Annotation Conf. Call 2015-10-27

From GO Wiki
Jump to navigation Jump to search

Meeting URL:


Annotation Consistency Exercise

Annotation consistency paper picked by ZFIN- PMID:24430697

Choice of paper-motivation: discussing the use of IGI as evidence code. When should we use IGI?

- mutation in gene A is rescued by a mutation in gene B
- mutation in gene A is rescued by over-expression of gene B

In this paper, how do we capture

- the repression of transcriptional activation activity of tbx5a by kctd10?
- the information about has2

Summary of the paper (PMID:24430697)

- mutants in kctd10 show heart defects (heart looping, atrioventricular canal development)

Note: phenotypes determined by

- analysis of heart morphology
- gene expression analysis of genes known to be expressed in specific part of the heart).

[Fig 1, 2, 3]

- downregulation of tbx5a or has 2 (using MO) rescue kctd10 mutant heart phenotypes. [Fig 4]

- in vitro experiments [Fig5]:

- luciferase reporter assay
kctd10 inhibits tbx5 transcriptional activity
- Co-IP ; Pull down assay:
tbx5a and kctd10 physically interact with each other.

Annotations: 9 sets of annotations received.

Analysis and points of discussion: (in no particular order)

- Most of us annotated kct10 to heart looping/atrioventricular canal development with IMP

- Use of IEP: One annotation used IEP (Inferred from Expression Pattern) as evidence code because the annotation based on altered expression of AVC marker genes (Fig 2)

kctd10 atrioventricular canal development GO:0036302 IEP

- “negative regulation of transcription from RNA polymerase II promoter” (GO:0000122) based on the change in gene expression assessed by in situ hybridization (Fig2). examples:

kctd10 negative regulation of transcription from RNA polymerase II promoter GO:0000122 IMP regulates_transcription_of(anf)|regulates_transcripition_of(amhc)|regulates_transcripion_of(Bmp4)|regulates_transcription_of(Notch1b)
kctd10 negative regulation of transcription from RNA polymerase II promoter GO:0000122 IMP regulates_transcription_of: tbx2b, occurs_in myocardium, happens_during cardiac chamber morphogenesis

- Is there enough evidence for the following annotations?

kctd10 eye development GO:0001654 IMP
kctd10 head development GO:0060322 IMP

It is possible that the eye/head phenotype is a secondary phenotype.

What is the timing of the different organs' development during zebrafish embryogenesis? Would this help to determine the likelihood of a secondary phenotype? Are there other mutations that affect heart development but not development of these other tissues? (Kimberly)

- The terms including “development”, “morphogenesis”, “formation” are confusing. What is the difference? Which one should we use in this context?

- Statements from paper:

Figure 1. (i-k) Both the myocardium (green) and the endocardium (red) are substantially smaller in the mutant and the morphant, and no cavity is visible inside the heart lumen.

Figure 4. (a,f) The endocardium of the mutant was extremely shrunken and the space between the myocardium and endocardium was much wider than that of the wild-type.


kctd10 endocardium development GO:0003157 IMP
kctd10 endocardium morphogenesis GO:0003160 IMP
kctd10 endocardium formation GO:0060214 IMP

- in Figure 3: hyaluronic acid staining was performed to visualize the cardiac jelly. Annotation:

kctd10 negative regulation of hyaluronan biosynthetic process GO:1900126 IMP

Is there enough evidence that kctd10 is involved in hyaluronan biosynthesis? Could it be possible that the increase in staining is a secondary effect?

- Fig4: injection of tbx5MO and has2MO rescue kctd10 mutant phenotype: Most of us annotated using IGI between tbx5 and kctd10, and between has2 and kctd10 heart looping/atrioventricular canal development

- Should kct10 be annotated as “positive/negative regulation” of heart looping / atrioventricular canal development?

- Annotation:

has2 negative regulation of heart looping GO:1901208 IGI kctd10

- Fig 5: regulation of transcription : which evidence code should be used?

kct10 negative regulation of sequence-specific DNA binding transcription factor activity GO:0043433 IDA
kct10 negative regulation of sequence-specific DNA binding transcription factor activity GO:0043433 IDA has_regulation_target(Tbx5a)
kct10 negative regulation of sequence-specific DNA binding transcription factor activity GO:0043433 IDA tbx5a has_regulation _target(Tbx5a)
kct10 negative regulation of sequence-specific DNA binding transcription factor activity GO:0043433 IGI tbx5a has_regulation _target(Tbx5a)
kct10 negative regulation of sequence-specific DNA binding transcription factor activity GO:0043433 IMP has_regulation_target: tbx5b

