Search in sources :

Example 1 with ProbeGroupGenerator

use of uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeGroupGenerator in project SeqMonk by s-andrews.

the class WindowedDifferencesFilter method generateProbeList.

/* (non-Javadoc)
	 * @see uk.ac.babraham.SeqMonk.Filters.ProbeFilter#generateProbeList()
	 */
@Override
protected void generateProbeList() {
    // We need to check that we don't add any probes more than once
    // so we need to keep a hash of the probes we've added to the
    // filtered list.
    Hashtable<Probe, Float> goingToAdd = new Hashtable<Probe, Float>();
    ProbeList newList = new ProbeList(startingList, "Filtered Probes", "", "Difference");
    Chromosome[] chromosomes = collection.genome().getAllChromosomes();
    for (int c = 0; c < chromosomes.length; c++) {
        progressUpdated("Processing windows on Chr" + chromosomes[c].name(), c, chromosomes.length);
        Probe[] probes = startingList.getProbesForChromosome(chromosomes[c]);
        ProbeGroupGenerator gen = null;
        if (windowType == DISTANCE_WINDOW) {
            gen = new ProbeWindowGenerator(probes, windowSize);
        } else if (windowType == CONSECUTIVE_WINDOW) {
            gen = new ConsecutiveProbeGenerator(probes, windowSize);
        } else if (windowType == FEATURE_WINDOW) {
            gen = new FeatureProbeGroupGenerator(probes, collection.genome().annotationCollection().getFeaturesForType(optionPanel.featureTypeBox.getSelectedItem().toString()));
        }
        while (true) {
            if (cancel) {
                cancel = false;
                progressCancelled();
                return;
            }
            Probe[] theseProbes = gen.nextSet();
            if (theseProbes == null) {
                break;
            }
            int count = 0;
            float d = 0;
            for (int s1 = 0; s1 < fromStores.length; s1++) {
                for (int s2 = 0; s2 < toStores.length; s2++) {
                    switch(combineType) {
                        case DifferencesFilter.AVERAGE:
                            d += getDifferenceValue(toStores[s2], fromStores[s1], theseProbes);
                            count++;
                            break;
                        case DifferencesFilter.MAXIMUM:
                            float dt1 = getDifferenceValue(toStores[s2], fromStores[s1], theseProbes);
                            if (count == 0 || dt1 > d)
                                d = dt1;
                            count++;
                            break;
                        case DifferencesFilter.MINIMUM:
                            float dt2 = getDifferenceValue(toStores[s2], fromStores[s1], theseProbes);
                            if (count == 0 || dt2 < d)
                                d = dt2;
                            count++;
                            break;
                    }
                }
            }
            if (combineType == DifferencesFilter.AVERAGE) {
                d /= count;
            }
            // Now we have the value we need to know if it passes the test
            if (upperLimit != null)
                if (d > upperLimit.doubleValue())
                    continue;
            if (lowerLimit != null)
                if (d < lowerLimit.doubleValue())
                    continue;
            for (int i = 0; i < theseProbes.length; i++) {
                if (goingToAdd.containsKey(theseProbes[i])) {
                    // Don't do anything if this probe is already there with a bigger difference
                    continue;
                // if (Math.abs(goingToAdd.get(theseProbes[i])) > Math.abs(d)) continue;
                }
                goingToAdd.put(theseProbes[i], d);
            }
        }
    }
    // Finally add all of the cached probes to the actual probe list
    Enumeration<Probe> en = goingToAdd.keys();
    while (en.hasMoreElements()) {
        Probe p = en.nextElement();
        newList.addProbe(p, goingToAdd.get(p));
    }
    filterFinished(newList);
}
Also used : ProbeList(uk.ac.babraham.SeqMonk.DataTypes.Probes.ProbeList) Hashtable(java.util.Hashtable) ProbeWindowGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeWindowGenerator) Chromosome(uk.ac.babraham.SeqMonk.DataTypes.Genome.Chromosome) Probe(uk.ac.babraham.SeqMonk.DataTypes.Probes.Probe) FeatureProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.FeatureProbeGroupGenerator) ConsecutiveProbeGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ConsecutiveProbeGenerator) ProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeGroupGenerator) FeatureProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.FeatureProbeGroupGenerator)

Example 2 with ProbeGroupGenerator

use of uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeGroupGenerator in project SeqMonk by s-andrews.

the class WindowedReplicateStatsFilter method generateProbeList.

