Re: [GUSDEV] [ApiDB] issue: ExternalNaSequence, VirtualSequence and NaSequence

[<pi...@pc...>]

I agree with this. In fact, the current database being used by our mammalian
DoTS build has incorporated a source_id into the assembly view of 
NASequence. I
believe (could be wrong on this) that this is the only view without a
source_id.   So this would clear the way to having source_id in the 
super class
for the "official" GUS schema. Are all the nasequenceimp views using the same
field for source_id? Are there other views missing a source_id?




Quoting Steve Fischer <sfi...@pc...>:

> folks-
>
> we've been again reminded of the issue that toxo stores its chromosome
> sequence in VirtualSequence while the other two use ExternalNaSequence.
>
> this means that we cannot re-use sql code, which is a headache.
>
> our proposed solution is to upgrade NaSequence view (the superclass) to
> have a source_id column.   We feel that this is correct semantically,
> and also solves our problem.
>
> the semantics are that:
>  (1) NaSequence already has an external_database_release_id, which
> makes no sense w/o source_id
>  (2) the pair (external_database_release_id, source_id) has "universal"
> semantics (ie, not just limited to "external" sequences).  that
> semantics is "dataset and stable_id"
>
> steve
> _______________________________________________
> ApiDB mailing list
> Ap...@pc...
> https://mail.pcbi.upenn.edu/mailman/listinfo/apidb
>