Semantic resources project/MouseModels/ArrayExpress Example

Alan suggested that ArrayExpress annotations might be one place where it would be useful to have URIs for mouse strains.

I don't see why they can't use an MGI identifier. But this got me curious. I tried to find a particular example of an ArrayExpress entry that could make use of some kind of mouse strain annotation.

At ArrayExpress, select species 'Mus Musculus', and look for annotations containing the word 'Alzheimer'. There are 7 such experiments. Pick one that provides Pubmed references to that we have a chance of tracking down MGI accessions: E-MTAB-1. Currently annotated with: StrainOrLine	C57BL/6, Tg6074, TgAPP23

The study itself (from 2007) has no associated Pubmed id, but the description refers to PNAS 94, 13287-13292 (pubmed 9371838) as describing Alzheimer model TgAPP23. 'Tg' means transgene, meaning there was never a cultivated strain with this genotype (if that's the right word). Find the Pubmed reference at MGI... this leads to three 'phenotypic alleles' Tg(THY1-APP)1Somm, 2Somm, 3Somm. The third of these has APP 23 listed as a synonym, so let's look at that one:

(what is a 'phenotypic allele' anyhow?)

Tg(Thy1-APP)3Somm 0 strains available, 0 cell lines available. Dead end... either the strain died out or was destroyed, or it lives on in a location beyond the reach of IMSR. Would it still be useful for a data integrator to create a URI for the strain? Would the URI have use anywhere other than annotations of that particular paper and that particular array (and secondary sources that aggregate them)? Is that a good enough reason to register such id in a central strain registry such as MGI or OBOF?