Annotation Conf. Call, May 26, 2015
- 1 Agenda
- 2 Unresolved questions from Curation Consistency exercise
- 3 Ion channels
- 4 Discussion
Unresolved questions from Curation Consistency exercise
April Mouse paper- http://www.ncbi.nlm.nih.gov/pubmed/?term=24349431
- Link to Annotations from the April exercise- http://wiki.geneontology.org/index.php/Annotation_Conf._Call,_April_28,_2015
- Is cystin regulating necdin? What would constitute evidence for this? The paper does show physical interaction between the two proteins and an effect on activity when co-transfected. How would you capture the fact that cystin has the antagonist effect when necdin is present but it has the opposite role when it is acting alone? Would you use the evidence code or Col-16?
- Would you use IGI as the evidence for how cystin interacts with necdin? Is cotransfection considered IGI? Both Rama and Midori thought this was not IGI based on current definition. Ruth, Rebecca and Rachael felt this would be IGI. Kimberly - I had originally used the IDA evidence code for the co-transfection experiment, but it seems that what we are really trying to capture with annotating this experiment is a functional interaction between two gene products. Although the type of experiment differs from traditional genetic experiments using multiply mutant strains, the conclusion is similar: these gene products functionally interact, even if we don't know the exact mechanism. The IGI evidence code therefore seems more appropriate to me now.
- Reason for use of IGI: 2 cDNA constructs (cystin and necdin) transfected into a cell line, the effect is only seen when both constructs are present, therefore the use of IGI code enables the cystin to be annotated to negative reg of transcription (child term), and the addition of the necdin protein ID in the WITH field (the reciprocal annotation would not be created).
- Reason against use of IGI: When only one of these genes is transfected into the cells the annotations created (pos reg of transcription, child term) are supported by the IDA evidence code not the IMP evidence code. The GOC documentation on IGI states: Includes any combination of alterations in the sequence (mutation) or expression of more than one gene/gene product.
- Reason for using Col-16: The evidence code for the single transfection is IDA, therefore the IGI code can’t be used. Using Col-16 enables this information to be captured.
- Reasons against using Col-16: There is no suitable relationship available to capture this data. There would be a lack of consistency in the application of evidence codes, e.g. If mutant and wt double transfectants were used then the IGI evidence code would be applied, and there would be information in the WITH field. Where as double wt transfectants would have no information in the WITH field.
Useful to look at GO evidence code definitions (only part of definitions included):
- IDA: For transfection experiments or other experiments where a gene from one organism or tissue is put into a system that is not its normal environment, the annotator should use the author's intent and interpretation of the experiment as a guide as to whether IMP or IDA is appropriate. When the author is comparing differences between alleles, regardless of the simplicity or complexity of the assay, IMP is appropriate. When the author is using an expression system as a way to investigate the normal function of a gene product, IDA is appropriate.
- IMP:transfection into a cell line, overexpression, or extopic expression of a gene where the effects of various alleles of a gene are compared to each other or to wild-type. For this type of experiment, annotate using IMP.
- Transfections into a cell line, overexpression, or ectopic expression of a gene when the expression system used is considered to be an assay system to address basic, normal functions of gene product even if it would not normally be expressed in that cell type or location. If the experiments were conducted to assess the normal function of the gene and the assay system is believed to reproduce this function, i.e., the authors would consider their experiment to be a direct assay, and not a comparison between various alleles of a gene, then the IDA code should be used. This is in contrast with a situation where overexpression affects the function or expression of the gene and that difference from normal is used to make an inference about the normal function; in this case use the IMP evidence code.
- IGI: Includes any combination of alterations in the sequence (mutation) or expression of more than one gene/gene product. ..... If there is a single mutation or difference between the two strains compared, use IMP. If there are multiple mutations or differences between the two strains compared, use IGI.
- Note that the IGI description mentions changes in expression or more than one gene product, not just mutation in the gene.
In this experiment (fig 6C) we are not told what the level of necdin and cystin are wrt normal levels. The transfection of either leads to an increase in transcription of Myc promoter (more of an increase with necdin than cystin). The transfection of both leads to an increase of Myc transcription, but this is reduced compared to that with necdin alone. This must imply that the level of cystin transfected in increases the normal level, if this was the normal level then the amount of Myc transcription would be the same as it is with necdin alone.