/* (non-Javadoc)
	 * @see uk.ac.babraham.SeqMonk.Filters.ProbeFilter#generateProbeList()
	 */
@Override
protected void generateProbeList() {
    Chromosome[] chromosomes = collection.genome().getAllChromosomes();
    Vector<ProbeGroupTTestValue> newListProbesVector = new Vector<ProbeGroupTTestValue>();
    for (int c = 0; c < chromosomes.length; c++) {
        progressUpdated("Processing windows on Chr" + chromosomes[c].name(), c, chromosomes.length);
        Probe[] probes = startingList.getProbesForChromosome(chromosomes[c]);
        ProbeGroupGenerator gen = null;
        if (windowType == DISTANCE_WINDOW) {
            gen = new ProbeWindowGenerator(probes, windowSize);
        } else if (windowType == CONSECUTIVE_WINDOW) {
            gen = new ConsecutiveProbeGenerator(probes, windowSize);
        } else if (windowType == FEATURE_WINDOW) {
            gen = new FeatureProbeGroupGenerator(probes, collection.genome().annotationCollection().getFeaturesForType(optionsPanel.featureTypeBox.getSelectedItem().toString()));
        }
        while (true) {
            if (cancel) {
                cancel = false;
                progressCancelled();
                return;
            }
            Probe[] theseProbes = gen.nextSet();
            if (theseProbes == null) {
                // System.err.println("List of probes was null");
                break;
            }
            // We need at least 3 probes in a set to calculate a p-value
            if (theseProbes.length < 3) {
                // System.err.println("Only "+theseProbes.length+" probes in the set");
                continue;
            }
            double[][] values = new double[stores.length][];
            for (int j = 0; j < stores.length; j++) {
                if (splitReplicateSets & stores[j] instanceof ReplicateSet) {
                    values[j] = new double[theseProbes.length * ((ReplicateSet) stores[j]).dataStores().length];
                } else {
                    values[j] = new double[theseProbes.length];
                }
            }
            for (int j = 0; j < stores.length; j++) {
                int index = 0;
                for (int i = 0; i < theseProbes.length; i++) {
                    try {
                        if (splitReplicateSets & stores[j] instanceof ReplicateSet) {
                            DataStore[] localStores = ((ReplicateSet) stores[j]).dataStores();
                            for (int l = 0; l < localStores.length; l++) {
                                values[j][index] = localStores[l].getValueForProbe(theseProbes[i]);
                                index++;
                            }
                        } else {
                            values[j][index] = stores[j].getValueForProbe(theseProbes[i]);
                            index++;
                        }
                    } catch (SeqMonkException e) {
                    }
                }
                if (index != values[j].length) {
                    throw new IllegalStateException("Didn't fill all values total=" + values[j].length + " index=" + index);
                }
            }
            double pValue = 0;
            try {
                if (stores.length == 1) {
                    pValue = TTest.calculatePValue(values[0], 0);
                } else if (stores.length == 2) {
                    pValue = TTest.calculatePValue(values[0], values[1]);
                } else {
