Talk:GRIMS data migration

From ICISWiki

Contents 1 GDATE, GLOCN 1.1 General definitions 1.2 Application to accessions in genebanks 1.3 Implications for ICIS applications 2 NDATE, NLOCN 2.1 Early discussion 2.2 Agreed solution 2.3 Implications for ICIS applications

GDATE, GLOCN

General definitions

The definitions given in ICIS do not cover all cases and must be generalised. GDATE and GLOCN = The date and location where the germplasm was created as a distinct unit of management with a new GID.

• For germplasm that acquires a GID because it is a newly-harvested seed sample (for example, as part of a breeding programme, or genebank multiplication programme, or harvested direct from a farmer’s field), GDATE and GLOCN are the date of harvest and the location of the plot or field where the seeds were harvested.
• For a sample of germplasm that is newly collected from a farm store, market place etc, GDATE and GLOCN are the date of collection and the location from which the sample was collected.
• For germplasm that acquires a GID because it is transferred from one organizational unit to another (for example, germplasm transfer between genebanks, or from farm or market place to genebank), GDATE is the date of acquisition by the receiving OU and GLOCN is the location of the source OU.

Application to accessions in genebanks

• For generation 0 with METHOD=IMPORT, the accession was obtained from a donor: the location is the location of the donor organization
• For generation 0 with METHOD=COLLECTED, the accession was derived directly from a sample newly collected from a farm, market place etc: the location is the location where the germplasm was collected.
• For generation 1..n, the location is the holding genebank that undertook the seed multiplication

N.B. generation 0 GIDs derived indirectly by maintenance methods from a sample collected from a farm, market place etc, should also have an attribute CLOCN pointing to the collection location. GPID1 should point to the GID of the first documented sample (which may be the original collected sample), and CLOCN of the accession = GLOCN of the first documented sample = collecting location.

Implications for ICIS applications

Change the terminology of ICIS applications. Change "germplasm date" to "germplasm creation date" and "germplasm location" to "germplasm creation location". Simililarly change "germplasm method" to "germplasm creation method"

NDATE, NLOCN

Early discussion

NDATE=”Date of name assignation”; NLOCN=” Identifier for the location where the name was assigned”. However, these definitions are ambiguous.

• Option 1: apply them to the original assignation of the name. For example, suppose IRGC has an accession obtained from USDA with a PI number. We put the PI number into IRIS as a DACCN of the IRGC accession. We say this name was assigned by USDA. Therefore this DACCN should have NLOCN=USDA and NDATE=date USDA assigned the PI number (available from GRIN web site).
• Option 2: apply them to the association of this name with this particular GID. For the same accession with a PI number as DACCN, we say it was GRC who stated that this PI number is associated with this particular GID. Therefore this DACCN should have NLOCN=IRRI, NDATE=acquisition date.
• Option 3: dissociate date from location, and say NLOCN=USDA, NDATE=acquisition date.

To my mind, option 3 is confusing and illogical, although it may be closest to past usage. According to Graham, option 2 is against the ICIS concept – NLOCN should not be IRRI for a PI number. Therefore I would favour option 1.

Agreed solution

Use option 1: NDATE and NLOCN = date and place where the name was first assigned. If the name identifies a MaintenanceNeighbourhood (e.g. accession or modern variety or inbred line), then NDATE and NLOCN is the date and place the name was assigned to the SourceGermplasm of the MaintenanceNeighbourhood (see GCP domain models). So

• If a GID has a name inherited from its source, that name should also inherit the NDATE and NLOCN of the source name. NDATE will then pre-date GDATE of the GID
• If a name is newly created and assigned to a GID, then
  • NDATE must not pre-date the GDATE of the GID. It may equal GDATE, or it may post-date it. For example generation 0 IRGC accessions are assigned an ACCNO only after successfully completing tests and initial seed multiplication; so the NDATE of the ACCNO is generally months or years later than the GDATE (the date of acquisition)
  • for the ACCNO of generation 0 GIDs of an accession, NLOCN normally <> GLOCN
  • for the MNNAM of generation 1..n GIDs of an accession, NLOCN normally = GLOCN

Implications for ICIS applications

Change the terminology of ICIS applications. Change "name date" to "date named" and "name location" to "naming location"