/**
* Finds the adaptor boundary around the read and returns the first base inside the adaptor that
* is closest to the read boundary. If the read is in the positive strand, this is the first base
* after the end of the fragment (Picard calls it 'insert'), if the read is in the negative
* strand, this is the first base before the beginning of the fragment.
*
* <p>There are two cases we need to treat here:
*
* <p>1) Our read is in the reverse strand :
*
* <p><----------------------| * |--------------------->
*
* <p>in these cases, the adaptor boundary is at the mate start (minus one)
*
* <p>2) Our read is in the forward strand :
*
* <p>|----------------------> * <----------------------|
*
* <p>in these cases the adaptor boundary is at the start of the read plus the inferred insert
* size (plus one)
*
* @param read the read being tested for the adaptor boundary
* @return the reference coordinate for the adaptor boundary (effectively the first base IN the
* adaptor, closest to the read. NULL if the read is unmapped or the mate is mapped to another
* contig.
*/
public static Integer getAdaptorBoundary(final SAMRecord read) {
final int MAXIMUM_ADAPTOR_LENGTH = 8;
final int insertSize =
Math.abs(
read
.getInferredInsertSize()); // the inferred insert size can be negative if the mate
// is mapped before the read (so we take the absolute
// value)
if (insertSize == 0
|| read
.getReadUnmappedFlag()) // no adaptors in reads with mates in another chromosome or
// unmapped pairs
return null;
Integer
adaptorBoundary; // the reference coordinate for the adaptor boundary (effectively the first
// base IN the adaptor, closest to the read)
if (read.getReadNegativeStrandFlag())
adaptorBoundary = read.getMateAlignmentStart() - 1; // case 1 (see header)
else adaptorBoundary = read.getAlignmentStart() + insertSize + 1; // case 2 (see header)
if ((adaptorBoundary < read.getAlignmentStart() - MAXIMUM_ADAPTOR_LENGTH)
|| (adaptorBoundary > read.getAlignmentEnd() + MAXIMUM_ADAPTOR_LENGTH))
adaptorBoundary =
null; // we are being conservative by not allowing the adaptor boundary to go beyond what
// we belive is the maximum size of an adaptor
return adaptorBoundary;
}
private void collectReadData(final SAMRecord record, final ReferenceSequence ref) {
metrics.TOTAL_READS++;
readLengthHistogram.increment(record.getReadBases().length);
if (!record.getReadFailsVendorQualityCheckFlag()) {
metrics.PF_READS++;
if (isNoiseRead(record)) metrics.PF_NOISE_READS++;
if (record.getReadUnmappedFlag()) {
// If the read is unmapped see if it's adapter sequence
final byte[] readBases = record.getReadBases();
if (!(record instanceof BAMRecord)) StringUtil.toUpperCase(readBases);
if (isAdapterSequence(readBases)) {
this.adapterReads++;
}
} else if (doRefMetrics) {
metrics.PF_READS_ALIGNED++;
if (!record.getReadNegativeStrandFlag()) numPositiveStrand++;
if (record.getReadPairedFlag() && !record.getMateUnmappedFlag()) {
metrics.READS_ALIGNED_IN_PAIRS++;
// Check that both ends have mapq > minimum
final Integer mateMq = record.getIntegerAttribute("MQ");
if (mateMq == null
|| mateMq >= MAPPING_QUALITY_THRESOLD
&& record.getMappingQuality() >= MAPPING_QUALITY_THRESOLD) {
++this.chimerasDenominator;
// With both reads mapped we can see if this pair is chimeric
if (Math.abs(record.getInferredInsertSize()) > maxInsertSize
|| !record.getReferenceIndex().equals(record.getMateReferenceIndex())) {
++this.chimeras;
}
}
}
}
}
}
public static Strand valueOfSAMRecord(SAMRecord sr) {
return sr.getReadNegativeStrandFlag() ? STRAND_NEGATIVE : STRAND_POSITIVE;
}
@Override
public int doWork(String[] args) {
File refFile = null;
com.github.lindenb.jvarkit.util.cli.GetOpt getopt =
new com.github.lindenb.jvarkit.util.cli.GetOpt();
int c;
while ((c = getopt.getopt(args, "hvL:r:")) != -1) {
