zzh
/
gatk-3.8
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

small cleanups for the GATK paper genotyper; switched to the managed output system. 

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@2156 348d0f76-0448-11de-a6fe-93d51630548a
This commit is contained in:
aaron 2009-11-25 08:04:13 +00:00
parent e1e5b35b19
commit cfbd9332b0
3 changed files with 21 additions and 88 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

37
java/src/org/broadinstitute/sting/playground/gatk/walkers/papergenotyper/GATKPaperGenotyper.java
View File

@ -10,7 +10,7 @@ import org.broadinstitute.sting.utils.pileup.ReadBackedPileup;
import org.broadinstitute.sting.utils.Utils; import org.broadinstitute.sting.utils.Utils;
import org.broadinstitute.sting.utils.cmdLine.Argument; import org.broadinstitute.sting.utils.cmdLine.Argument;
import java.io.File; import java.io.PrintWriter;
/** /**
@ -21,13 +21,14 @@ import java.io.File;
* A simple Bayesian genotyper, that output a text based call format. Intended to be used only as an * A simple Bayesian genotyper, that output a text based call format. Intended to be used only as an
* example in the GATK publication. * example in the GATK publication.
*/ */
public class GATKPaperGenotyper extends LocusWalker<SimpleCall, SimpleCallList> implements TreeReducible<SimpleCallList> { public class GATKPaperGenotyper extends LocusWalker<SimpleCall, Integer> implements TreeReducible<Integer> {
// the possible diploid genotype strings // the possible diploid genotype strings
private static enum GENOTYPE { AA, AC, AG, AT, CC, CG, CT, GG, GT, TT } private static enum GENOTYPE { AA, AC, AG, AT, CC, CG, CT, GG, GT, TT }
// where to write the genotyping data to
@Argument(fullName = "call_location", shortName = "cl", doc = "File to which calls should be written", required = true) @Argument(fullName = "call_location", shortName = "cl", doc = "File to which calls should be written", required = true)
private File LOCATION = new File("genotyping.output"); public PrintWriter outputStream;
/** /**
* our map function, which takes the reads spanning this locus, any associated reference ordered data, * our map function, which takes the reads spanning this locus, any associated reference ordered data,
@ -45,12 +46,13 @@ public class GATKPaperGenotyper extends LocusWalker<SimpleCall, SimpleCallList>
double likelihoods[] = DiploidGenotypePriors.getReferencePolarizedPriors(ref.getBase(), double likelihoods[] = DiploidGenotypePriors.getReferencePolarizedPriors(ref.getBase(),
DiploidGenotypePriors.HUMAN_HETEROZYGOSITY, DiploidGenotypePriors.HUMAN_HETEROZYGOSITY,
DiploidGenotypePriors.PROB_OF_TRISTATE_GENOTYPE); DiploidGenotypePriors.PROB_OF_TRISTATE_GENOTYPE);
byte bases[] = pileup.getBases();
byte quals[] = pileup.getQuals();
for (GENOTYPE genotype : GENOTYPE.values()) for (GENOTYPE genotype : GENOTYPE.values())
for (int index = 0; index < pileup.getBases().length; index++) { for (int index = 0; index < bases.length; index++) {
if (pileup.getQuals()[index] > 0) { if (quals[index] > 0) {
double epsilon = Math.pow(10, pileup.getQuals()[index] / -10.0); double epsilon = Math.pow(10, quals[index] / -10.0);
byte pileupBase = pileup.getBases()[index]; byte pileupBase = bases[index];
for (char genotypeBase : genotype.toString().toCharArray()) for (char genotypeBase : genotype.toString().toCharArray())
if (genotypeBase == pileupBase) if (genotypeBase == pileupBase)
likelihoods[genotype.ordinal()] += Math.log10(0.5 * ((1 - epsilon) + epsilon / 3)); likelihoods[genotype.ordinal()] += Math.log10(0.5 * ((1 - epsilon) + epsilon / 3));
@ -76,8 +78,8 @@ public class GATKPaperGenotyper extends LocusWalker<SimpleCall, SimpleCallList>
* *
* @return Initial value of reduce. * @return Initial value of reduce.
*/ */
public SimpleCallList reduceInit() { public Integer reduceInit() {
return new SimpleCallList(LOCATION); return 0;
} }
/** /**
@ -87,9 +89,9 @@ public class GATKPaperGenotyper extends LocusWalker<SimpleCall, SimpleCallList>
* @param sum accumulator for the reduce. * @param sum accumulator for the reduce.
* @return accumulator with result of the map taken into account. * @return accumulator with result of the map taken into account.
*/ */
public SimpleCallList reduce(SimpleCall value, SimpleCallList sum) { public Integer reduce(SimpleCall value, Integer sum) {
if (value != null) sum.add(value); outputStream.println(value.toString());
return sum; return sum + 1;
} }
/** /**
@ -99,17 +101,16 @@ public class GATKPaperGenotyper extends LocusWalker<SimpleCall, SimpleCallList>
* @param rhs 'right-most' portion of data in the composite reduce. * @param rhs 'right-most' portion of data in the composite reduce.
* @return The composite reduce type. * @return The composite reduce type.
*/ */
public SimpleCallList treeReduce(SimpleCallList lhs, SimpleCallList rhs) { public Integer treeReduce(Integer lhs, Integer rhs) {
lhs.addAll(rhs); return lhs + rhs;
return lhs;
} }
/** /**
* when we finish traversing, close the result list * when we finish traversing, close the result list
* @param result the final reduce result * @param result the final reduce result
*/ */
public void onTraversalDone(SimpleCallList result) { public void onTraversalDone(Integer result) {
result.close(); out.println("Simple Genotyper genotyped " + result + "Loci.");
} }
} }

