private void writeDifferences(
final List<VariantContext> source1Alleles, final List<VariantContext> source2Alleles) {
int currentIndex1 = 0, currentIndex2 = 0;
final int size1 = source1Alleles.size(), size2 = source2Alleles.size();
VariantContext current1 = source1Alleles.get(0);
VariantContext current2 = source2Alleles.get(0);
while (currentIndex1 < size1 || currentIndex2 < size2) {
if (current1 == null) {
writeOne(current2, source2, null);
currentIndex2++;
current2 = (currentIndex2 < size2 ? source2Alleles.get(currentIndex2) : null);
} else if (current2 == null) {
writeOne(current1, source1, null);
currentIndex1++;
current1 = (currentIndex1 < size1 ? source1Alleles.get(currentIndex1) : null);
} else {
final GenomeLoc loc1 = getToolkit().getGenomeLocParser().createGenomeLoc(current1);
final GenomeLoc loc2 = getToolkit().getGenomeLocParser().createGenomeLoc(current2);
if (loc1.getStart() == loc2.getStart() || loc1.overlapsP(loc2)) {
String status;
if (loc1.getStart() == loc2.getStart()) {
final String allele1 = current1.getAlternateAllele(0).getBaseString();
final String allele2 = current2.getAlternateAllele(0).getBaseString();
if (allele1.indexOf(allele2) != -1 || allele2.indexOf(allele1) != -1)
status = ONE_ALLELE_SUBSET_OF_OTHER_STATUS;
else status = SAME_START_DIFFERENT_ALLELES_STATUS;
} else {
status = OVERLAPPING_EVENTS_STATUS;
}
writeOne(current1, INTERSECTION_SET, status);
currentIndex1++;
currentIndex2++;
current1 = (currentIndex1 < size1 ? source1Alleles.get(currentIndex1) : null);
current2 = (currentIndex2 < size2 ? source2Alleles.get(currentIndex2) : null);
} else if (loc1.isBefore(loc2)) {
writeOne(current1, source1, null);
currentIndex1++;
current1 = (currentIndex1 < size1 ? source1Alleles.get(currentIndex1) : null);
} else {
writeOne(current2, source2, null);
currentIndex2++;
current2 = (currentIndex2 < size2 ? source2Alleles.get(currentIndex2) : null);
}
}
}
}
private byte[] generateHaplotype(
final List<VariantContext> sourceVCs, final ReferenceContext refContext) {
final StringBuilder sb = new StringBuilder();
final int startPos = refContext.getWindow().getStart();
int currentPos = startPos;
final byte[] reference = refContext.getBases();
for (final VariantContext vc : sourceVCs) {
// add any missing reference context
int vcStart = vc.getStart();
final int refAlleleLength = vc.getReference().length();
if (refAlleleLength
== vc.getEnd()
- vc.getStart()) // this is a deletion (whereas for other events the padding base
// isn't part of the position)
vcStart++;
while (currentPos < vcStart) sb.append((char) reference[currentPos++ - startPos]);
// add the alt allele
sb.append(vc.getAlternateAllele(0).getBaseString());
// skip the reference allele
currentPos += refAlleleLength;
}
// add any missing reference context
final int stopPos = refContext.getWindow().getStop();
while (currentPos < stopPos) sb.append((char) reference[currentPos++ - startPos]);
return sb.toString().getBytes();
}
@Test
public void shouldPreserveSymbolicAlleleCase() {
VCFFileReader reader =
new VCFFileReader(new File(VariantBaseTest.variantTestDataRoot + "breakpoint.vcf"), false);
VariantContext variant = reader.iterator().next();
reader.close();
// VCF v4.1 s1.4.5
// Tools processing VCF files are not required to preserve case in the allele String, except for
// IDs, which are case sensitive.