switch (c) {
case 'h':
printUsage();
return 0;
case 'v':
System.out.println(getVersion());
return 0;
case 'L':
getLogger().setLevel(java.util.logging.Level.parse(getopt.getOptArg()));
break;
case 'r':
refFile = new File(getopt.getOptArg());
break;
case ':':
System.err.println("Missing argument for option -" + getopt.getOptOpt());
return -1;
default:
System.err.println("Unknown option -" + getopt.getOptOpt());
return -1;
}
}
if (refFile == null) {
error("Undefined REF file");
return -1;
}
File bamFile = null;
if (getopt.getOptInd() + 1 != args.length) {
info("reading from stdin.");
} else {
bamFile = new File(args[getopt.getOptInd()]);
}
IndexedFastaSequenceFile indexedFastaSequenceFile = null;
SAMFileReader samFileReader = null;
try {
GenomicSequence genomicSequence = null;
indexedFastaSequenceFile = new IndexedFastaSequenceFile(refFile);
SAMFileReader.setDefaultValidationStringency(ValidationStringency.SILENT);
samFileReader = null;
if (bamFile == null) {
samFileReader = new SAMFileReader(System.in);
} else {
samFileReader = new SAMFileReader(bamFile);
}
XMLOutputFactory xmlfactory = XMLOutputFactory.newInstance();
XMLStreamWriter w = xmlfactory.createXMLStreamWriter(System.out, "UTF-8");
w.writeStartDocument("UTF-8", "1.0");
w.writeStartElement("sam");
w.writeComment(getProgramCommandLine());
w.writeAttribute("ref", (bamFile == null ? "stdin" : bamFile.getPath()));
w.writeAttribute("bam", args[1]);
SAMRecordIterator iter = samFileReader.iterator();
while (iter.hasNext()) {
SAMRecord rec = iter.next();
final byte readbases[] = rec.getReadBases();
w.writeStartElement("read");
w.writeStartElement("name");
w.writeCharacters(rec.getReadName());
w.writeEndElement();
w.writeStartElement("sequence");
w.writeCharacters(new String(readbases));
w.writeEndElement();
w.writeStartElement("flags");
w.writeAttribute("paired", String.valueOf(rec.getReadPairedFlag()));
w.writeAttribute(
"failsVendorQual", String.valueOf(rec.getReadFailsVendorQualityCheckFlag()));
w.writeAttribute("mapped", String.valueOf(!rec.getReadUnmappedFlag()));
w.writeAttribute("strand", (rec.getReadNegativeStrandFlag() ? "-" : "+"));
if (rec.getReadPairedFlag()) {
w.writeAttribute("mate-mapped", String.valueOf(!rec.getMateUnmappedFlag()));
w.writeAttribute("mate-strand", (rec.getMateNegativeStrandFlag() ? "-" : "+"));
w.writeAttribute("proper-pair", String.valueOf(rec.getProperPairFlag()));
}
w.writeCharacters(String.valueOf(rec.getFlags()));
w.writeEndElement();
if (!rec.getReadUnmappedFlag()) {
w.writeStartElement("qual");
w.writeCharacters(String.valueOf(rec.getMappingQuality()));
w.writeEndElement();
w.writeStartElement("chrom");
w.writeAttribute("index", String.valueOf(rec.getReferenceIndex()));
w.writeCharacters(rec.getReferenceName());
w.writeEndElement();
w.writeStartElement("pos");
w.writeCharacters(String.valueOf(rec.getAlignmentStart()));
w.writeEndElement();
w.writeStartElement("cigar");
w.writeCharacters(rec.getCigarString());
w.writeEndElement();
}
if (!rec.getMateUnmappedFlag()) {
w.writeStartElement("mate-chrom");
w.writeAttribute("index", String.valueOf(rec.getMateReferenceIndex()));
w.writeCharacters(rec.getMateReferenceName());
w.writeEndElement();
w.writeStartElement("mate-pos");
w.writeCharacters(String.valueOf(rec.getMateAlignmentStart()));
w.writeEndElement();
}
if (!rec.getReadUnmappedFlag()) {
if (genomicSequence == null
|| genomicSequence.getChrom().equals(rec.getReferenceName())) {
genomicSequence = new GenomicSequence(indexedFastaSequenceFile, rec.getReferenceName());
}
w.writeStartElement("align");
int readIndex = 0;
int refIndex = rec.getAlignmentStart();
for (final CigarElement e : rec.getCigar().getCigarElements()) {
switch (e.getOperator()) {
case H:
break; // ignore hard clips
case P:
break; // ignore pads
case I: // cont.
