Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Ambiguous mapping of some EMAPA terms #664

Open
fbastian opened this Issue Mar 27, 2015 · 8 comments

Comments

Projects
None yet
3 participants
@fbastian
Copy link

fbastian commented Mar 27, 2015

Some EMAPA terms are xrefs of several UBERON terms, so it is difficult for us to map (I know we can use taxon equivalent classes, but when there are only xrefs available, we use xrefs).

Is it possible to try to disambiguate some of these xrefs? I guess some are intended, but maybe not all.

2 xref: EMAPA:16098
2 xref: EMAPA:16115
2 xref: EMAPA:16133
2 xref: EMAPA:16276
2 xref: EMAPA:16291
2 xref: EMAPA:16294
2 xref: EMAPA:16483
2 xref: EMAPA:16564
2 xref: EMAPA:16589
2 xref: EMAPA:16678
2 xref: EMAPA:16748
2 xref: EMAPA:16762
2 xref: EMAPA:16775
2 xref: EMAPA:16850
2 xref: EMAPA:16914
2 xref: EMAPA:17213
2 xref: EMAPA:17272
2 xref: EMAPA:17298
2 xref: EMAPA:17301
2 xref: EMAPA:17339
2 xref: EMAPA:17342
2 xref: EMAPA:17416
2 xref: EMAPA:17423
2 xref: EMAPA:17426
2 xref: EMAPA:17492
2 xref: EMAPA:17495
2 xref: EMAPA:17498
2 xref: EMAPA:17501
2 xref: EMAPA:17596
2 xref: EMAPA:17681
2 xref: EMAPA:17823
2 xref: EMAPA:18025
2 xref: EMAPA:18223
2 xref: EMAPA:18577
2 xref: EMAPA:19133
2 xref: EMAPA:28130
2 xref: EMAPA:28346
2 xref: EMAPA:28370
2 xref: EMAPA:29794
2 xref: EMAPA:30763
2 xref: EMAPA:30997
2 xref: EMAPA:32628
2 xref: EMAPA:32637
2 xref: EMAPA:32645
2 xref: EMAPA:32719
2 xref: EMAPA:32730
2 xref: EMAPA:32759
2 xref: EMAPA:32947
2 xref: EMAPA:35133
2 xref: EMAPA:35313
2 xref: EMAPA:35505
3 xref: EMAPA:16139
3 xref: EMAPA:17748
3 xref: EMAPA:18606
3 xref: EMAPA:28393
3 xref: EMAPA:32705
@cmungall

This comment has been minimized.

Copy link
Member

cmungall commented Mar 27, 2015

I'll do a round of filtering on these and then recruit @tfhayamizu for any questions that come up

@cmungall

This comment has been minimized.

Copy link
Member

cmungall commented Mar 27, 2015

Hmm, tricky. On the one hand we want to rely on MA, which provides nice xrefs:

[Term]
id: MA:0000577
name: head connective tissue
namespace: adult_mouse_anatomy.gxd
xref: EMAPS:1609828
is_a: MA:0000570  ! head or neck connective tissue
relationship: part_of MA:0000023 ! head

EMAPS:1609828 = EMAPA:16098 ! head mesenchyme, at stage 28

So this is the closest match to EMAPA in MA (MA is adult). We take this as the xref for UBERON:0003566 'head connective tissue'

However, we have independently aligned this to UBERON:0005253 'head mesenchyme', which is the more appropriate mapping for a pan-stage perspective.

We could always merge the two uberon classes here, it's an almost transformation_of... but this could be problematic

Need to think about the optimal strategy for sharing the work here.

I'll look at the others a bit more systematically.

@tfhayamizu

This comment has been minimized.

Copy link

tfhayamizu commented Mar 27, 2015

This will likely be an issue for "mesenchyme" and "connective tissue" across the board -- terms used synonymously, although not really correct. In most of literature (at least for the mouse, probably human), embryos have mesenchyme and adults have connective tissue. Crossing the divide is tricky.

cmungall added a commit that referenced this issue Mar 27, 2015

@cmungall

This comment has been minimized.

Copy link
Member

cmungall commented Mar 27, 2015

report:
https://github.com/obophenotype/uberon/blob/master/scratch/ubr-to-many-emapa.tsv

we have a mixture here. In some cases, the EMAPAs represent the same structure at different stages, where Uberon collapses these stages. It's useful if confusing to maintain a 2-1 mapping to Uberon here.

