Understanding Picards HsMetrics

Question

[finswimmer]

Hello,

I have some problems understanding the output of picards CollectHsMetrics. Because I don't have a bait file i uses my bed file with the target regions to produce the interval list.

java -jar picard.jar BedToIntervalList I=target.bed SD=genome.dict O=target.interval_list

This target interval list I now use for the BAIT parameter and the TARGET parameter:

java -jar picard.jar CollectHsMetrics I=Sample.bam O=hsmetrix.txt BI=target.interval_list TI=target.interval_list

In my understanding the results for ON_BAIT_BASES and ON_TARGET_BASES have to be the same as the interval list is the same (and also all other values where is differ between bait and target). But in my case I have this result:

ON_BAIT_BASES: 1012933970

ON_TARGET_BASES: 678789480

Please help me understanding where this difference come from.

Thanks a lot.

fin swimmer

Answer 1

[Pierre Lindenbaum]

In my understanding the 'bait' is the read and the 'target' is the reference . If you have insertions, clipped bases in your reads, there will be a higher number of bases for the reads.

looking at the code in picard:

                int onBaitBases = 0;

                if (!probes.isEmpty()) {
                    for (final Interval bait : probes) {
                        for (final AlignmentBlock block : record.getAlignmentBlocks()) {
                            final int end = CoordMath.getEnd(block.getReferenceStart(), block.getLength());

                            for (int pos = block.getReferenceStart(); pos <= end; ++pos) {
                                if (pos >= bait.getStart() && pos <= bait.getEnd()) ++onBaitBases;
                            }
                        }
                    }

[finswimmer]

Hello Pierre, that makes it a little bit clearer.

So ON_TARGET means here only those bases of read which are realy mapped to a target position, and ON_BAIT counts also those bases that are not strictly mapped to a reference position but "somewhere between" (insertions) or are clipped?

Is the ratio of my 2 values a common one?

fin swimmer