search

dnanexus-rnd / GLnexus

Scalable gVCF merging and joint variant calling for population sequencing projects
Apache License 2.0
145 stars 37 forks source link

glnexus_cli should have -o to write out a file instead of stdout #175

Open aardes opened 5 years ago

aardes commented 5 years ago

Hi,

First of all, Great Job.

(Issue) I am trying to use GLnexus, not sure why, but when I run the command under the screen command, the output does not generate. Is there any -out option to define the output file rather than ">"

(Suggestion) Imagin that I run GLnexus on 50 samples, after a week I have 10 new samples, it will be a nice option if GLnexus only go over the new samples and new positions. As I learned currently I need to call all samples again.

Kind regards Amin

mlin commented 5 years ago

Thanks! Yea, I can't think of any excuse for the lack of -o option as an alternative to dumping BCF on standard output. Feeding in such a command-line argument here possibly may just work.

The open-source driver program indeed runs starting over from the gVCF files every time. The DNAnexus-native version supports generating big callsets more incrementally. Some of the machinery for that is actually in the codebase here but not used by the driver program. The state management becomes tricky when big enough to need coordination among multiple compute nodes.