/**
* Returns the coverage distribution of a single read within the desired region.
*
* <p>Note: This function counts DELETIONS as coverage (since the main purpose is to downsample
* reads for variant regions, and deletions count as variants)
*
* @param read the read to get the coverage distribution of
* @param startLocation the first reference coordinate of the region (inclusive)
* @param stopLocation the last reference coordinate of the region (inclusive)
* @return an array with the coverage of each position from startLocation to stopLocation
*/
public static int[] getCoverageDistributionOfRead(
GATKSAMRecord read, int startLocation, int stopLocation) {
int[] coverage = new int[stopLocation - startLocation + 1];
int refLocation = read.getSoftStart();
for (CigarElement cigarElement : read.getCigar().getCigarElements()) {
switch (cigarElement.getOperator()) {
case S:
case M:
case EQ:
case N:
case X:
case D:
for (int i = 0; i < cigarElement.getLength(); i++) {
if (refLocation >= startLocation && refLocation <= stopLocation) {
int baseCount =
read.isReducedRead()
? read.getReducedCount(refLocation - read.getSoftStart())
: 1;
coverage[refLocation - startLocation] +=
baseCount; // this may be a reduced read, so add the proper number of bases
}
refLocation++;
}
break;
case P:
case I:
case H:
break;
}
if (refLocation > stopLocation) break;
}
return coverage;
}
public static int getMeanRepresentativeReadCount(GATKSAMRecord read) {
if (!read.isReducedRead()) return 1;
// compute mean representative read counts
final byte[] counts = read.getReducedReadCounts();
return (int) Math.round((double) MathUtils.sum(counts) / counts.length);
}