Preserving "chr" prefix in output records (bug fix)

[John Marshall]

Hello,

When running CAVA over VCF files that have chromosome names like "chr5", we've noticed that CAVA removes the "chr" prefixes in the output files. The VCF headers are copied as is from the input (so retain their <ID=chr5,... prefixes), but the records are output as

5 14370 id ...etc...

and thus are no longer consistent with the headers. I've written a small patch to fix this, which may be of interest:

https://github.com/RahmanTeam/CAVA/pull/3 <https://github.com/RahmanTeam/CAVA/pull/3>

Thanks for considering this patch,

John

[Shawn Yost]

Hi John,

  Thank you for bringing this to our attention.  We appreciate you writing a patch to fix this and we will fix this bug in the latest version of CAVA and have it updated on GitHub in the next couple of days. Hope everything else with CAVA is working great for you.

Thanks,

  Shawn

[John Marshall]

On 6 Jun 2017, at 17:06, Shawn Yost <yost...@gmail.com> wrote:
> Thank you for bringing this to our attention. We appreciate you writing a patch to fix this and we will fix this bug in the latest version of CAVA and have it updated on GitHub in the next couple of days. Hope everything else with CAVA is working great for you.

Thanks for the update. There is one other tiny CAVA problem I've noticed: the output VCF files have an extra tab character at the end of the record lines, which might cause trouble for some tools and is actually forbidden in VCF 4.3. That pull request has more description of the problem and a suggested fix:

https://github.com/RahmanTeam/CAVA/pull/3/commits/67b574c64412cbf37a073e583f815416365a640d

Cheers,

John

[Márton Münz]

Hi John,

Thank you for spotting these. We have now fixed both issues (preserving chr prefix, if any, and ensuring that there are no extra tab characters in output VCF records). The update is released on GitHub as v1.2.2, please see here:

https://github.com/RahmanTeam/CAVA/releases/tag/v1.2.2

Best wishes,

Márton