Category:RefGenome Electronic Jamboree 2010-02 NIPBL

See the talk page for discussion. See also Help:Annotation Jamborees

Orthoset

Ortholog set at PPOD

Participants

Name	Group	Organism(s)/Genome(s)	Notes
Stan Laulederkind	RGD	Rat
Ruth Lovering	BHF-UCL	Human
Varsha Khodiyar	BHF-UCL	Human	participated in annotation but not attending conference call
Yasmin Alam-Faruque	GOA	Human
Emily Dimmer	GOA	Human
Rachael Huntley	GOA	Human
Dianna Fisk	SGD	S. cerevisiae
Petra Fey	dictyBase	D. discoideum
Pascale Gaudet	dictyBase	D. discoideum
Li Ni	MGI	Mouse
Susan Tweedie	flyBase	D. melanogaster
Kimberly Van Auken	WormBase	C. elegans
Ranjana Kishore	WormBase	C. elegans
Donghui Li	TAIR	A. thaliana
Lakshmi Pillai	AgBase	G. gallus
edit table

Annotations

Aspect	ID	Term

Notes

As Wiki down adding comments here:

Term Requests

Accepted:

1. NTR: long-range enhancer-promoter communication - ID: 2952117 (Susan) covered by new term id: GO:0090203 name: transcriptional activation by gene looping

Unresolved:

1. NTR:transcriptionally active chromatin binding - ID: 2952055 (Susan) see comment below on ChIP experiments relevant to Drosophila Nipb, PMID:17965872^[1] binding to transcribed vs untranscribed regions of the DNA (Hetero- and euchromatin interactions)? There is GO:0043566 structure-specific DNA binding and GO:0031490 chromatin DNA binding
   Or should this just be captured through new component terms eg: New term: transcriptionally active chromatin
2. Placement of Transcription cofactor activity - ID: 2952243 (Ruth) will be dealt by Karen as part of overhaul of transcription factors.
   UK curators are not convinced that GO:0003714 transcription corepressor activity should be a child of GO:0003712 Transcription cofactor activity.
3. protein-promoter complex ID: 2953747 (Ruth)
4. NTR: enhancer binding - ID: 2952064 (Susan)
5. uterus morphogenesis - ID: 2940885 (Varsha)
6. NTR: abdominal wall muscle cell development - ID: 2947823 (Yasmin)
7. Broaden the definition & name of cellular component "cohesin loading complex" (GO:032116) - ID: 2956225 (Dianna)

Other Issues:

1. IMP annotation on developmental defects associated with the disease Cornelia de Lange Syndrome (CdLS) PMID:19242925^[2], however, in many ways this is an IC because I wouldn't have made this without reading a previous paper showing that this gene is involved in gene transcription regulation (Ruth).

2. Some development terms have embryonic and some do not. When should embryonic apply to a dev term? Just to a specifically embryonic structure which is not present in the adult? Or an adult struture (e.g. arm) which develops in utero (Varsha).

3. ChIP experiments should not solely be used as evidence for DNA binding as crossing linking between proteins can also occur (Rachael).

4. protein C/N-terminus binding. Can this term be applied to proteins binding a range of amino acids located at the C/N-terminal end region of a protein, or proteins binding the actual last amino acid?

Definition for GO:0008022 protein C-terminus binding: Interacting selectively and non-covalently with a protein C-terminus, the end of any peptide chain at which the 1-carboxy function of a constituent amino acid is not attached in peptide linkage to another amino-acid residue.

5. TAIR has annotated this protein to 'sister chromatid cohesion' based on PMID 19228337^[3]. Upon looking at the paper again, a more accurate term would be the child term 'meiotic sister chromatid cohesion' GO:0051177. We'll correct this. (Donghui).

Minutes

Minutes from Dianna (SGD); other participants should feel free to fix any errors or omissions!

Action items

1. Look into defining chromatin more explicitly
2. Look into removing protein C/N-terminus binding terms
3. Look into assessing whether embryonic-specific development terms are genuinely useful or not.

Discussion topics

Accepted:

1. NTR: long-range enhancer-promoter communication - ID: 2952117 (Susan) covered by new term id: GO:0090203 name: transcriptional activation by gene looping
No time for discussion of this topic.

Unresolved:

1. NTR:transcriptionally active chromatin binding - ID: 2952055 (Susan)
   There were a couple of general issues that came up in the discussion of this new term:
   1. When should things be annotated to a function term vs. a component term? That is, should this be annotated to both a "chromatin binding" function term AND a "chromatin" component term? Although some people seemed fine with annotating to seemingly overlapping component and function terms, the conceptual conflict here is the idea that anything that is part of the chromatin would intrinsically have the function of chromatin binding.
      • Can we define chromatin more explicitly? For example, should it be defined in a static state with explicit protein components? If so, should transcription factors and RNA be considered part of chromatin?
      • Should we remove the "chromatin binding" terms? This seemed like a solution to some participants, but Dianna felt strongly that this term should remain, arguing that for some complexes (e.g. cohesin, condensin) chromatin binding, in a dynamic manner, really is their job.
   2. What is the difference between the chromosome and chromatin terms? Between the chromosome binding, chromatin binding, and chromatin DNA binding terms? When would you choose one of these terms over the other? Are there experiments that can distinguish these terms?
      • Jim argued the chromatin and chromosome were not equivalent terms; that chromatin implies structural features.
      • Dianna argued that the divide between chromosome and chromatin was an ontology construct to make things work for bacteria, since prokaryotes don't have chromatin in the same sense that eukaryotes do. With this logic, eukaryotes should always be annotated to the chromatin terms.
      • Nobody volunteered examples of how to distinguish these experimentally
2. Placement of Transcription cofactor activity - ID: 2952243 (Ruth)
   The issue is that not all co-repressors bind directly to co-factors, as indicated by the ontology structure. Everyone was happy with having Karen consider this in her in-process transcription revamp.
3. protein-promoter complex ID: 2953747 (Ruth)
   There is some discussion of this on the sourceforge item. The proposed solution is to just use nuclear chromatin and the SO terms, promoter or enchancer, in column 16. Ruth feels that this leaves us with a GO annotation that is two broad. We did not have time for a full discussion of this item, but Ruth encourages people to comment on the sourceforge item.
4. NTR: enhancer binding - ID: 2952064 (Susan)
   No time for discussion of this item.
5. uterus morphogenesis - ID: 2940885 (Varsha)
   No time for discussion of this item.
6. NTR: abdominal wall muscle cell development - ID: 2947823 (Yasmin)
   No time for discussion of this item.
7. Broaden the definition & name of cellular component "cohesin loading complex" (GO:032116) - ID: 2956225 (Dianna)
   Midori agreed that this definition was too broad and will be happy to update with whatever we decide on. There did not seem to be strong preferences for the name of the complex. Based on another discussion, it is agreed that it's important to include members of the complex in the definition, whenever possible, making sure that it's clear these are examples.

Other Issues:

1. IMP annotation on developmental defects associated with the disease Cornelia de Lange Syndrome (CdLS) PMID:19242925^[2], however, in many ways this is an IC because I wouldn't have made this without reading a previous paper showing that this gene is involved in gene transcription regulation (Ruth).
   This item was discussed in the context of a couple of larger issues:
   • When is it appropriate to annotate to high level development (or morphology) terms, more refined molecular terms, or both? Tanya explained that it's first determined by what evidence exists. It's not uncommon for the initial publications to describe a mutant phenotype, with a developmental defect, and then later publications to describe much more explicit functions or processes. You should always annotate based on whatever evidence is available. Once you've done that, the question becomes, "When do we keep or remove the phenotype-based annotations?" At TAIR, their policy is to keep the developmental terms if they think that their users would expect to see them. Some participants suggested that one would expect all orthologs to have the same development-type annotations, across organisms. Others disagreed with this expectation.
   • Is it possible to have some sort of automated system to alert us when old annotations should be removed? It was pointed out that all groups will add GO annotations based on new publications without necessarily reviewing the older annotations. No one has the resources to constantly review existing annotations. There was some discussion of how PAINT might be able to alert people to instances where there has been a fundamental change in the annotation of genes, since they get alerts whenever annotations are added to or removed from a reviewed family. It's possible that groups could receive notices when this happens, or when their gene has annotations that the PAINT group has decided not to propagate (these may indicate errors or "over-annotations").
2. Some development terms have embryonic and some do not. Why? (Varsha).
   Midori explained that , historically, there are embryonic development terms only for things that actually occur in the embryo, regardless of whether or not development continues in adults. It's possible that we should reexamine these terms and decide what benefit we get from having them. There is also a group on people working on human development right now, so maybe they can contribute to the assessment of these terms.
3. ChIP experiments should not solely be used as evidence for DNA binding as crossing linking between proteins can also occur (Rachael).
   Everyone agreed that, when these experiments are done with in vivo samples, this was the correct way to treat these experiment types.
4. protein C/N-terminus binding.
   It was pointed out that protein binding is usually dependent on a protein-binding domain, and whether that domain is at the C- or N-terminus is kind of irrelevant to the function. People agreed that these terms were generally useless and didn't actually convey any more functional information than the parent term. Participants will go back to their groups and see how and why these terms are being used, and we'll look into removing them and mapping everything up to the parent term, protein binding.
5. TAIR has annotated this protein to 'sister chromatid cohesion' based on PMID 19228337^[3]. Upon looking at the paper again, a more accurate term would be the child term 'meiotic sister chromatid cohesion' GO:0051177. We'll correct this. (Donghui).
   No discussion of this item.
6. When can you ISS to a complex term?
   Do people make ISS annotations to complexes? If so what are your criteria? For many groups, this set of papers was a nice example of when researchers seem to regard a fact as common knowledge (in this case that NIBPL orthologs are part of the cohesin loading complex), even when it hasn't been explicitly demonstrated. This left many curators in a situation where they would have liked to ISS to the human or yeast data, but were unsure exactly what was acceptable practice. It was suggested that if all components of the complex are conserved, then it is appropriate to make an ISS. When you're using an ISS in this way, you can use internal GO reference number 24. Based on this policy, it's important to try to include a list of complex members in complex term definitions, whenever possible. It was also suggested that, if people don't want to do this ISS themselves, they can just wait for the PAINT propagation.

References

See Help:References for how to manage references in GONUTS.