Provide more reproducible details on provenance

[jbingham]

Provide more reproducible details on provenance:

(1) Reads

• exactly which BAMs were imported (e.g., the ftp path)
• how the reads were imported to readstore (e.g. a sample command line referencing world-readable 1,000 genomes BAMs in GCS would be explanatory)
• description of any data cleaning / data transformation that happens during read store's import processing (though this should be written on a non-dataset specific page)

(2) Variant Store import

• how the variants were imported to readstore (e.g. a sample command line referencing world-readable 1,000 genomes VCFs in GCS would be explanatory; I have these in a non-public GCS bucket currently)
• any data cleaning / data transformation that happens during variant store's import processing (though this should be written on a non-dataset specific page) --- this is important, variant store does silently throw away information if the values in the INFO fields for "the same variant" differ per input file

The idea is that there should be enough information here for someone else to follow along and reproduce the exact same end result.

[pgrosu]

Hi Jonathan,

I totally agree! Transparency and clarity of all the steps performed - starting from raw data all way to analysis - will make a lot of the scientific community very happy and grow the community even more.

Thank you :) Paul

[deflaux]

This is well underway with https://cloud.google.com/genomics/public-data from @doll