case S:
{
final int length = e.getLength();
for (int i = 0; i < length; ++i) {
w.writeEmptyElement(e.getOperator().name());
w.writeAttribute("read-index", String.valueOf(readIndex + 1));
if (readIndex >= 0 && readIndex < readbases.length) {
w.writeAttribute("read-base", String.valueOf((char) (readbases[readIndex])));
}
readIndex++;
}
break;
}
case N: // cont. -- reference skip
case D:
{
final int length = e.getLength();
for (int i = 0; i < length; ++i) {
w.writeEmptyElement(e.getOperator().name());
w.writeAttribute("ref-index", String.valueOf(refIndex));
if (refIndex >= 1 && refIndex <= genomicSequence.length()) {
w.writeAttribute(
"ref-base", String.valueOf(genomicSequence.charAt(refIndex - 1)));
}
refIndex++;
}
break;
}
case M:
case EQ:
case X:
{
final int length = e.getLength();
for (int i = 0; i < length; ++i) {
w.writeEmptyElement(e.getOperator().name());
char baseRead = '\0';
if (readIndex >= 0 && readIndex < readbases.length) {
baseRead = (char) (rec.getReadBases()[readIndex]);
w.writeAttribute("read-index", String.valueOf(readIndex + 1));
w.writeAttribute("read-base", String.valueOf(baseRead));
}
w.writeAttribute("ref-index", String.valueOf(refIndex));
if (refIndex >= 1 && refIndex <= genomicSequence.length()) {
char baseRef = genomicSequence.charAt(refIndex - 1);
w.writeAttribute("ref-base", String.valueOf(baseRef));
if (Character.toUpperCase(baseRef) != Character.toUpperCase(baseRead)) {
w.writeAttribute("mismatch", "true");
}
}
refIndex++;
readIndex++;
}
break;
}
default:
throw new IllegalStateException(
"Case statement didn't deal with cigar op: " + e.getOperator());
}
}
}
w.writeEndElement();
w.writeEndElement();
iter.close();
w.writeEndElement();
}
w.writeEndElement();
w.writeEndDocument();
w.flush();
w.close();
} catch (Exception err) {
error(err);
return -1;
} finally {
CloserUtil.close(samFileReader);
CloserUtil.close(indexedFastaSequenceFile);
}
return 0;
}
private void collectQualityData(final SAMRecord record, final ReferenceSequence reference) {
// If the read isnt an aligned PF read then look at the read for no-calls
if (record.getReadUnmappedFlag()
|| record.getReadFailsVendorQualityCheckFlag()
|| !doRefMetrics) {
final byte[] readBases = record.getReadBases();
for (int i = 0; i < readBases.length; i++) {
if (SequenceUtil.isNoCall(readBases[i])) {
badCycleHistogram.increment(
CoordMath.getCycle(record.getReadNegativeStrandFlag(), readBases.length, i));
}
}
} else if (!record.getReadFailsVendorQualityCheckFlag()) {
final boolean highQualityMapping = isHighQualityMapping(record);
if (highQualityMapping) metrics.PF_HQ_ALIGNED_READS++;
final byte[] readBases = record.getReadBases();
final byte[] refBases = reference.getBases();
final byte[] qualities = record.getBaseQualities();
final int refLength = refBases.length;
long mismatchCount = 0;
long hqMismatchCount = 0;
for (final AlignmentBlock alignmentBlock : record.getAlignmentBlocks()) {
final int readIndex = alignmentBlock.getReadStart() - 1;
final int refIndex = alignmentBlock.getReferenceStart() - 1;
final int length = alignmentBlock.getLength();
for (int i = 0; i < length && refIndex + i < refLength; ++i) {
final int readBaseIndex = readIndex + i;
boolean mismatch =
!SequenceUtil.basesEqual(readBases[readBaseIndex], refBases[refIndex + i]);
boolean bisulfiteBase = false;
if (mismatch && isBisulfiteSequenced) {
if ((record.getReadNegativeStrandFlag()
&& (refBases[refIndex + i] == 'G' || refBases[refIndex + i] == 'g')
&& (readBases[readBaseIndex] == 'A' || readBases[readBaseIndex] == 'a'))
|| ((!record.getReadNegativeStrandFlag())
&& (refBases[refIndex + i] == 'C' || refBases[refIndex + i] == 'c')
&& (readBases[readBaseIndex] == 'T')
|| readBases[readBaseIndex] == 't')) {
bisulfiteBase = true;
mismatch = false;
}
}
if (mismatch) mismatchCount++;
metrics.PF_ALIGNED_BASES++;
if (!bisulfiteBase) nonBisulfiteAlignedBases++;
if (highQualityMapping) {
metrics.PF_HQ_ALIGNED_BASES++;
if (!bisulfiteBase) hqNonBisulfiteAlignedBases++;
if (qualities[readBaseIndex] >= BASE_QUALITY_THRESHOLD)