This would then support the annotation using the IGI evidence code. However, with this information is it now more appropriate to annotate the single transfection expts to IMP?
Note that in the ECO ontology: over expression analysis evidence and ectopic expression evidence (and RNAi data) are considered 'experimental phenotypic evidence' and 'genetic interaction evidence' does not include transfections.
- > + evidence
- >> + experimental evidence
- >>> + experimental phenotypic evidence
- >>>> + genetic interaction evidence
- >>>> + mutant phenotype evidence
- >>>> - RNAi evidence-anti-sense experiment evidence
- >>>> - over expression analysis evidence
- >>>> - ectopic expression evidence
- RNA pol II binding- this is redundant with RNA Pol II txn factor complex annotation. Is there a link between the two? This is not always true, so no. It is not clear why this is not always true from the minutes, is it because the definition for GO:0090575 RNA polymerase II transcription factor complex does not state RNA pol II binding is essential for these complexes: A transcription factor complex that acts at promoters of genes transcribed by RNA polymerase II?
Is the decision here to annotate to both if this is shown? Does this paper support the annotation to GO:0090575 RNA polymerase II transcription factor complex?
- I don’t understand what is meant in the minutes by: annotating to chromtin binding and putting Myc1 in col-16 is not right. It is upstream. David pointed out that we can include C16 information to show the region of either chromatin binding or if nuclear chromatin is used then C16 info can be included here. How can this be done? From the table annotations is the correct option: chromatin binding: has_input (Myc,promoter ; SO:0000167; 1 annotation) in which case what ID should be used for Myc1? nuclear chromatin (coincident_with('GN box' or Myc P1 promoter' SO:new). It would be good to have this clearly stated. coincident_with does not have any C16 usage examples (nor is it listed) on the C16 relationship wiki.
- Chromatin- is not just histones and DNA. Should Txn factors be annotated to chromatin? Would you make a CC annotation to chromatin over a MF annotation to chromatin binding when you have ChiP evidence? Karen mentioned that in the paper she worked on with the Norway folks they decided all ChiP experiments should be CC. David mentioned that the way he decides between the two is : When you are binding chromatin you are not part of the chromatin. Is txn factor part of chromatin? We need to resolve this issue at the next GOC meeting (Midori has some rules, so does David). We will collect some use cases and present it at the GOC meeting. What should be included in the definition of chromatin? Note that we also have terms like 'transcriptionally active chromatin' and 'transcriptionally silent chromatin'.
- histone modifying complexes? - these are not children of chromatin in the CC
- chromtin re-modeling complexes? - these are also not children of chromatin in the CC
- transcription factors?
- cystin binds to necdin, and also this binding is required in order for cystin to repress necdin positive regulation of transcription. I would like to use the GO term (or child) GO:0016564 transcription repressor activity. However this is obsolete. The GOC has introduced a rule that in order to annotate to an inhibitor or activator activity the inhibited/activated protein has to be bound by the inhibitor/activator. Why not have this rule for new child terms to the GO:0016564 transcription repressor activity term, eg: new term: transcription factor binding transcription repressor activity; new term: DNA binding transcription repressor activity. Note that GO:0003714 transcription corepressor activity describes repressing transcription factor in the definition: Interacting selectively and non-covalently with a repressing transcription factor and also with the basal transcription machinery in order to stop, prevent, or reduce the frequency, rate or extent of transcription. Cofactors generally do not bind the template nucleic acid, but rather mediate protein-protein interactions between repressive transcription factors and the basal transcription machinery.
- Should transfected experiments ever be annotated to the cell type that the constructs are transfected into? In this case, the authors had demonstrated that cystin and necdin are expressed in these cell types therefore their intention was to elucidate the role of these proteins in these cells. UCL team have not added cell type information in C16, however there was some discussion about whether or not to do this as the protein data was not based on the endogenous protein, and we did not reach an agreement.
March Yeast paper- http://www.ncbi.nlm.nih.gov/pubmed/12048186
- Use of in_the_presence_of and in_the_absence_of relationships in col-16.
