/**
* Calculates the reference coordinate for the end of the read taking into account soft clips but
* not hard clips.
*
* <p>Note: getUnclippedEnd() adds soft and hard clips, this function only adds soft clips.
*
* @param read the read
* @param cigar the read's cigar
* <p>Note: this overload of the function takes the cigar as input for speed because getCigar
* is an expensive operation. Most callers should use the overload that does not take the
* cigar.
* @return the unclipped end of the read taking soft clips (but not hard clips) into account
*/
public static int getSoftEnd(final GATKRead read, final Cigar cigar) {
Utils.nonNull(read, "read");
Utils.nonNull(cigar, "cigar");
boolean foundAlignedBase = false;
int softEnd = read.getEnd();
final List<CigarElement> cigs = cigar.getCigarElements();
for (int i = cigs.size() - 1; i >= 0; --i) {
final CigarElement cig = cigs.get(i);
final CigarOperator op = cig.getOperator();
if (op
== CigarOperator
.SOFT_CLIP) { // assumes the soft clip that we found is at the end of the aligned read
softEnd += cig.getLength();
} else if (op != CigarOperator.HARD_CLIP) {
foundAlignedBase = true;
break;
}
}
if (!foundAlignedBase) { // for example 64H14S, the soft end is actually the same as the
// alignment end
softEnd = read.getEnd();
}
return softEnd;
}
/**
* Can the adaptor sequence of read be reliably removed from the read based on the alignment of
* read and its mate?
*
* @param read the read to check
* @return true if it can, false otherwise
*/
public static boolean hasWellDefinedFragmentSize(final GATKRead read) {
if (read.getFragmentLength() == 0)
// no adaptors in reads with mates in another chromosome or unmapped pairs
{
return false;
}
if (!read.isPaired())
// only reads that are paired can be adaptor trimmed
{
return false;
}
if (read.isUnmapped() || read.mateIsUnmapped())
// only reads when both reads are mapped can be trimmed
{
return false;
}
// if ( ! read.isProperlyPaired() )
// // note this flag isn't always set properly in BAMs, can will stop us from
// eliminating some proper pairs
// // reads that aren't part of a proper pair (i.e., have strange alignments) can't
// be trimmed
// return false;
if (read.isReverseStrand() == read.mateIsReverseStrand())
// sanity check on isProperlyPaired to ensure that read1 and read2 aren't on the same strand
{
return false;
}
if (read.isReverseStrand()) {
// we're on the negative strand, so our read runs right to left
return read.getEnd() > read.getMateStart();
} else {
// we're on the positive strand, so our mate should be to our right (his start + insert size
// should be past our start)
return read.getStart() <= read.getMateStart() + read.getFragmentLength();
}
}
/**
* Is a base inside a read?
*
* @param read the read to evaluate
* @param referenceCoordinate the reference coordinate of the base to test
* @return true if it is inside the read, false otherwise.
*/
public static boolean isInsideRead(final GATKRead read, final int referenceCoordinate) {
return referenceCoordinate >= read.getStart() && referenceCoordinate <= read.getEnd();
}