metrics.PF_HQ_ALIGNED_Q20_BASES++;
if (mismatch) hqMismatchCount++;
}
if (mismatch || SequenceUtil.isNoCall(readBases[readBaseIndex])) {
badCycleHistogram.increment(
CoordMath.getCycle(record.getReadNegativeStrandFlag(), readBases.length, i));
}
}
}
mismatchHistogram.increment(mismatchCount);
hqMismatchHistogram.increment(hqMismatchCount);
// Add any insertions and/or deletions to the global count
for (final CigarElement elem : record.getCigar().getCigarElements()) {
final CigarOperator op = elem.getOperator();
if (op == CigarOperator.INSERTION || op == CigarOperator.DELETION) ++this.indels;
}
}
}
public void find_coverage(SAMResource sres) {
int start_base = sres.region.range.start;
int end_base = sres.region.range.end;
int coverage_len = (end_base - start_base) + 1;
int i, end, ref_i, read_i, len;
int[] coverage = new int[coverage_len];
Arrays.fill(coverage, 0);
WorkingFile wf = null;
if (outfile != null) {
try {
wf = new WorkingFile(outfile);
ps = wf.getPrintStream();
} catch (Exception e) {
System.err.println("I/O error: " + e); // debug
e.printStackTrace();
System.exit(1);
}
}
try {
//
// gather coverage info:
//
CloseableIterator<SAMRecord> iterator = sres.get_iterator();
int read_count = 0;
int ref_min = -1;
int ref_max = -1;
while (iterator.hasNext()) {
SAMRecord sr = iterator.next();
read_count++;
// System.err.println(sr.getReadName() + ": " + sr.getAlignmentStart() + "-" +
// sr.getAlignmentEnd()); // debug
if (sr.getReadUnmappedFlag()) continue;
if (sr.getDuplicateReadFlag()) {
if (verbose_mode)
System.err.println(
sr.getReadName()
+ "."
+ (sr.getReadNegativeStrandFlag() ? "R" : "F")
+ " ignoring, duplicate");
continue;
}
byte[] read = sr.getReadBases();
byte[] quals = sr.getBaseQualities();
for (AlignmentBlock ab : sr.getAlignmentBlocks()) {
len = ab.getLength();
read_i = ab.getReadStart() - 1;
ref_i = ab.getReferenceStart() - start_base;
if (ref_min == -1 || ref_i < ref_min) ref_min = ref_i;
for (i = read_i, end = read_i + len; i < end; i++, ref_i++) {
if (ref_i >= 0 && ref_i < coverage_len) {
if (quals[i] >= MIN_QUALITY) {
if (verbose_mode)
System.err.println(
sr.getReadName()
+ "."
+ (sr.getReadNegativeStrandFlag() ? "R" : "F")
+ " hit at "
+ (ref_i + start_base)
+ " as="
+ sr.getAlignmentStart()
+ " ae="
+ sr.getAlignmentEnd());
coverage[ref_i]++;
} else if (verbose_mode) {
System.err.println(
sr.getReadName()
+ "."
+ (sr.getReadNegativeStrandFlag() ? "R" : "F")
+ " qual_reject at "
+ (ref_i + start_base)
+ " as="
+ sr.getAlignmentStart()
+ " ae="
+ sr.getAlignmentEnd());
}
}
}
if (ref_max == -1 || ref_i > ref_max) ref_max = ref_i;
}
}
sres.close();
System.err.println(
"records:"
+ read_count
+ " ref_min:"
+ (ref_min + start_base)
+ " ref_max:"
+ (ref_max + start_base)); // debug
//
// report coverage info:
//
for (i = 0; i < coverage.length; i++) {
if (name != null) ps.print(name + ",");
ps.println((i + start_base) + "," + coverage[i]); // debug
}
if (wf != null) wf.finish();
} catch (Exception e) {
System.err.println("ERROR: " + e); // debug
e.printStackTrace();
}
}
public static void align(
Graph graph,
SAMRecord rec,
Node recNode,
ReferenceSequence sequence,
SAMProgramRecord programRecord,
int offset,
AlleleCoverageCutoffs alleleCoverageCutoffs,
boolean correctBases,
boolean useSequenceQualities,
int MAXIMUM_TOTAL_COVERAGE,
int MAX_HEAP_SIZE)
throws Exception {
int i;
AlignHeapNode curAlignHeapNode = null;
AlignHeapNode nextAlignHeapNode = null;
AlignHeapNode bestAlignHeapNode = null;
AlignHeap heap = null;
String read = null; // could be cs
String readBases = null; // always nt
String qualities = null; // could be cq
SRMAUtil.Space space = SRMAUtil.Space.NTSPACE;
ListIterator<NodeRecord> iter = null;
AlignHeapNodeComparator comp = null;
int alignmentStart = -1;
int numStartNodesAdded = 0;
boolean strand = rec.getReadNegativeStrandFlag(); // false -> forward, true -> reverse
String softClipStartBases = null;
String softClipStartQualities = null;
String softClipEndBases = null;
String softClipEndQualities = null;
// Debugging stuff
