Talk:GRIMS glossary
From ICISWiki
GRIMS Concepts and Terminology - Discussion
This page is used to discuss the GRIMS concepts and terminology and use the result of the discussion to improve the definitions and explanations in the corresponding "article" page GRIMS Concepts and Terminology.
To facilitate the dicussion, please "sign" each comment/question/suggestion. Signing is done by moving the cursor to a new line immediately below your comment, and then click on the second-last button on top of the editor window. This will produce a --~~~~ string, which displays on the saved page as your user name followed by a timestamp.
The visual layout of a discussion can be improved by using a single colon : as the first character of a paragraph, which results in a one-level indend (= first level response). A double-colon :: results in a two-level indent, (= second-level response) etc.
Using section editing (edit link a the right side of the screen) reduces concurrency problems (i.e. more than one person editing at the same time).
Once a discussion has come to a conclusing, the text on the article page should be modified. The discussion section could then also have a last comment, effectively closing the discussion.
Germplasm sample and Germplasm neighbourhood: do these concepts relate to biological reality or to its representation in ICIS?
--Rhamilton 22:47, 18 May 2006 (PHT) The existing definitions are not sufficiently clear to answer this question. The conflict identified in the next discussion point results from this lack of clarity, showing the importance of answering the question.
Option 1: biological reality
In this case a sample is a biological entity and so must be defined biologically; and a neighbourhood is a set of these biological samples. For example, IR36 is a neighbourhood spanning many samples, regardless of how many GIDs are used to represent those samples.
Option 2: representation in ICIS of biological reality
In this case a sample is a database entity, namely a GID (so a GID is a sample), and a neighbourhood is a set of related GIDs. Therefore the user's choice of how to document germplasm in ICIS determines whether the germplasm is considered one sample or a neighbourhood. For example,
- INGER's IR36 is one Germplasm sample, even though it spans many biological samples from around 30 generations, because INGER chose to represent them all using one GID
- GRC's IR36 is a Germplasm neighbourhood, because GRC chose to represent each generation as a separate GID.
Conflicting definitions for GID and Germplasm sample
--Rhamilton 21:52, 18 May 2006 (PHT)
One GID identifies one Germplasm sample. One Germplasm is “a packet of seeds … in a collection” … that specialists would not want to mix. Definitions of several dependent entities (e.g. GDATE) also require this narrow definition of sample.
Yet GIDs are used in a much broader sense – for germplasm outside collections and for collections of germplasm. For example, IR36 in INGER has gone through 30 generations of management and progressive change, but is documented in IRIS under a single GID. So there is a conflict between the definitions of GID and Germplasm sample, and we must change one or the other. Which one?
Is INGER's IR36 one Germplasm sample? Is a released variety one Germplasm sample no matter how often it is grown and selected by farmers? If the answer is yes, then we must generalize the definition of Germplasm sample, to clarify that it includes germplasm outside collections as well as inside collections, and it includes genetically distinct but related sets of germplasm as well as single samples.
Alternatively, is INGER's IR36 a Germplasm neighbourhood, even though it is represented by only on GID? If yes, then we must generalize the definition of GID to clarify that it can include germplasm neighbourhoods as well as germplasm samples, and germplasm outside collections as well as germplasm inside collections
My preference is the second - generalize the definition of GID. I think it is illogical to classify INGER's set of seed packets of IR36 as one sample because INGER chose to assign only one GID, and yet to classify GRC's accession as a Germplasm Neighbourhood because GRC chooses to manage the generations separately. Especially illogical because GRC's accession will encompass less genetic variation than the 30 years' worth of INGER IR36 generations.
I have provisionally edited the GID definition to conform with the second definition above.
