Bug Bulletin: we have identified a bug that affects indexing when producing gzipped VCFs. This will be fixed in the upcoming 3.2 release; in the meantime you need to reindex gzipped VCFs using Tabix.

BaseCoverageDistribution

blueskypyblueskypy Posts: 182Member
edited June 2013 in Ask the team

In the output grp file,

#:GATKReport.v1.1:1
#:GATKTable:3:880:%s:%s:%s:;
#:GATKTable:BaseCoverageDistribution:A simplified GATK table report
Coverage  Count    Filtered
       0  2859049   2932784
       1   856997    837791
       2   288587    276253
       3    95618     91703

what's the meaning of the three columns?

Thanks,

Post edited by Geraldine_VdAuwera on

Answers

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 5,235Administrator, GSA Member admin

    I'll have the tool author (@Carneiro) confirm, but as I recall:

    1. The first column, Coverage, is the depth of coverage corresponding to the bin (ie the first line is the set of loci that are covered by 0 reads, the second is the set covered by 1 read, etc);

    2. The second, Count, is the number of loci in the bin (without any filtering);

    3. The third, Filtered, is the number of loci in the bin after applying quality filtering to exclude bad reads.

    Geraldine Van der Auwera, PhD

  • blueskypyblueskypy Posts: 182Member
    edited June 2013

    hi, Geraldine, Thanks for the quick response! Two questions:

    1. if the coverage is 2, can I interpret it as the so-called 2x coverage?

    2. By your explanation, the Count should > Filtered; why Count < Filtered at coverage 0?

    Post edited by Geraldine_VdAuwera on
  • blueskypyblueskypy Posts: 182Member

    Well, I found Count < Filtered at some other coverages as well. For example: 23 7910 8119

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 5,235Administrator, GSA Member admin
    1. The 2x-style expression is a convention to express the overall coverage of a dataset, so I'm not sure it's appropriate to use it in this context. If you do use it, make sure to communicate clearly what you mean, to avoid any unfortunate misunderstandings.

    2. Hmm, I may have misremembered. Based on the tech doc it looks like it might be the count of filtered reads (not including good reads). I'll ask @Carneiro to confirm, but I reckon that makes sense. But if so, there's an awful lot of low-quality reads in your data, at least based on the low-value bins you posted...

    Geraldine Van der Auwera, PhD

  • blueskypyblueskypy Posts: 182Member

    Thanks so much, Geraldine! Please confirm!

  • avidLearneravidLearner Posts: 9Member

    Hi Geraldine. Did you get a chance to confirm @blueskypy 's second question - whether the filtered column includes the good reads? I just ran this tool and for most rows, the values for the 3rd column are higher.

  • Geraldine_VdAuweraGeraldine_VdAuwera Posts: 5,235Administrator, GSA Member admin

    Hi @avidLearner,

    It seems I forgot to report back, so here it is.

    Here "filtered" doesn't refer to the count of filtered reads. This output gives two separate coverage distributions (to be plotted as a histogram), one for data that passes internal filters for minimum_mapping_quality and minimum_base_quality, and another one for the data that gets filtered out on those criteria. The first column is the bin, which represents the amount of coverage, and the second and third columns are the numbers of positions in the analysis where coverage was equal to that amount, one for the "passing filters" distribution, and one for the "failing filters" distribution. There is no column that gives total count. I'll see if we can change the column names to avoid confusion.

    Geraldine Van der Auwera, PhD

Sign In or Register to comment.