String readName = rec.getReadName();
assert SRMAUtil.Space.COLORSPACE != space;
// Get space
read = (String) rec.getAttribute("CS");
if (null == read) {
// Use base space
space = SRMAUtil.Space.NTSPACE;
} else {
// assumes CS and CQ are always in sequencing order
space = SRMAUtil.Space.COLORSPACE;
}
// Get read and qualities
if (space == SRMAUtil.Space.NTSPACE) {
byte tmpRead[] = rec.getReadString().getBytes();
byte tmpQualities[] = rec.getBaseQualityString().getBytes();
// Reverse once
if (strand) { // reverse
SAMRecordUtil.reverseArray(tmpRead);
SAMRecordUtil.reverseArray(tmpQualities);
}
read = new String(tmpRead);
readBases = new String(tmpRead);
qualities = new String(tmpQualities);
// Reverse again
if (strand) { // reverse
SAMRecordUtil.reverseArray(tmpRead);
SAMRecordUtil.reverseArray(tmpQualities);
}
} else {
byte tmpRead[] = rec.getReadString().getBytes();
// Reverse once
if (strand) { // reverse
SAMRecordUtil.reverseArray(tmpRead);
}
readBases = new String(tmpRead);
// Reverse again
if (strand) { // reverse
SAMRecordUtil.reverseArray(tmpRead);
}
read = SRMAUtil.normalizeColorSpaceRead(read);
qualities = (String) rec.getAttribute("CQ");
// Some aligners include a quality value for the adapter. A quality value
// IMHO should not be given for an unobserved (assumed) peice of data. Trim
// the first quality in this case
if (qualities.length() == 1 + read.length()) { // trim the first quality
qualities = qualities.substring(1);
}
}
// Reverse back
if (readBases.length() <= 0) {
throw new Exception("Error. The current alignment has no bases.");
}
if (read.length() <= 0) {
throw new Exception("Error. The current alignment has no bases.");
}
if (qualities.length() <= 0) {
throw new Exception("Error. The current alignment has no qualities.");
}
if (readBases.length() != read.length()) {
if (space == SRMAUtil.Space.COLORSPACE) {
throw new Exception(
"Error. The current alignment's read bases length does not match the length of the colors in the CS tag ["
+ rec.getReadName()
+ "].");
} else {
throw new Exception("Error. Internal error: readBases.length() != read.length()");
}
}
// Deal with soft-clipping
// - save the soft clipped sequence for latter
{
List<CigarElement> cigarElements = null;
cigarElements = rec.getCigar().getCigarElements();
CigarElement e1 = cigarElements.get(0); // first
CigarElement e2 = cigarElements.get(cigarElements.size() - 1); // last
// Soft-clipped
if (CigarOperator.S == e1.getOperator()) {
if (space == SRMAUtil.Space.COLORSPACE) {
throw new Exception(
"Error. Soft clipping with color-space data not currently supported.");
}
int l = e1.getLength();
if (strand) { // reverse
softClipStartBases = readBases.substring(readBases.length() - l);
softClipStartQualities = qualities.substring(qualities.length() - l);
readBases = readBases.substring(0, readBases.length() - l);
read = read.substring(0, read.length() - l);
qualities = qualities.substring(0, qualities.length() - l);
} else {
softClipStartBases = readBases.substring(0, l - 1);
softClipStartQualities = qualities.substring(0, l - 1);
readBases = readBases.substring(l);
read = read.substring(l);
qualities = qualities.substring(l);
}
}
if (CigarOperator.S == e2.getOperator()) {
if (space == SRMAUtil.Space.COLORSPACE) {
throw new Exception(
"Error. Soft clipping with color-space data not currently supported.");
}
int l = e2.getLength();
if (strand) { // reverse
softClipEndBases = readBases.substring(0, l - 1);
softClipEndQualities = qualities.substring(0, l - 1);
readBases = readBases.substring(l);
read = read.substring(l);
qualities = qualities.substring(l);
} else {
softClipEndBases = readBases.substring(readBases.length() - l);
softClipEndQualities = qualities.substring(qualities.length() - l);
readBases = readBases.substring(0, readBases.length() - l);
read = read.substring(0, read.length() - l);
qualities = qualities.substring(0, qualities.length() - l);
}
}
}
// Remove mate pair information
Align.removeMateInfo(rec);
comp =
new AlignHeapNodeComparator(
