"variant merging" documentation

[googlegenomics]

"variant merging" documentation

[nmousavi]

re error: This is due to an incompatible header data. Looks like 'GQ' field (Genotype Quality) is defined as a single Integer value in one VCF file, while another VCF file defines it as any-number (i.e., '.') of Integers. The pipeline fails because it doesn't know how to build the Bigquery schema in this case.

Solnution:
Make the definition of 'GQ' field consistent across the files. Or

[arostamianfar]

To clarify: using --representative_header_file will not actually work in this case as explained in the thread. This is b/c the parser still returns a single number for the GQ field while the 'correct' representation is to have a list. We realize it is frustrating to use Variant Transforms when dealing with malformed/incompatible VCF files, so we are actively working on automatically fixing such cases in Issue #101. In the meantime, changing the VCF header in all files to have Number='.' would work. You would also need to ensure that other fields (e.g. AD) are properly defined in all files.

RE documentation: Sorry for the delay on this! It has been on my TODO list and I have started on a draft. Should be ready by EOW. Also, tracked on Issue #76.