use of com.github.lindenb.jvarkit.util.vcf.ContigPos in project jvarkit by lindenb.
the class DepthOfCoverage method doWork.
@Override
public int doWork(final List<String> args) {
PrintWriter out = null;
if (this.auto_mask && this.faidx == null) {
LOG.error("Cannot auto mask if REF is not defined");
return -1;
}
if (this.maskBed != null && this.includeBed != null) {
LOG.error("both --mask and --bed both defined");
return -1;
}
ReferenceSequenceFile referenceSequenceFile = null;
try {
final Predicate<String> isRejectContigRegex;
if (!StringUtils.isBlank(this.skipContigExpr)) {
final Pattern pat = Pattern.compile(this.skipContigExpr);
isRejectContigRegex = S -> pat.matcher(S).matches();
} else {
isRejectContigRegex = S -> false;
}
final SamReaderFactory srf = super.createSamReaderFactory();
if (this.faidx != null) {
srf.referenceSequence(this.faidx);
srf.setUseAsyncIo(this.asyncIo);
referenceSequenceFile = ReferenceSequenceFileFactory.getReferenceSequenceFile(this.faidx);
}
out = super.openPathOrStdoutAsPrintWriter(this.outputFile);
out.print("#BAM\tSample\tContig\tContig-Length\tMasked-Contig-Length\tCount\tDepth\tMedian\tMin\tMax\tMaxPos");
for (RangeOfIntegers.Range r : this.summaryCov.getRanges()) {
if (r.getMinInclusive() == null)
continue;
out.print("\t");
out.print(r.toString());
}
out.println();
for (final Path path : IOUtils.unrollPaths(args)) {
try (final SamReader sr = srf.open(path)) {
if (!sr.hasIndex()) {
LOG.error("File " + path + " is not indexed.");
return -1;
}
final SAMFileHeader header = sr.getFileHeader();
final SAMSequenceDictionary dict = SequenceDictionaryUtils.extractRequired(header);
final Set<String> rejectContigSet = dict.getSequences().stream().map(SSR -> SSR.getSequenceName()).filter(isRejectContigRegex).collect(Collectors.toCollection(HashSet::new));
rejectContigSet.addAll(dict.getSequences().stream().filter(SSR -> SSR.getSequenceLength() < this.skipContigLength).map(SSR -> SSR.getSequenceName()).collect(Collectors.toCollection(HashSet::new)));
if (!header.getSortOrder().equals(SAMFileHeader.SortOrder.coordinate)) {
LOG.error("file is not sorted on coordinate :" + header.getSortOrder() + " " + path);
return -1;
}
final QueryInterval[] intervals;
if (this.useBamIndexFlag && this.includeBed != null) {
if (!sr.hasIndex()) {
LOG.error("Bam is not indexed. " + path);
return -1;
}
final ContigNameConverter contigNameConverter = ContigNameConverter.fromOneDictionary(dict);
final List<QueryInterval> L = new ArrayList<>();
try (BedLineReader br = new BedLineReader(this.includeBed)) {
while (br.hasNext()) {
final BedLine bed = br.next();
final String ctg = contigNameConverter.apply(bed.getContig());
if (StringUtils.isBlank(ctg))
continue;
final int tid = dict.getSequenceIndex(ctg);
if (tid < 0)
continue;
L.add(new QueryInterval(tid, bed.getStart(), bed.getEnd()));
}
}
intervals = QueryInterval.optimizeIntervals(L.toArray(new QueryInterval[L.size()]));
} else {
intervals = null;
}
Integer minCov = null;
Integer maxCov = null;
ContigPos maxCovPosition = null;
long count_raw_bases = 0L;
long count_bases = 0L;
long sum_coverage = 0L;
final DiscreteMedian<Integer> discreteMedian_wg = new DiscreteMedian<>();
final Counter<RangeOfIntegers.Range> countMap_wg = new Counter<>();
final String sample = header.getReadGroups().stream().map(RG -> RG.getSample()).filter(S -> !StringUtils.isBlank(S)).findFirst().orElse(path.toString());
