What format of sample information to feed the pipleline?

[zhujack]

Hi Sean and all,

Storing and parsing sample information from a JSON format for a workflow provides a flexible and platform independent way to handle complex sample information. Ideally all sample information should be retrieved from a database and a JSON format is automatically generated. Since we don't have a database, yet and it's not easy to generate a JSON format of sample information manually when dealing many samples, a text format sample information table is probably needed at this point. What do you think?

Jack

[seandavi]

I agree. What use cases do you think we need to capture? Right now, three columns (sample, library, and "unit") can be converted to the current JSON format. Is there any other sample information that we need to have available? You had mentioned capture region (bedfile) as another possibility (which would mean some adjustments to the current JSON schema). Any others?

[zhujack]

What I have in mind now is the following additional fields:

run_folder: Illumina run root folder name

reference: what to be compared for somatic variant calling and this should be a sample name that is used as reference for a given sample

genome_partition_file: bed track in GATK format