Skip to content

GLIS DOI support

GLIS will provide DOI (https://www.doi.org) registration service for genebanks and other PGRFA (perhaps also GRFA, PGR) holders.

The DOI assigned by GLIS will be in the regexp format 10\.18730/[A-Z0-9]+ (production GLIS) and 10\.0155/[A-Z0-9]+ (test environment). It is possible (and likely) that individual organizations assign their own DOI to accessions (10\.[0-9]+/.+ format).

In this ticket we add support for DOI only and ignore other PUID types uploaded to Genesys.

GLIS DOI extras

Checksum: the last character of the GLIS DOI is a checksum and can be one of *, ~, $, =, U and [A-Z0-9].

Characters I, L, O, U are not in the "DOI" part of the DOI.

http://www.crockford.com/wrmg/base32.html

The "big picture"

Genebank_Genesys_GLIS

Key principles

  1. For genebank material (accessions) there is one DOI assigned to one accession.
  2. From the moment an accession has a DOI assigned, that record can not be deleted from Genesys (REST delete should return 403 Forbidden).
  3. Any other URL (like the current /acn/id/[0-9]+) must permanently redirect to the DOI'zed accession URL in Genesys.

Accession URL

For accessions with a DOI registered in Genesys, we will use the following URL template: https://www.genesys-pgr.org/[DOI] for example https://www.genesys-pgr.org/10.18730/A3C9. Every URL to an accession with DOI must use the DOI'zed URL.

Managing accession passport data

We add a unique DOI field (doi varchar(255) null) to the accession record.

Genebanks obtain DOI for their accessions and register the DOIs with their accession passport data. On upload of passport data to Genesys, the DOI is included with the passport data (next to INSTCODE and ACCENUMB).

The DOI takes absolute priority when identifying accession records. We now use the combination of INSTCODE, ACCENUMB and depending on the institute also GENUS to identify the record in our system. With the DOI provided, we only need to look up the record by the unique DOI value.

  1. When DOI does not exist Genesys, record lookup is based on current approach (INSTCODE, ACCENUMB, GENUS?).
  2. When DOI exists in Genesys we are updating that one single accession record.

There is a possibility in current GLIS that the same DOI would be propagated between PGRFA holders and new holders opt to reuse the DOI assigned by the source genebank. This can only be resolved by resolving the DOI against GLIS and obtaining information on the registered holder. Non-GLIS DOI resolution may not return the same information or even use the same format. This also should be explored in the future.

Holder validation

Only on the first registration of DOI in Genesys we would need to resolve the DOI and retrieve registration data from GLIS. We need to verify that the WIEWS code of the holding institute matches the provided data.

Validation is implemented in glis-client#21 (closed)

Edited by Matija Obreza