Derived Data Extension

[jdeck88]

In connection with the Ira Moana Project and useful for many other efforts is a way to generically connect Tissues with downstream analytical data. This could be genotype files, OTU tables, or ASV files. We are not proposing to store the raw data but instead provide a means for linking to the derived data files for each tissue. This then, becomes a more convenient method for linking metadata with analysis pipelines, especially with regards to R package development.

Here is a list of the proposed data attributes of the derivedData Class:

• tissueID - the foreign key
• derivedGeneticDataFilename - filename of data file with derived data. could be a compressed archive.
• derivedGeneticDataURI(required) - URI pointer to derived data, suggested to use DOI
• derivedGeneticDataType(required) - (controlled vocabulary: microsats, SNPs, OTUs, ASVs, sequences).
• derivedGeneticDataFormat - (controlled vocabulary: genpop, fasta)
• derivedGeneticDataRemarks - Additional detail and possible fields to describe the dataset.

Note here is that we're not trying to build processing of this data into GEOME-- rather, just collect metadata on associated files and enable linking.

[liblig]

Before you action any of these John, please check back in with me regarding a few vocab changes

[jdeck88]

will do... i'll message you off-list to setup a meeting time.

[jdeck88]

Recap from John and Libby conversation on July 8:

1. One Idea is to create these fields as tissue data properties and call them: derivedGeneticDataFilename, derivedGeneticDataURI, etc... with Pipe delimiters for multiple attributes

2. Another idea is to replicate other metadata on row and just vary the derived data.

The 2nd option above is more true to the GEOME model.

[jdeck88]

See also: https://tools.gbif.org/dwca-validator/extension.do?id=http://rs.gbif.org/terms/1.0/DNADerivedData

https://docs.gbif-uat.org/publishing-dna-derived-data/1.0/en/