4
java/src/org/broadinstitute/sting/playground/gatk/walkers/papergenotyper/SimpleCall.java
View File

@ -8,7 +8,7 @@ import org.broadinstitute.sting.utils.GenomeLoc;
* Date: Nov 19, 2009 * Date: Nov 19, 2009
* Time: 2:07:25 AM * Time: 2:07:25 AM
* *
* This simple call class stores the data for the per-locus calls of the GATKPaperGenotyper. * This is a simple call class that stores the data for the per-locus calls of the GATKPaperGenotyper.
* *
*/ */
class SimpleCall { class SimpleCall {
@ -23,6 +23,6 @@ class SimpleCall {
} }
public String toString() { public String toString() {
return String.format("Location %s : %s with LOD %.2f", loc, genotype, LOD); return String.format("%s : %s with LOD %.4f", loc, genotype, LOD);
} }
} }

68
java/src/org/broadinstitute/sting/playground/gatk/walkers/papergenotyper/SimpleCallList.java
View File

@ -1,68 +0,0 @@
package org.broadinstitute.sting.playground.gatk.walkers.papergenotyper;
import org.broadinstitute.sting.utils.StingException;
import java.io.*;
import java.util.AbstractList;
import java.util.ArrayList;
import java.util.Collection;
/**
* Created by IntelliJ IDEA.
* User: aaronmckenna
* Date: Nov 19, 2009
* Time: 12:50:20 AM
*
* A simple class, that dumps the records to disk when we've hit a threshold.
* This class makes the GATKPaperGenotyper much simpler to take in for the reader.
*/
class SimpleCallList extends AbstractList<SimpleCall> {
private File outputLocation;
private ArrayList<SimpleCall> list = new ArrayList<SimpleCall>();
private int WRITE_LIMIT = 100000;
public SimpleCallList(File writeTo) {
outputLocation = writeTo;
}
public boolean add(SimpleCall call) {
boolean added = list.add(call);
writeIfNeeded();
return added;
}
public boolean addAll(Collection<? extends SimpleCall> otherCalls) {
boolean added = list.addAll(otherCalls);
writeIfNeeded();
return added;
}
public void writeIfNeeded() {
synchronized(list) {
if (list.size() > WRITE_LIMIT) {
try {
PrintWriter writer = new PrintWriter(new FileWriter(outputLocation, true));
for (SimpleCall call : list) writer.println(call.toString());
writer.close();
} catch (IOException e) {
throw new StingException("Unable to write to file " + outputLocation);
}
list.clear();
}
}
}
@Override
public int size() {
return list.size();
}
@Override
public SimpleCall get(int index) {
return list.get(index);
}
public void close() {
WRITE_LIMIT = 0;
writeIfNeeded();
}
}