                    pValue = AnovaTest.calculatePValue(values);
                }
            } catch (SeqMonkException e) {
                throw new IllegalStateException(e);
            }
            newListProbesVector.add(new ProbeGroupTTestValue(theseProbes, pValue));
        }
    }
    ProbeGroupTTestValue[] newListProbes = newListProbesVector.toArray(new ProbeGroupTTestValue[0]);
    // Do the multi-testing correction if necessary
    if (multiTest) {
        BenjHochFDR.calculateQValues(newListProbes);
    }
    ProbeList newList;
    // We need to handle duplicate hits internally since probe lists can't do
    // this themselves any more.
    Hashtable<Probe, Float> newListTemp = new Hashtable<Probe, Float>();
    if (multiTest) {
        newList = new ProbeList(startingList, "", "", "Q-value");
        for (int i = 0; i < newListProbes.length; i++) {
            if (newListProbes[i].q <= cutoff) {
                Probe[] passedProbes = newListProbes[i].probes;
                for (int p = 0; p < passedProbes.length; p++) {
                    if (newListTemp.containsKey(passedProbes[p])) {
                        // We always give a probe the lowest possible q-value
                        if (newListTemp.get(passedProbes[p]) <= newListProbes[i].q) {
                            continue;
                        }
                    }
                    newListTemp.put(passedProbes[p], (float) newListProbes[i].q);
                }
            }
        }
    } else {
        newList = new ProbeList(startingList, "", "", "P-value");
        for (int i = 0; i < newListProbes.length; i++) {
            if (newListProbes[i].p <= cutoff) {
                Probe[] passedProbes = newListProbes[i].probes;
                for (int p = 0; p < passedProbes.length; p++) {
                    if (newListTemp.containsKey(passedProbes[p])) {
                        // We always give a probe the lowest possible p-value
                        if (newListTemp.get(passedProbes[p]) <= newListProbes[i].p) {
                            continue;
                        }
                    }
                    newListTemp.put(passedProbes[p], (float) newListProbes[i].p);
                }
            }
        }
    }
    // Add the cached hits to the new list
    Enumeration<Probe> en = newListTemp.keys();
    while (en.hasMoreElements()) {
        Probe p = en.nextElement();
        newList.addProbe(p, newListTemp.get(p));
    }
    filterFinished(newList);
}
Also used : ReplicateSet(uk.ac.babraham.SeqMonk.DataTypes.ReplicateSet) ProbeWindowGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeWindowGenerator) Probe(uk.ac.babraham.SeqMonk.DataTypes.Probes.Probe) FeatureProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.FeatureProbeGroupGenerator) DataStore(uk.ac.babraham.SeqMonk.DataTypes.DataStore) SeqMonkException(uk.ac.babraham.SeqMonk.SeqMonkException) Vector(java.util.Vector) ProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeGroupGenerator) FeatureProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.FeatureProbeGroupGenerator) ProbeList(uk.ac.babraham.SeqMonk.DataTypes.Probes.ProbeList) Hashtable(java.util.Hashtable) Chromosome(uk.ac.babraham.SeqMonk.DataTypes.Genome.Chromosome) ProbeGroupTTestValue(uk.ac.babraham.SeqMonk.Analysis.Statistics.ProbeGroupTTestValue) ConsecutiveProbeGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ConsecutiveProbeGenerator)

Example 3 with ProbeGroupGenerator

use of uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeGroupGenerator in project SeqMonk by s-andrews.

the class WindowedValuesFilter method generateProbeList.

/* (non-Javadoc)
	 * @see uk.ac.babraham.SeqMonk.Filters.ProbeFilter#generateProbeList()
	 */
@Override
protected void generateProbeList() {
    ProbeList newList = new ProbeList(startingList, "Filtered Probes", "", null);
    Chromosome[] chromosomes = collection.genome().getAllChromosomes();
    for (int c = 0; c < chromosomes.length; c++) {
        HashSet<Probe> passedProbes = new HashSet<Probe>();
        progressUpdated("Processing windows on Chr" + chromosomes[c].name(), c, chromosomes.length);
        Probe[] probes = startingList.getProbesForChromosome(chromosomes[c]);
        ProbeGroupGenerator gen = null;
        if (windowType == DISTANCE_WINDOW) {
            gen = new ProbeWindowGenerator(probes, windowSize);
        } else if (windowType == CONSECUTIVE_WINDOW) {
            gen = new ConsecutiveProbeGenerator(probes, windowSize);
        } else if (windowType == FEATURE_WINDOW) {
            gen = new FeatureProbeGroupGenerator(probes, collection.genome().annotationCollection().getFeaturesForType(optionsPanel.featureTypeBox.getSelectedItem().toString()));
        } else {
            progressExceptionReceived(new SeqMonkException("No window type known with number " + windowType));
            return;
        }
        while (true) {
            if (cancel) {
                cancel = false;
                progressCancelled();
                return;
            }
            Probe[] theseProbes = gen.nextSet();
            if (theseProbes == null) {
                break;
            }
            int count = 0;
            for (int s = 0; s < stores.length; s++) {
                double totalValue = 0;
                for (int i = 0; i < theseProbes.length; i++) {
                    // Get the values for the probes in this set
                    try {
                        totalValue += stores[s].getValueForProbe(theseProbes[i]);
                    } catch (SeqMonkException e) {
                    }
                }
                totalValue /= theseProbes.length;
                // Now we have the value we need to know if it passes the test
                if (upperLimit != null)
                    if (totalValue > upperLimit.doubleValue())
                        continue;
                if (lowerLimit != null)
                    if (totalValue < lowerLimit.doubleValue())
                        continue;
                // This one passes, we can add it to the count
                ++count;
            }
            // probe to the probe set.
            switch(limitType) {
                case EXACTLY:
                    if (count == storesLimit)
                        for (int i = 0; i < theseProbes.length; i++) {
                            if (passedProbes.contains(theseProbes[i]))
                                continue;
                            newList.addProbe(theseProbes[i], null);
                            passedProbes.add(theseProbes[i]);
                        }
                    break;
                case AT_LEAST:
                    if (count >= storesLimit)
                        for (int i = 0; i < theseProbes.length; i++) {
                            if (passedProbes.contains(theseProbes[i]))
                                continue;
                            newList.addProbe(theseProbes[i], null);
                            passedProbes.add(theseProbes[i]);
                        }
                    break;
                case NO_MORE_THAN:
                    if (count <= storesLimit)
                        for (int i = 0; i < theseProbes.length; i++) {
                            if (passedProbes.contains(theseProbes[i]))
                                continue;
                            newList.addProbe(theseProbes[i], null);
                            passedProbes.add(theseProbes[i]);
                        }
                    break;
            }
        }
    }
    filterFinished(newList);
}
Also used : ProbeList(uk.ac.babraham.SeqMonk.DataTypes.Probes.ProbeList) ProbeWindowGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeWindowGenerator) Chromosome(uk.ac.babraham.SeqMonk.DataTypes.Genome.Chromosome) Probe(uk.ac.babraham.SeqMonk.DataTypes.Probes.Probe) FeatureProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.FeatureProbeGroupGenerator) ConsecutiveProbeGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ConsecutiveProbeGenerator) SeqMonkException(uk.ac.babraham.SeqMonk.SeqMonkException) ProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeGroupGenerator) FeatureProbeGroupGenerator(uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.FeatureProbeGroupGenerator) HashSet(java.util.HashSet)

Aggregations

Chromosome (uk.ac.babraham.SeqMonk.DataTypes.Genome.Chromosome)3 Probe (uk.ac.babraham.SeqMonk.DataTypes.Probes.Probe)3 ProbeList (uk.ac.babraham.SeqMonk.DataTypes.Probes.ProbeList)3 ConsecutiveProbeGenerator (uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ConsecutiveProbeGenerator)3 FeatureProbeGroupGenerator (uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.FeatureProbeGroupGenerator)3 ProbeGroupGenerator (uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeGroupGenerator)3 ProbeWindowGenerator (uk.ac.babraham.SeqMonk.Filters.ProbeGroupGenerator.ProbeWindowGenerator)3 Hashtable (java.util.Hashtable)2 SeqMonkException (uk.ac.babraham.SeqMonk.SeqMonkException)2 HashSet (java.util.HashSet)1 Vector (java.util.Vector)1 ProbeGroupTTestValue (uk.ac.babraham.SeqMonk.Analysis.Statistics.ProbeGroupTTestValue)1 DataStore (uk.ac.babraham.SeqMonk.DataTypes.DataStore)1 ReplicateSet (uk.ac.babraham.SeqMonk.DataTypes.ReplicateSet)1