(strand) ? AlignHeap.HeapType.MAXHEAP : AlignHeap.HeapType.MINHEAP);
// Bound by original alignment if possible
bestAlignHeapNode =
Align.boundWithOriginalAlignment(
rec,
graph,
recNode,
comp,
strand,
read,
qualities,
readBases,
space,
sequence,
alleleCoverageCutoffs,
useSequenceQualities,
MAXIMUM_TOTAL_COVERAGE,
MAX_HEAP_SIZE);
/*
System.err.println("readName="+rec.getReadName());
if(null != bestAlignHeapNode) {
System.err.println("\nFOUND BEST:" + rec.toString());
}
else {
System.err.println("\nNOT FOUND (BEST): " + rec.toString());
}
Align.updateSAM(rec, programRecord, bestAlignHeapNode, space, read, qualities, softClipStartBases, softClipStartQualities, softClipEndBases, softClipEndQualities, strand, correctBases);
return;
*/
heap = new AlignHeap((strand) ? AlignHeap.HeapType.MAXHEAP : AlignHeap.HeapType.MINHEAP);
// Add start nodes
if (strand) { // reverse
alignmentStart = rec.getAlignmentEnd();
for (i = alignmentStart + offset; alignmentStart - offset <= i; i--) {
int position = graph.getPriorityQueueIndexAtPositionOrBefore(i);
PriorityQueue<Node> startNodeQueue = graph.getPriorityQueue(position);
if (0 != position && null != startNodeQueue) {
Iterator<Node> startNodeQueueIter = startNodeQueue.iterator();
while (startNodeQueueIter.hasNext()) {
Node startNode = startNodeQueueIter.next();
int f = passFilters(graph, startNode, alleleCoverageCutoffs, MAXIMUM_TOTAL_COVERAGE);
if (0 == f) {
heap.add(
new AlignHeapNode(
null,
startNode,
startNode.coverage,
read.charAt(0),
qualities.charAt(0),
useSequenceQualities,
space));
} else if (f < 0) {
return;
}
if (startNode.position < i) {
i = startNode.position;
}
numStartNodesAdded++;
}
}
}
} else {
alignmentStart = rec.getAlignmentStart();
for (i = alignmentStart - offset; i <= alignmentStart + offset; i++) {
int position = graph.getPriorityQueueIndexAtPositionOrGreater(i);
PriorityQueue<Node> startNodeQueue = graph.getPriorityQueue(position);
if (0 != position && null != startNodeQueue) {
Iterator<Node> startNodeQueueIter = startNodeQueue.iterator();
while (startNodeQueueIter.hasNext()) {
Node startNode = startNodeQueueIter.next();
int f = passFilters(graph, startNode, alleleCoverageCutoffs, MAXIMUM_TOTAL_COVERAGE);
if (0 == f) {
heap.add(
new AlignHeapNode(
null,
startNode,
startNode.coverage,
read.charAt(0),
qualities.charAt(0),
useSequenceQualities,
space));
} else if (f < 0) {
return;
}
if (i < startNode.position) {
i = startNode.position;
}
numStartNodesAdded++;
}
}
}
}
if (numStartNodesAdded == 0) {
throw new Exception("Did not add any start nodes!");
}
// Get first node off the heap
curAlignHeapNode = heap.poll();
while (null != curAlignHeapNode) {
if (MAX_HEAP_SIZE <= heap.size()) {
// too many to consider
return;
}
// System.err.println("strand:" + strand + "\tsize:" + heap.size() + "\talignmentStart:" +
// alignmentStart + "\toffset:" + offset + "\treadOffset:" + curAlignHeapNode.readOffset);
// System.err.print("size:" + heap.size() + ":" + curAlignHeapNode.readOffset + ":" +
// curAlignHeapNode.score + ":" + curAlignHeapNode.alleleCoverageSum + ":" +
// curAlignHeapNode.startPosition + "\t");
// curAlignHeapNode.node.print(System.err);
// System.err.print("\rposition:" + curAlignHeapNode.node.position + "\treadOffset:" +
// curAlignHeapNode.readOffset);
// Remove all non-insertions with the same contig/pos/read-offset/type/base and lower score
nextAlignHeapNode = heap.peek();
while (Node.INSERTION != curAlignHeapNode.node.type
&& null != nextAlignHeapNode
&& 0 == comp.compare(curAlignHeapNode, nextAlignHeapNode)) {
if (curAlignHeapNode.score < nextAlignHeapNode.score
|| (curAlignHeapNode.score == nextAlignHeapNode.score
&& curAlignHeapNode.alleleCoverageSum < nextAlignHeapNode.alleleCoverageSum)) {
// Update current node
curAlignHeapNode = heap.poll();
} else {
// Ignore next node
heap.poll();
}
nextAlignHeapNode = heap.peek();
}
nextAlignHeapNode = null;
// Check if the alignment is complete
if (curAlignHeapNode.readOffset == read.length() - 1) {
// All read bases examined, store if has the best alignment.