- protein binding: which evidence code should be used? examples:

kct10 protein binding GO:0005515 IPI tbx5a
kct10 transcription factor binding GO:0008134 IPI tbx5a
kct10 activating transcription factor binding GO:0033613 IPI
kct10 RNA polymerase II activating transcription factor bindin GO:0001102 IDA has_direct_input Tbx5
kct10 transcriptional repressor activity, RNA polymerase II transcription factor binding GO:0001191 IPI tbx5a


  • On call: Chris, David H, Doug, Edith, Kimberly, Li, Midori, Moni, Paul T., Petra, Robert, Ruth, Sabrina, Shur-Jen, Stacia, Stan, Tanya
  • Rationale for picking this paper:
    • Use of IGI mutation or over expression of one gene rescues phenotype for another
    • Annotating transcriptional repression experiments
  • All agreed to annotate Kctd10 to heart looping
    • Question about use of annotation extension happens_during
      • How does happens_during relate to the part_of relation in the ontology?
      • happens_during means coincident in time
      • during is going to be deprecated
      • but exists_during and happens_during will remain
      • no demonstrated causal relation exists between happens_during and the process
      • in this case, not necessary to add the happens_during because of the existing ontology structure
  • Due to the transcription component of the process, should we have included ‘regulation of heart looping’?
    • In this case, transcription is part_of the development of the structure, so not necessary to use the ‘regulation of heart looping”.
    • Developmental processes can be complex, so sometimes it’s hard to draw the line when deciding what is development and what is regulation of development.
    • Ask: is the action of that gene part of the biological program?
  • Annotating kctd10 to negative regulation of transcription based on Figure 2.
    • Question: is in situ hybridization data enough evidence to make a transcription annotation?
    • Paul T: why are the authors performing this experiment? They were looking at these markers to establish that kctd10 affects AVC development, and not something more upstream.
    • Sabrina: would not have made this annotation
    • David H: thought that the transcription evidence was elsewhere
    • Ruth: did make these annotations, but found the overexpression phenotype informative in the context of the rest of the paper
    • Paul T: also found the overexpression phenotype information significant
    • David H: Run-ons are best evidence for regulation of transcription, but those experiments are not generally performed and people take other experiments (e.g., in situ, Northerns, etc.) as evidence for regulation of transcription
    • David H: also made annotation to anatomical boundary formation using annotation extensions
    • Ruth: we are always encouraged to annotate to our level of understanding
  • Some annotations made to eye and head development
    • Sabrina: not enough evidence to show that the gene was involved in eye and head development
    • Based on developmental timing and co-occurrence of phenotypes, the pectoral fin phenotype may indicate a primary defect, while the eye and head development is more likely to be secondary. As the heart did not form correctly there has to be concern that developmental processes elsewhere in the embryo might be affected by the lack of sufficient blood supply to that tissue/organ.
  • Development vs Morphogenesis vs Formation
    • For example, endocardium terms
    • What should be used in this case? There was a broad range of term selection here.
    • These terms are children of one another in the ontology, but development is the least granular term and includes the other two. Morphogenesis specifically refers to the shape of the structure; formation is just the very beginning of morphogenesis, i.e., getting the shape from nothing. Where formation begins and end can be a difficult call, but morphogenesis and formation both have to do with the shaping of a structure.
    • In this case, what would be the correct term to use?
    • If you aren’t sure, then development is the correct term.
  • Role of hyaluronan biosynthetic process
    • Does kctd10 negatively regulate the hyaluronan biosynthetic process?
    • Because of the effect on Has2 biosynthesis, kdtc10 could be annotated to regulation of hyaluronan biosynthetic process.
    • Paul T: does this make more sense to capture this as a regulator of Has2 molecular function
    • David H: if you have access to noctua, there are some models for this paper on noctua
    • Discuss one of the LEGO models on the next call?
  • In vitro transcription studies
    • Pretty much everyone agreed on the term, but there was different evidence code usage: IDA, IGI, and IMP with and without annotation extensions
    • Ruth: would have used IGI, but you lose the semantics in the with column (extends the evidence code) while the annotation extension extends the GO term. But there needs to be consistent capture of ID information. In some cases it might be appropriate to add to both the with column and the annotation extension column.
  • Protein binding
    • Mostly used IPI but there was an IDA with has_direct_input annotation extension
    • When you know the two entities interacting (e.g., have a database identifier) use IPI
    • Otherwise, you can use IDA (e.g., actin binding)
    • Protein binding annotations made using IDA will not load in Cytoscape
    • Is there an annotation issue here?