how to proceed to create a persistent identifier for a sample of a specimen used in chemical analysis?

[myrmoteras]

issue:

the Swiss Digital Botanical Gardens Initiative is analysis the chemical diversity in plants, using as starting point plants in botanical gardens. The need a persistent identifier for the sample to obtain from a living plant in the botanical garden, in order to cite it in publications where they describe the chemical substances, evenutally linked to a barcode, physical obbject

question:

which identifier and service to use?

current situation

At the moment a sample is taken, digital images are made and uploaded to iNaturalist, from where an identifier is obtained that is used to cite the specimen. However, iNaturalist does not allow cultivated plants to upload to GBIF, where the sample (occurrence) should be uploaded.

challenges

• iNaturalist provides immediately a PID for the specimen
• iNaturalist does not allow creating it a research grade observation that is moved to GBIF
• create a link between the sample and the living organism in the botanical garden, and throug this possible to a specimen in the original location, to have the entire chain of provenance
• no service available that is customisable enough for the DBGI
• how to best model and which terms to use for a sample collected for the chemical analysis, that eventually can be cited in the publication of the chemical products.

possible solutions

• upload via the local GBIF node (in this case GBIF-CH)
• establish the DBGI as a data publisher including the option to add metadata that is pertinent for the project

[dagendresen]

In my opinion, you need a NEW sample identifier for the new sample taken (for now materialSampleID, hopefully soon materialEntityID will be approved), which can be linked to the specimen ID for the botanical garden specimen using parentMaterialSampleID. The end-point for the botanical garden specimen ID would link to the corresponding occurrenceID for the original sampling from nature (the new sample end-point does not really need to link to this original source occurrenceID?).

iNaturalist provides immediately a PID for the specimen

I disagree, iNaturalist creates identifiers for species occurrences, and NOT for specimens

In my opinion, the new sampling event when the botanical garden living specimen was sampled inside the garden creates a NEW dwc:Occurrence event. And reusing any previous occurrenceID to identify this new occurrence (of the living specimen) would be nonsensical!

The new iNaturalist ID could be such a new occurrenceID. But does iNaturalist provide any appropriate real end-point for this ID? So simply using a new UUID would work more or less almost the same? (If the iNaturalist ID is transferred to GBIF or not is less important? - unless GBIF would [start to] provide a persistent end-point for the occurrenceID).

[ymgan]

other links that Dagmar posted in Zoom meeting:

• https://doi.org/10.1093/database/bax096
• https://www.igsn.org/