// System.err.print(curAlignHeapNode.alleleCoverageSum + ":" + curAlignHeapNode.score +
// ":");
// System.err.print(curAlignHeapNode.startPosition + ":");
// curAlignHeapNode.node.print(System.err);
if (null == bestAlignHeapNode
|| bestAlignHeapNode.score < curAlignHeapNode.score
|| (bestAlignHeapNode.score == curAlignHeapNode.score
&& bestAlignHeapNode.alleleCoverageSum < curAlignHeapNode.alleleCoverageSum)) {
bestAlignHeapNode = curAlignHeapNode;
}
} else if (null != bestAlignHeapNode && curAlignHeapNode.score < bestAlignHeapNode.score) {
// ignore, under the assumption that scores can only become more negative.
} else {
if (strand) { // reverse
// Go to all the "prev" nodes
iter = curAlignHeapNode.node.prev.listIterator();
} else { // forward
// Go to all "next" nodes
iter = curAlignHeapNode.node.next.listIterator();
}
while (iter.hasNext()) {
NodeRecord next = iter.next();
int f =
passFilters(
graph, next.node, next.coverage, alleleCoverageCutoffs, MAXIMUM_TOTAL_COVERAGE);
if (0 == f) {
heap.add(
new AlignHeapNode(
curAlignHeapNode,
next.node,
next.coverage,
read.charAt(curAlignHeapNode.readOffset + 1),
qualities.charAt(curAlignHeapNode.readOffset + 1),
useSequenceQualities,
space));
} else if (f < 0) {
return;
}
}
iter = null;
}
// Get next node
curAlignHeapNode = heap.poll();
}
// Recover alignment
Align.updateSAM(
rec,
sequence,
programRecord,
bestAlignHeapNode,
space,
read,
qualities,
softClipStartBases,
softClipStartQualities,
softClipEndBases,
softClipEndQualities,
strand,
correctBases);
}
@Override
public int doWork(String[] args) {
boolean repair_missing_read = false;
SortingCollectionFactory<MappedFastq> sortingFactory =
new SortingCollectionFactory<MappedFastq>();
File forwardFile = null;
File reverseFile = null;
com.github.lindenb.jvarkit.util.cli.GetOpt opt =
new com.github.lindenb.jvarkit.util.cli.GetOpt();
int c;
sortingFactory.setComponentType(MappedFastq.class);
sortingFactory.setCodec(new MappedFastqCodec());
sortingFactory.setComparator(new MappedFastqComparator());
while ((c = opt.getopt(args, super.getGetOptDefault() + "F:R:N:r")) != -1) {
switch (c) {
case 'F':
forwardFile = new File(opt.getOptArg());
break;
case 'R':
reverseFile = new File(opt.getOptArg());
break;
case 't':
addTmpDirectory(new File(opt.getOptArg()));
break;
case 'N':
sortingFactory.setMaxRecordsInRAM(Math.max(Integer.parseInt(opt.getOptArg()), 100));
break;
case 'r':
repair_missing_read = true;
break;
case ':':
System.err.println("Missing argument for option -" + opt.getOptOpt());
return -1;
default:
{
switch (handleOtherOptions(c, opt, args)) {
case EXIT_FAILURE:
return -1;
case EXIT_SUCCESS:
return 0;
default:
break;
}
}
}
}
SAMFileReader sfr = null;
SortingCollection<MappedFastq> fastqCollection = null;
try {
sortingFactory.setTmpDirs(this.getTmpDirectories());
fastqCollection = sortingFactory.make();
fastqCollection.setDestructiveIteration(true);
boolean found_single = false;
boolean found_paired = false;
long non_primary_alignmaned_flag = 0L;
if (opt.getOptInd() == args.length) {
info("Reading from stdin");
sfr = new SAMFileReader(System.in);
} else if (opt.getOptInd() + 1 == args.length) {
String filename = args[opt.getOptInd()];
sfr = new SAMFileReader(new File(filename));
} else {
error(getMessageBundle("illegal.number.of.arguments"));
return -1;
}
sfr.setValidationStringency(ValidationStringency.LENIENT);
SAMRecordIterator iter = sfr.iterator();
SAMSequenceDictionaryProgress progress =
new SAMSequenceDictionaryProgress(sfr.getFileHeader().getSequenceDictionary());
while (iter.hasNext()) {
SAMRecord rec = iter.next();
progress.watch(rec);
if (rec.isSecondaryOrSupplementary()) {
if (non_primary_alignmaned_flag == 0) {