Assert.assertTrue(variant.getAlternateAllele(0).getDisplayString().contains("chr12"));
}
@Override
public void runCommand() {
logger.info("MergeVCFColumnsCommand");
/*
* Assumptions
* (1) Only two vcfs that are sorted with the same contig order
* (2) if contigs on it same order, then we will just skip that contig
* (3) No overlapping samples allowed
*
* Output:
* A vcf where intersecting sites are merged together and will only return biallelic markers
* the info field will be cleared
* the only GT FORMAT field will be there
*/
Collection<File> vcfs = applicationOptions.getVcfs();
String outfile = applicationOptions.getOutFile();
if (vcfs.size() != 2) {
throw new IllegalArgumentException("This function requires exactly two vcfs");
}
Iterator<File> vcfFileIter = vcfs.iterator();
File vcf1 = vcfFileIter.next();
File vcf2 = vcfFileIter.next();
VCFFileReader reader1 = new VCFFileReader(vcf1, false);
VCFFileReader reader2 = new VCFFileReader(vcf2, false);
Iterator<VariantContext> iter1 = reader1.iterator();
Iterator<VariantContext> iter2 = reader2.iterator();
VariantContextComparator comparator = new VariantContextComparator();
/*
* Merge headers
*/
VCFHeader header1 = reader1.getFileHeader();
VCFHeader header2 = reader2.getFileHeader();
List<String> samples1 = header1.getGenotypeSamples();
List<String> samples2 = header2.getGenotypeSamples();
List<String> mergedSamples = new ArrayList<>(samples1.size() + samples2.size());
mergedSamples.addAll(samples1);
mergedSamples.addAll(samples2);
// Validate that there are no duplicates
HashSet<String> sampleSet = new HashSet<String>();
for (String id : mergedSamples) {
if (sampleSet.contains(id)) {
throw new IllegalArgumentException("Duplicate id found: " + id);
} else {
sampleSet.add(id);
}
}
HashSet<VCFHeaderLine> meta = new HashSet<>();
meta.add(new VCFFormatHeaderLine("GT", 1, VCFHeaderLineType.String, "GT"));
meta.addAll(header1.getContigLines());
VCFHeader mergedHeader = new VCFHeader(meta, mergedSamples);
/*
* Create encoder
*/
VCFEncoder encoder = new VCFEncoder(mergedHeader, false, false);
BufferedWriter writer = null;
try {
if (outfile.endsWith(".gz")) {
BlockCompressedOutputStream outstream = new BlockCompressedOutputStream(new File(outfile));
writer = new BufferedWriter(new OutputStreamWriter(outstream));
} else {
writer =
Files.newBufferedWriter(
Paths.get(outfile),
Charset.defaultCharset(),
StandardOpenOption.CREATE,
StandardOpenOption.TRUNCATE_EXISTING);
}
/*
* Write header
*/
VCFHeaderWriter.writeHeader(writer, mergedHeader);
logger.info("Wrote header");
VariantContext previous1 = null;
VariantContext previous2 = null;
int count = 0;
int countFile1 = 0;
int countFile2 = 0;
boolean usePrevious1 = false;
boolean usePrevious2 = false;
while (iter1.hasNext() || iter2.hasNext()) {
if ((iter1.hasNext() || usePrevious1) && (iter2.hasNext() || usePrevious2)) {
VariantContext variant1 = null;
VariantContext variant2 = null;
// if(usePrevious1 == true && usePrevious2 == true &&
// comparator.compare(previous1,previous2) != 0) {
// //then skip both
// usePrevious1 = false;
// usePrevious2 = false;
// }
if (usePrevious1) {
variant1 = previous1;
} else {
variant1 = iter1.next();
countFile1++;
}
if (usePrevious2) {
variant2 = previous2;
} else {
variant2 = iter2.next();
countFile2++;
}
// check that variants are ordered correctly
if (previous1 != null
&& previous1 != variant1
&& comparator.compare(previous1, variant1) > 0) {
throw new IllegalStateException(
previous1.getContig()
+ ":"
+ previous1.getStart()
+ " > "
+ variant1.getContig()
+ ":"
+ variant1.getStart());
}
if (previous2 != null
&& previous2 != variant2
