Hi Esteban,
The distinction that might not be clear is that a target might be
non-zero-coverage, i.e. some of the target has coverage, but that does not mean
that there are not loci in a target region that have zero coverage.
Excuse me if I'm stating the obvious, but if the mean depth is 10 and a
particular locus has depth 1, then the amount of sequencing needed to get that
locus up to depth 10 is estimated at 10 times the amount of sequencing
represented by your BAM. However, if a particular locus has depth 0, i.e. no
coverage, then it is not clear how much more sequencing would be needed to
bring it up to 10.
If the goal is to bring 80% of the bases in targets that have non-zero coverage
up to the target depth, then at least 80% of the bases in the total non-zero
coverage target territory must be covered. If not, you get infinity for
FOLD_80_BASE_PENALTY.
-Alec
On May 27, 2014, at 1:09 PM, Esteban Czwan <ecz...@gmail.com> wrote:
> Hi Alec,
>
> Thank you for your answer.
>
> I am afraid I cannot fully understand your comments. Here's the definition of
> the fold 80 base penalty from Picard:
>
> FOLD_80_BASE_PENALTY: The fold over-coverage necessary to raise 80% of bases
> in "non-zero-cvg" targets to the mean coverage level in those targets.
>
> According to the definition it is not required that more than 80% of the
> bases in the target should have coverage.
>
> What I understand is that the fold 80 base penalty is the fold over-coverage
> needed to raise 80% of bases in covered targets to the mean coverage. As far
> as I see it, it should not matter how many bases are actually covered in the
> target.
>
> What do you think?
>
> Thank you all once again.
>
> Best wishes,
>
> Esteban
>
> El martes, 27 de mayo de 2014, Alec Wysoker <al...@broadinstitute.org>
> escribió:
> Hi Esteban,
>
> Less than 80% of the bases in targets that have non-zero coverage have any
> coverage at all. The program can't know how many times more sequencing would
> be required to raise 80% of bases to mean coverage, so it returns infinity.
>
> -Alec
>
> On May 27, 2014, at 5:54 AM, Esteban Czwan <ecz...@gmail.com> wrote:
>
> > Dear all,
> >
> > I calculated Picard HsMetrics and the FOLD_80_BASE_PENALTY metric had
> > a question mark ("?").
> >
> > I have attached the output file to this email.
> >
> > Do you know what might have happened?
> >
> > Thank you and best wishes,
> >
> > Esteban
> > <EXOMEV3_KAPA_1_microgram_subsample_calculate_hs_metrics.txt>------------------------------------------------------------------------------
> > The best possible search technologies are now affordable for all companies.
> > Download your FREE open source Enterprise Search Engine today!
> > Our experts will assist you in its installation for $59/mo, no commitment.
> > Test it for FREE on our Cloud platform anytime!
> > http://pubads.g.doubleclick.net/gampad/clk?id=145328191&iu=/4140/ostg.clktrk_______________________________________________
> > Samtools-help mailing list
> > Samtools-help@lists.sourceforge.net
> > https://lists.sourceforge.net/lists/listinfo/samtools-help
>
------------------------------------------------------------------------------
The best possible search technologies are now affordable for all companies.
Download your FREE open source Enterprise Search Engine today!
Our experts will assist you in its installation for $59/mo, no commitment.
Test it for FREE on our Cloud platform anytime!
http://pubads.g.doubleclick.net/gampad/clk?id=145328191&iu=/4140/ostg.clktrk
_______________________________________________
Samtools-help mailing list
Samtools-help@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/samtools-help