warning("SKIPPING NON-PRIMARY " + (non_primary_alignmaned_flag + 1) + " ALIGNMENTS");
}
non_primary_alignmaned_flag++;
continue;
}
MappedFastq m = new MappedFastq();
m.name = rec.getReadName();
if (m.name == null) m.name = "";
m.hash = m.name.hashCode();
m.seq = rec.getReadString();
if (m.seq.equals(SAMRecord.NULL_SEQUENCE_STRING)) m.seq = "";
m.qual = rec.getBaseQualityString();
if (m.qual.equals(SAMRecord.NULL_QUALS_STRING)) m.qual = "";
if (!rec.getReadUnmappedFlag() && rec.getReadNegativeStrandFlag()) {
m.seq = AcidNucleics.reverseComplement(m.seq);
m.qual = new StringBuilder(m.qual).reverse().toString();
}
if (m.seq.length() != m.qual.length()) {
error("length(seq)!=length(qual) in " + m.name);
continue;
}
if (m.seq.isEmpty() && m.qual.isEmpty()) {
m.seq = "N";
m.qual = "#";
}
if (rec.getReadPairedFlag()) {
found_paired = true;
if (found_single) {
sfr.close();
throw new PicardException("input is a mix of paired/singled reads");
}
m.side = (byte) (rec.getSecondOfPairFlag() ? 2 : 1);
} else {
found_single = true;
if (found_paired) {
sfr.close();
throw new PicardException("input is a mix of paired/singled reads");
}
m.side = (byte) 0;
}
fastqCollection.add(m);
}
iter.close();
CloserUtil.close(iter);
CloserUtil.close(sfr);
progress.finish();
fastqCollection.doneAdding();
info("Done reading.");
if (found_paired) {
FastqWriter fqw1 = null;
FastqWriter fqw2 = null;
if (forwardFile != null) {
info("Writing to " + forwardFile);
fqw1 = new BasicFastqWriter(forwardFile);
} else {
info("Writing to stdout");
fqw1 = new BasicFastqWriter(new PrintStream(System.out));
}
if (reverseFile != null) {
info("Writing to " + reverseFile);
fqw2 = new BasicFastqWriter(reverseFile);
} else {
info("Writing to interlaced stdout");
fqw2 = fqw1;
}
List<MappedFastq> row = new ArrayList<MappedFastq>();
CloseableIterator<MappedFastq> r = fastqCollection.iterator();
for (; ; ) {
MappedFastq curr = null;
if (r.hasNext()) curr = r.next();
if (curr == null || (!row.isEmpty() && !row.get(0).name.equals(curr.name))) {
if (!row.isEmpty()) {
if (row.size() > 2) {
warning("WTF :" + row);
}
boolean found_F = false;
boolean found_R = false;
for (MappedFastq m : row) {
switch ((int) m.side) {
case 1:
if (found_F)
throw new PicardException("two forward reads found for " + row.get(0).name);
found_F = true;
echo(fqw1, m);
break;
case 2:
if (found_R)
throw new PicardException("two reverse reads found for " + row.get(0).name);
found_R = true;
echo(fqw2, m);
break;
default:
throw new IllegalStateException("uh???");
}
}
if (!found_F) {
if (repair_missing_read) {
warning("forward not found for " + row.get(0));
MappedFastq pad = new MappedFastq();
pad.side = (byte) 1;
pad.name = row.get(0).name;
pad.seq = "N";
pad.qual = "#";
echo(fqw1, pad);
} else {
throw new PicardException("forward not found for " + row);
}
}
if (!found_R) {
if (repair_missing_read) {
warning("reverse not found for " + row.get(0));
MappedFastq pad = new MappedFastq();
pad.side = (byte) 2;
pad.name = row.get(0).name;
pad.seq = "N";
pad.qual = "#";
echo(fqw2, pad);
} else {
throw new PicardException("reverse not found for " + row);
}
}
}
if (curr == null) break;
row.clear();
}
row.add(curr);
}
r.close();
fqw1.close();
fqw2.close();
} else if (found_single) {
FastqWriter fqw1 = null;
if (forwardFile != null) {
info("Writing to " + forwardFile);
fqw1 = new BasicFastqWriter(forwardFile);
} else {
info("Writing to stdout");
fqw1 = new BasicFastqWriter(new PrintStream(System.out));
}
CloseableIterator<MappedFastq> r = fastqCollection.iterator();
while (r.hasNext()) {
echo(fqw1, r.next());
}
r.close();
fqw1.close();
}
return 0;
} catch (Exception err) {
error(err);
return -1;
} finally {
if (fastqCollection != null) fastqCollection.cleanup();
}
}