int[] coverage = null;
String prevContig = null;
BitSet mask = null;
final ProgressFactory.Watcher<SAMRecord> progress = ProgressFactory.newInstance().dictionary(dict).logger(LOG).build();
try (CloseableIterator<SAMRecord> iter = intervals == null ? sr.iterator() : sr.queryOverlapping(intervals)) {
for (; ; ) {
final SAMRecord rec = iter.hasNext() ? progress.apply(iter.next()) : null;
if (rec != null) {
if (!SAMRecordDefaultFilter.accept(rec, this.mapping_quality))
continue;
if (rejectContigSet.contains(rec.getContig()))
continue;
}
if (rec == null || !rec.getContig().equals(prevContig)) {
if (coverage != null) {
// DUMP
long count_bases_ctg = 0L;
long sum_coverage_ctg = 0L;
Integer minV_ctg = null;
Integer maxV_ctg = null;
ContigPos maxPos_ctg = null;
final DiscreteMedian<Integer> discreteMedian_ctg = new DiscreteMedian<>();
final Counter<RangeOfIntegers.Range> countMap_ctg = new Counter<>();
for (int i = 0; i < coverage.length; i++) {
if (mask.get(i))
continue;
final int covi = coverage[i];
if (covi > this.max_depth)
continue;
if (minV_ctg == null || minV_ctg.intValue() > covi)
minV_ctg = covi;
if (maxV_ctg == null || maxV_ctg.intValue() < covi) {
maxV_ctg = covi;
maxPos_ctg = new ContigPos(prevContig, i + 1);
}
countMap_ctg.incr(this.summaryCov.getRange(covi));
count_bases_ctg++;
sum_coverage_ctg += covi;
discreteMedian_ctg.add(covi);
}
out.print(path);
out.print("\t");
out.print(sample);
out.print("\t");
out.print(prevContig);
out.print("\t");
out.print(coverage.length);
out.print("\t");
out.print(count_bases_ctg);
out.print("\t");
out.print(sum_coverage_ctg);
out.print("\t");
if (count_bases_ctg > 0) {
out.printf("%.2f", sum_coverage_ctg / (double) count_bases_ctg);
} else {
out.print("N/A");
}
out.print("\t");
final OptionalDouble median = discreteMedian_ctg.getMedian();
if (median.isPresent()) {
out.print(median.getAsDouble());
} else {
out.print("N/A");
}
out.print("\t");
if (minV_ctg != null) {
out.print(minV_ctg);
} else {
out.print("N/A");
}
out.print("\t");
if (maxV_ctg != null) {
out.print(maxV_ctg);
out.print("\t");
out.print(maxPos_ctg);
} else {
out.print("N/A\tN/A");
}
for (final RangeOfIntegers.Range r : this.summaryCov.getRanges()) {
if (r.getMinInclusive() == null)
continue;
out.print("\t");
out.print(countMap_ctg.count(r));
if (!countMap_ctg.isEmpty()) {
out.print(" ");
out.printf("(%.2f%%)", (countMap_ctg.count(r) / (countMap_ctg.getTotal() * 1.0)) * 100.0);
}
}
out.println();
if (minCov == null || (minV_ctg != null && minV_ctg.compareTo(minCov) < 0))
minCov = minV_ctg;
if (maxCov == null || (maxV_ctg != null && maxV_ctg.compareTo(maxCov) > 0)) {
maxCov = maxV_ctg;
maxCovPosition = maxPos_ctg;
}
count_bases += count_bases_ctg;
sum_coverage += sum_coverage_ctg;
count_raw_bases += coverage.length;
discreteMedian_wg.add(discreteMedian_ctg);
countMap_wg.putAll(countMap_ctg);
}
coverage = null;
mask = null;
// /
System.gc();
if (rec == null)
break;
final SAMSequenceRecord ssr = Objects.requireNonNull(dict.getSequence(rec.getContig()));
coverage = new int[ssr.getSequenceLength()];
mask = new BitSet(ssr.getSequenceLength());
if (this.auto_mask && referenceSequenceFile != null) {
final byte[] refSeq = Objects.requireNonNull(referenceSequenceFile.getSequence(ssr.getSequenceName())).getBases();
for (int i = 0; i < refSeq.length; i++) {
if (AcidNucleics.isATGC(refSeq[i]))