In other cases (e.g.

EMAPA:17596 ! otic capsule UBERON:0005410 ! cartilaginous otic capsule __n/a__
EMAPA:17596 ! otic capsule UBERON:0005411 ! bony otic capsule MA:0000252 ! otic capsule
) uberon has different representations for different stages, and these could be better mapped to the more precise EMAPA stage based on timings.

Sometimes two MAs will point to a single EMAPA, where I think they should point to one, e.g.

EMAPA:18577 ! sciatic nerve UBERON:0001322 ! sciatic nerve MA:0001172 ! sciatic nerve
EMAPA:18577 ! sciatic nerve UBERON:0002004 ! trunk of sciatic nerve MA:0001178 ! sciatic nerve trunk
- my preference would be to fix this in MA

If someone wants to fork the github repo and annotate the table with what they think should be done, I can integrate this back in. Otherwise I will likely fix these gradually, putting requests in upstream trackers, with no fixed timeline...

note to self for regenerating table:

blip-findall -r MA -i z.pro  -r uberonp -r EMAPA "m(X),entity_xref(U,X),(entity_xref(U,M),entity_xref(M,E),atom_concat('EMAPS:',N,E),atom_concat(ID,'28',N),atom_concat('EMAPA:',ID,X)->S=M;S='__n/a__')" -select "x(X,U,S)" -label -no_pred | sort -u > scratch/ubr-to-many-emapa.tsv
@cmungall

This comment has been minimized.

Copy link
Member

cmungall commented Mar 28, 2015

For the anterior prostate/coagulating gland, I budded off #665

@cmungall

This comment has been minimized.

Copy link
Member

cmungall commented Mar 28, 2015

@tfhayamizu would you agree with the following:

EMAPA:18577 ! sciatic nerve UBERON:0001322 ! sciatic nerve  MA:0001172 ! sciatic nerve  Y
EMAPA:18577 ! sciatic nerve UBERON:0002004 ! trunk of sciatic nerve MA:0001178 ! sciatic nerve trunk    N - Remove xref in MA, the above one is sufficient?

EMAPA:32705 ! limb mesenchyme   UBERON:0003587 ! limb connective tissue MA:0000689 ! limb connective tissue TRANSFORMATION_OF
EMAPA:32705 ! limb mesenchyme   UBERON:0009749 ! limb mesenchyme    __n/a__ TRANSFORMATION_OF
EMAPA:32705 ! limb mesenchyme   UBERON:0010740 ! bone of appendage girdle complex   MA:0000688 ! limb bone  N - remove xref in MA

EMAPA:35505 ! long bone epiphyseal plate    UBERON:0002516 ! epiphyseal plate   MA:0002839 ! long bone epiphyseal plate Y
EMAPA:35505 ! long bone epiphyseal plate    UBERON:0006861 ! diaphysis proper   MA:0002803 ! long bone diaphysis    N - Remove xref in MA. Also review 'diaphysis' vs 'diaphysis proper'.
@tfhayamizu

This comment has been minimized.

Copy link

tfhayamizu commented Mar 28, 2015

Fine with me. Looks like there are some errors in the file. I will make the changes to the MA xrefs. And we clearly need to re-evaluate the mesenchyme-connective tissue relationship within the context of the EMAPA overall. Develops-from relationships have been part of the plan for a long time, but it's not clear whether this case fits that.

cmungall added a commit that referenced this issue Mar 28, 2015

cmungall added a commit that referenced this issue Mar 28, 2015

ensured transformation_of holds between all CT and mesenchymes. Note …
…that we still have the unsatisfactory situation of MA xreffing the earlier structure, TBD. Obsoleted skeleton of embryo as this was not useful. Made post-anal tail follow transformation-of edges. We retain the embryonic tail as its a useful grouping for humans and other chordates that lack the adult post-anal tail. #664
@cmungall

This comment has been minimized.

Copy link
Member

cmungall commented Mar 28, 2015

Also, for

EMAPA:30763 ! ovary tertiary follicle   UBERON:0000036 ! secondary ovarian follicle MA:0002909 ! ovary antral follicle  N - I think antral is considered 3rd but confusing 

See https://sourceforge.net/p/geneontology/ontology-requests/11601/

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
You can’t perform that action at this time.