&& comparator.compare(previous2, variant2) > 0) {
throw new IllegalStateException(
previous2.getContig()
+ ":"
+ previous2.getStart()
+ " > "
+ variant2.getContig()
+ ":"
+ variant2.getStart());
}
int cmp = comparator.compare(variant1, variant2);
if (cmp < 0) {
// logger.info("Skipping VCF1: " + variant1.getContig() + ":" + variant1.getStart() +
// "\t" + variant1.getReference().toString() + "\t" + variant1.getAlternateAlleles());
if (usePrevious1 == true && usePrevious2 == true) {
// variant1 < variant2
// we need to go to next variant in vcf1
usePrevious1 = false;
}
usePrevious2 = true;
} else if (cmp > 0) {
if (usePrevious1 == true && usePrevious2 == true) {
// variant1 > variant2
// we need to go to next variant in vcf2
usePrevious2 = false;
}
usePrevious1 = true;
// logger.info("Skipping VCF2: " + variant2.getContig() + ":" + variant2.getStart() +
// "\t" + variant2.getReference().toString() + "\t" + variant2.getAlternateAlleles());
} else {
// they equal position
usePrevious1 = false;
usePrevious2 = false;
if (variant1.isBiallelic()
&& variant2.isBiallelic()
&& variant1.getReference().equals(variant2.getReference())
&& variant1.getAlternateAllele(0).equals(variant2.getAlternateAllele(0))) {
// TODO: Finish merging
// both variants are bialleleic and the reference and alternative alleles match
count++;
if (count % 10000 == 0) {
logger.info(count + " mergeable variants found");
}
VariantContext merged = VariantContextMerger.merge(variant1, variant2);
writer.write(encoder.encode(merged));
writer.write("\n");
} else {
// skip if they do not equal
// logger.info("Skipping: " + variant1.getContig() + ":" + variant1.getStart() +
// "\t" + variant1.getReference().toString() + "\t" +
// variant1.getAlternateAlleles());
// logger.info("Skipping: " + variant2.getContig() + ":" + variant2.getStart() +
// "\t" + variant2.getReference().toString() + "\t" +
// variant2.getAlternateAlleles());
}
}
previous1 = variant1;
previous2 = variant2;
} else if (iter1.hasNext()) {
// just skip remaining variants
VariantContext current = iter1.next();
countFile1++;
if (previous1 != null && current != null && comparator.compare(previous1, current) > 0) {
throw new IllegalStateException(
previous1.getContig()
+ ":"
+ previous1.getStart()
+ " > "
+ current.getContig()
+ ":"
+ current.getStart());
}
previous1 = current;
// logger.info("Skipping: " + previous1.getContig() + ":" + previous1.getStart() + "\t" +
// previous1.getReference().toString() + "\t" + previous1.getAlternateAlleles());
} else if (iter2.hasNext()) {
// just skip remaining variants
// fixed bug/ was iter1 changed to iter2
VariantContext current = iter2.next();
countFile2++;
if (previous2 != null && current != null && comparator.compare(previous2, current) > 0) {
throw new IllegalStateException(
previous2.getContig()
+ ":"
+ previous2.getStart()
+ " > "
+ current.getContig()
+ ":"
+ current.getStart());
}
previous2 = current;
// logger.info("Skipping: " + previous2.getContig() + ":" + previous2.getStart() + "\t" +
// previous2.getReference().toString() + "\t" + previous2.getAlternateAlleles());
} else {
throw new IllegalStateException("Error should not of reached this point");
}
}
reader1.close();
reader2.close();
logger.info(count + " merged variants");
logger.info(countFile1 + " variants in " + vcf1.getAbsolutePath());
logger.info(countFile2 + " variants in " + vcf2.getAbsolutePath());
} catch (Exception e) {
e.printStackTrace();
} finally {
if (writer != null) {
try {
logger.info("Flushing writer");
writer.close();
} catch (IOException e) {
// TODO Auto-generated catch block
e.printStackTrace();
}
}
}
logger.info("finished merging vcfs");
}
