the current submit-a-dataset form has three analysis subtypes: these would be split into three separate bundles, so we can configure the fields for each separately.
The question is: What metadata do we collect, how do we store it in chado, and how do we make a good UX?
Currently for the genome, questions are asked about the tissue type, geographic location, sex. This info all "technically" should go in chado.biomaterial, not chado.analysis. However, fields are typically designed to describe the record youre creating: so tripal_HQ would make it really easy to put this info in analysisprop, harder in biomaterialprop.
consider the current data source information for genome assembly:
We have some possible solutions:
Require users to create a biomaterial first
A bad solution no matter how you slice it. More forms.
Require accessions
Analysis submission is greatly simplified. Users fill out Chado fields we wouldnt expect in the ncbi accession, and provide the accession. Then we create something that pulls the accession and creates an analysis and biomaterial record from the provided accesssions.
Attach properties to analysis instead
Easiest, most straight forward solution, but, technically not "correct" for Chado. Possible to migrate down the road.
Provide a new biomaterial field that creates a biomaterial on the fly
This is do-able, although not typically how Tripal thinks about it. It wants someone to create a biomaterial first, then create an analysis and link it somehow. We would create a field that, while attached to analysis, asks all the questions needed to create the biomaterial.
the current submit-a-dataset form has three analysis subtypes: these would be split into three separate bundles, so we can configure the fields for each separately.
The question is: What metadata do we collect, how do we store it in chado, and how do we make a good UX?
Currently for the genome, questions are asked about the tissue type, geographic location, sex. This info all "technically" should go in chado.biomaterial, not chado.analysis. However, fields are typically designed to describe the record youre creating: so tripal_HQ would make it really easy to put this info in analysisprop, harder in biomaterialprop.
consider the current data source information for genome assembly:
We have some possible solutions:
Require users to create a biomaterial first
A bad solution no matter how you slice it. More forms.
Require accessions
Analysis submission is greatly simplified. Users fill out Chado fields we wouldnt expect in the ncbi accession, and provide the accession. Then we create something that pulls the accession and creates an analysis and biomaterial record from the provided accesssions.
Attach properties to analysis instead
Easiest, most straight forward solution, but, technically not "correct" for Chado. Possible to migrate down the road.
Provide a new biomaterial field that creates a biomaterial on the fly
This is do-able, although not typically how Tripal thinks about it. It wants someone to create a biomaterial first, then create an analysis and link it somehow. We would create a field that, while attached to analysis, asks all the questions needed to create the biomaterial.
Could be a good idea, could be too clunky.