continue;
mask.set(i);
}
}
/* read mask */
if (this.maskBed != null) {
final ContigNameConverter contigNameConverter = ContigNameConverter.fromOneDictionary(dict);
try (BedLineReader br = new BedLineReader(this.maskBed)) {
while (br.hasNext()) {
final BedLine bed = br.next();
if (bed == null)
continue;
String ctg = contigNameConverter.apply(bed.getContig());
if (StringUtils.isBlank(ctg))
continue;
if (!rec.getContig().equals(ctg))
continue;
for (int p1 = bed.getStart(); p1 <= bed.getEnd() && p1 <= coverage.length; ++p1) {
mask.set(p1 - 1);
}
}
}
} else if (this.includeBed != null) {
final List<Locatable> list = new ArrayList<>();
final ContigNameConverter contigNameConverter = ContigNameConverter.fromOneDictionary(dict);
try (BedLineReader br = new BedLineReader(this.includeBed)) {
while (br.hasNext()) {
final BedLine bed = br.next();
if (bed == null)
continue;
final String ctg = contigNameConverter.apply(bed.getContig());
if (StringUtils.isBlank(ctg))
continue;
if (!rec.getContig().equals(ctg))
continue;
list.add(new SimpleInterval(ctg, bed.getStart(), bed.getEnd()));
}
}
// sort on starts
Collections.sort(list, (A, B) -> Integer.compare(A.getStart(), B.getStart()));
int p1 = 1;
while (p1 <= coverage.length) {
while (!list.isEmpty() && list.get(0).getEnd() < p1) {
list.remove(0);
}
if (!list.isEmpty() && list.get(0).getStart() <= p1 && p1 <= list.get(0).getEnd()) {
++p1;
continue;
}
mask.set(p1 - 1);
p1++;
}
}
prevContig = rec.getContig();
}
int max_end1 = coverage.length;
if (!this.disable_paired_overlap_flag && rec.getReadPairedFlag() && !rec.getMateUnmappedFlag() && rec.getReferenceIndex().equals(rec.getMateReferenceIndex()) && rec.getAlignmentStart() < rec.getMateAlignmentStart() && rec.getAlignmentEnd() > rec.getMateAlignmentStart()) {
max_end1 = rec.getMateAlignmentStart() - 1;
}
for (final AlignmentBlock block : rec.getAlignmentBlocks()) {
final int pos1 = block.getReferenceStart();
final int len = block.getLength();
for (int i = 0; i < len; i++) {
if (pos1 + i - 1 >= 0 && pos1 + i <= max_end1) {
coverage[pos1 + i - 1]++;
}
}
}
}
/* end rec */
}
/* end iter */
progress.close();
out.print(path);
out.print("\t");
out.print(sample);
out.print("\t");
out.print(SAMRecord.NO_ALIGNMENT_REFERENCE_NAME);
out.print("\t");
out.print(count_raw_bases);
out.print("\t");
out.print(count_bases);
out.print("\t");
out.print(sum_coverage);
out.print("\t");
if (count_bases > 0) {
out.printf("%.2f", sum_coverage / (double) count_bases);
} else {
out.print("N/A");
}
out.print("\t");
final OptionalDouble median = discreteMedian_wg.getMedian();
if (median.isPresent()) {
out.print(median.getAsDouble());
} else {
out.print("N/A");
}
out.print("\t");
if (minCov != null) {
out.print(minCov);
} else {
out.print("N/A");
}
out.print("\t");
if (maxCov != null) {
out.print(maxCov + "\t" + maxCovPosition);
} else {
out.print("N/A\tN/A");
}
for (final RangeOfIntegers.Range r : this.summaryCov.getRanges()) {
if (r.getMinInclusive() == null)
continue;
out.print("\t");
out.print(countMap_wg.count(r));
if (!countMap_wg.isEmpty()) {
out.print(" ");
out.printf("(%.2f%%)", (countMap_wg.count(r) / (countMap_wg.getTotal() * 1.0)) * 100.0);
}
}
out.println();
}
}
out.flush();
out.close();
return 0;
} catch (final Exception err) {
LOG.error(err);
return -1;
} finally {
CloserUtil.close(referenceSequenceFile);
}
}
Aggregations