout.NORTH);
JPanel variancePanel = new JPanel(new GridBagLayout());
variancePanel.setBorder(BorderFactory.createTitledBorder("Variance Component Estimation"));
gbc.gridy = 0;
variancePanel.add(btnP3D, gbc);
gbc.gridy++;
if(weightedSize==0) {
variancePanel.add(btnEachMarker, gbc);
}
theContentPane.add(variancePanel, BorderLayout.CENTER);
//panel for choosing a data type, if there is more than one
int numberOfTypes = 0;
for (boolean b : hasTypes) if (b) numberOfTypes++;
if (numberOfTypes > 1) {
bgType = new ButtonGroup();
JPanel typePanel = new JPanel(new GridBagLayout());
typePanel.setBorder(BorderFactory.createTitledBorder("Choose genotype data"));
gbc.gridy = 0;
boolean initialValue = true;
if (hasTypes[0]) {
JRadioButton btnDiscrete = new JRadioButton("Discrete type, e.g. SNP", initialValue);
btnDiscrete.setActionCommand("discrete");
btnDiscrete.addActionListener(this);
initialValue = false;
bgType.add(btnDiscrete);
typePanel.add(btnDiscrete);
gbc.gridy++;
}
if (hasTypes[1]) {
JRadioButton btnRef = new JRadioButton("Numeric genotype", initialValue);
btnRef.setActionCommand("reference");
btnRef.addActionListener(this);
initialValue = false;
bgType.add(btnRef);
typePanel.add(btnRef);
gbc.gridy++;
}
if (hasTypes[2]) {
JRadioButton btnAllele = new JRadioButton("Allele probabilities", initialValue);
btnAllele.setActionCommand("allele");
btnAllele.addActionListener(this);
initialValue = false;
bgType.add(btnAllele);
typePanel.add(btnAllele);
gbc.gridy++;
}
}
//the help me button
JButton btnHelpme = new JButton("Help Me Choose");
final String msg = "For faster analysis, impute marker values before running MLM and use P3D.\n"
+ "With imputed marker values (no missing data), P3D will be very fast but compression will actually increase execution time somewhat.\n"
+ "However, because compression will improve the overall model fit, it should still be used.\n"
+ "If there is missing marker data, compression with P3D will probably be faster than P3D alone.\n"
+ "With small to moderate sized data sets any of the methods should give reasonable performance. \n"
+ "With large data sets consider imputing marker data then using EMMA with both compression and P3D";
btnHelpme.addActionListener(new ActionListener() {
@Override
public void actionPerformed(ActionEvent arg0) {
JOptionPane.showMessageDialog(pframe, msg, "EMMA methods", JOptionPane.INFORMATION_MESSAGE);
}
});
//the run and cancel buttons
JButton btnRun = new JButton("Run");
JButton btnCancel = new JButton("Cancel");
btnRun.setActionCommand("run");
btnRun.addActionListener(this);
btnCancel.setActionCommand("cancel");
btnCancel.addActionListener(this);
Box buttonBox = Box.createHorizontalBox();
buttonBox.add(Box.createGlue());
buttonBox.add(btnRun);
buttonBox.add(Box.createHorizontalStrut(50));
buttonBox.add(btnCancel);
buttonBox.add(Box.createHorizontalStrut(50));
buttonBox.add(btnHelpme);
buttonBox.add(Box.createGlue());
theContentPane.add(buttonBox, BorderLayout.SOUTH);
this.pack();
setLocationRelativeTo(getParent());
this.setVisible(true);
}
public boolean useP3D() {
return useP3D;
}
public WeightedMLMPlugin.CompressionType getCompressionType() {
return compressionType;
}
public boolean useDiscrete() { return useDiscrete; }
public boolean useRefProb() { return useRefprob; }
public boolean useAlleleProb() { return useAlleleprob; }
double getCompressionLevel() {
double comp;
try {
comp = Double.parseDouble(txtCustom.getText());
} catch (Exception e) {
comp = Double.NaN;
}
return comp;
}
@Override
public void actionPerformed(ActionEvent e) {
if (e.getActionCommand().equals("run")) {
runClicked = true;
this.setVisible(false);
} else if (e.getActionCommand().equals("cancel")) {
this.setVisible(false);
} else if (e.getActionCommand().equals("Optimum")) {
compressionType = WeightedMLMPlugin.CompressionType.Optimum;
} else if (e.getActionCommand().equals("Custom")) {
compressionType = WeightedMLMPlugin.CompressionType.Custom;
} else if (e.getActionCommand().equals("None")) {
compressionType = WeightedMLMPlugin.CompressionType.None;
} else if (e.getActionCommand().equals("Eachmarker")) {
useP3D = false;
} else if (e.getActionCommand().equals("P3D")) {
useP3D = true;
} else if (e.getActionCommand().equals("discrete")) {
useDiscrete = true;
useRefprob = false;
useAlleleprob = false;
} else if (e.getActionCommand().equals("reference")) {
useDiscrete = false;
useRefprob = true;
useAlleleprob = false;
} else if (e.getActionCommand().equals("allele")) {
useDiscrete = false;
useRefprob = false;
useAlleleprob = true;
}
}
public static void main(String[] args) {
WeightedMLMOptionDialog mod = new WeightedMLMOptionDialog(null, new boolean[]{true, true, true},0);
mod.dispose();
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy