#234 gene allele association bug

[Anonymous]

Reported by lreiser:
Hello
I have a real problem with the logic of the function
that allows you to make an association between an
allele and a gene. I have a list of alleles from a
paper describing ACR4-Arabidopsis CRinkly4.

Problem 1: Alleles associated to one gene model for a
locus are NOT necessarily associated to the other gene
models.. for whatever reason.

Problem 2: Currently you have to make this association
manually- one allele at a time.

Problem 3. This association relies on typing in the
NAME of the gene model to associate.

That is the big problem-as we know NAMES are not
unique.So because there is another gene called ACR4 in
the database- the mutations become associated to the
wrong gene model (which oddly enough also gets
assocaited to the AGI gene models for that same
incorrect locus).

It seems to me we need to have one of two functions.

1. Propagate alleles to ALL gene models for this same
locus function

or
2. The ability to add a UNIQUE identifier (e.g.
pub_gene_id) rather than a name in this field.

I dont know what the long term plan is for propagating
alleles to all gene models is-if it will be done as
part of assignment - if not- if via pub- it would be
nice to make this easier. But the most important aspect
is to resolve the ambiguity.. I only caught this
because one allele became associated to more than one
AGI locus id...

Also I dont know what germplasms are in Pub but for all
of these alleles there is a corresponding tilling line-
none of which are listed in the 'associated
germplasms'. Id prefer to update the record for an
existing germplasm with phenotype data, rather than
create a potentially duplicate record.

[Danny Yoo]

Logged In: YES
user_id=49843

tanya says: There is a long-term plan for propagating
alleles to all gene models. It's not in place yet. We may
be able to attack problem 3 by adding one stage for
verifying what genes will be used to attach to the allele.