No announcement yet.

Caller oddity or bug?

  • Filter
  • Time
  • Show
Clear All
new posts

  • Caller oddity or bug?

    I've been looking for SNPs that have a few no-calls out of
    about 20 for reliable ones. I also looked at indels and
    found none.

    But for SNPs I did find several under L176.1 and YP326 =
    6777158 G > C .

    What is odd is that the vast majority of no-calls
    in these, which are

    22228615 G>T
    22269858 C>G
    22270062 T>C
    22270127 C>T
    22271724 G>T
    22271726 C>T

    are, when looked at in the bam files, apparently
    perfectly good calls. There are a few cases
    where there are numbers of ancestral/derived
    like 20/80 or 6/20 that are thus iffy. The "read quality"
    shown by Bamview is quite OK.

    A striking example is 22228615G>T . ALL
    of one person’s bam reads (22 of them) are T!
    Yet there is a one-position gap there in the bed file.
    And here the real quality is all top-notch. And
    there's no sign of misalignment. And it fits phylogenetically.

    Does anybody have any idea what's going on?
    I've written Mittelman about this and hopefully his
    lab people will know.

    Doug McDonald

  • #2
    Thomas Krahn made a warning about positions starting 222... here:

    "...Whenever you see a hg19 ChrY position starting with 222 you should be aware of this 125bp repeat region. Essentially you can scratch this region from further phylogenetic consideration.


    I've not worked out the exact extent of the region though.


    • #3
      Well, that answers it. This event is recent enough
      (15th century) that it is probably reliable enough for our
      purposes. With the NGS methods it seems to be OK if you read
      the bam file. We'll just have to be very careful, and not use it without a BigY.