[DOCKTESTERS] BWA-Mem final

[Miguel Vazquez]

Thank you for this extra effort Jonas, I think this closes this discussion
very nicely.

Best regards

Miguel

On Thu, May 4, 2017 at 11:26 AM, Jonas Demeulemeester <
Jonas.Demeulemeester at crick.ac.uk> wrote:

> Hi all,
>
> Apologies for the delay in reporting back on this, I was out and we’re
> currently in the process of upgrading our compute infrastructure to
> OpenStack 10.
> Before documenting our findings, I ran one last test on the BWA-Mem docker
> to confirm that it is indeed the read order in the BAMs that was causing
> the 3–4% mismatch rates, and not some artefact/error of our BAM resetting
> procedures.
>
> Briefly, I simply shuffled the read order in the original unmapped
> lane-level BAMs for DO51057 (keeping pairs together) and remapped them
> using the BWA-Mem docker (BAMs fed in the same order).
> Results for the normal are:
>
> Lines: 1124984372
> Matches: 1049339853
> Misses: 41820223
> Soft: 33824296
>
>
> and for the tumour:
>
> Lines: 1010510325
> Matches: 933112344 <933%2011%2023%2044>
> Misses: 38856341
> Soft: 38541640
>
> Which show 3.7% and 3.8% mismatch rates, respectively – approx. two orders
> of magnitude higher than when using the same BAM files with the reads in
> the original order.
> These rates are similar to what we saw in the initial tests when running
> on our own unaligned BAM files generated from the final mapped ones.
> Taken together with our previous results, this confirms read order within
> BAMs as the main source of the original discrepancies.
>
> I’ll now proceed to document the results on the BWA-Mem workflow docker
> github page and in the manuscript, and will continue with the other docker
> containers as well.
>
> Cheers,
> Jonas
>
>
>
> _________________________________
> Jonas Demeulemeester, PhD
> Postdoctoral Researcher
> The Francis Crick Institute
> 1 Midland Road
> London
> NW1 1AT
>
> *T:* +44 (0)20 3796 2594 <+44%2020%203796%202594>
> M: +44 (0)7482 070730 <+44%207482%20070730>
> *E:* jonas.demeulemeester at crick.ac.uk
> *W:* www.crick.ac.uk
>
> The Francis Crick Institute Limited is a registered charity in England and
> Wales no. 1140062 and a company registered in England and Wales no.
> 06885462, with its registered office at 1 Midland Road London NW1 1AT
>
> _______________________________________________
> docktesters mailing list
> docktesters at lists.icgc.org
> https://lists.icgc.org/mailman/listinfo/docktesters
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.icgc.org/mailman/private/docktesters/attachments/20170504/f771a7f0/attachment-0001.html>