- these relationships are slated for obsoletion. We could capture the data using different relationships
- BFA1 has inhibitory role on the Tem1 GTPase when it is acting alone. It is okay to say that it is acting alone and no need to use in_the_absence_of: Bub2.
- Bub2 gets GAP activity that can activate Tem1 GTPase (when present together as a complex with Bfa1) . part_of: BFA1-BUB2 complex, has_regulation_target:Tem1p
- Would you annotate Bfa1 to GAP activity or contributes to GAP activity? Data in the paper points to BFA1 being an inhibitor by itself. So may it shouldn't get this term. May be its role is to bridge Tem1 and Bub2 but there is not enough data in the paper to give it a protein binding, bridging annotation
Update on ion channels. Background: http://wiki.geneontology.org/index.php/Annotation_Conf._Call,_May_12,_2015#Ion_channels-_Ontology_overhaul_.28David.29_2
Based on discussion at the previous annotation call and among editors, below is a revised proposal. This hasn't been implemented in the ontology yet.
1) The recently created terms:
- non-selective anion channel activity
- non-selective cation channel activity
will be obsoleted (reason: added in error). Unless (see very end of proposal).
2) All existing terms under 'ion channel activity' will be meant to be agnostic to the notion of selectivity or non-selectivity, that is, their definitions will simply state that the channel is capable of passing a given ion, regardless as to whether only that specific ion can pass through, or whether other ions can pass as well:
- we will remove 'selective' from the definition (as in "Enables the selective, facilitated diffusion of...");
- we will remove synonyms containing the 'specific' string (e.g. 'chloride-specific channel activity');
- we will add 'non-selective anion channel activity' as a narrow synonym to 'anion channel activity';
- we will add 'non-selective cation channel activity' as a narrow synonym to 'cation channel activity'.
3) We will create new terms for selective channels. This should be done on a case-by-case basis. We could either draw a list of known selective channels, or use the list of non-selective channels that Huaiyu had offered to provide. Also see point 4) below about annotations.
- These new terms would be called e.g. 'selective calcium channel activity' and
- would be given exact synonyms e.g. 'calcium-specific channel activity'.
- The definitions would read "Enables the selective, facilitated diffusion of...".
If we go this route, all current annotations would be correct, though some of them (the ones to selective channels) would not be as granular as they could. These are the ones we'd want to re-annotate. General rules for annotating ion channels should be:
- If a paper gives evidence that a channel is selective for a given ion (or if the curator knows from prior knowledge), the channel component(s) should be annotated to a selective term, e.g. 'selective calcium channel activity'. If the term doesn't exist, the curator should request it.
- If a paper shows that a channel passes a given ion, but does not specify if the channel is selective for that ion or not (or if the curator does not know from prior knowledge), the channel component(s) should be annotated to one of the existing term as appropriate, e.g. 'calcium channel activity'.
- If a paper shows that a channel passes more than one ion and is therefore non-selective (e.g. TRP channels let sodium, potassium and calcium through), the channel component(s) should be annotated to as many existing terms as appropriate (e.g. all three 'sodium channel activity', 'potassium channel activity' and 'calcium channel activity' provided the paper shows experiments for all of them!).
5) End results:
- Selective channel proteins could be pulled out as such => gain of information;
- Non-selective channels would not be pulled out as such; the assumption would be that if selectivity isn't specified, the channel is either non-selective or of unknown specificity. If this is unwelcome, we could leave the recently created terms 'non-selective anion channel activity' and 'non-selective cation channel activity' in the ontology, and ask curators to move their existing annotations there if they belong to the list of non-selective channels provided by Huaiyu.
Any comments please?
Issues from April consistency paper
- we discussed the annotation of cystin to GO:43433 (negative regulation of sequence-specific DNA binding transcription factor activity). Cystin negatively regulates the transcription activity of necdin. We can't use GO:1227 (RNA polymerase II transcription regulatory region sequence-specific DNA binding transcription factor activity involved in negative regulation of transcription)because the data doesn't support this term. We can't use IPI as evidence code because physical interaction alone doesn't establish the regulation part. For the annotation to GO:43433, don't put Myc in col-16 because cystin is regulating the transcriptiopn factor (necdin) not Myc.
- discussion about use of IGI: some curators felt it was okay to use IGI with cotransfection experiments. Transfections are over expressions, so it is fine to use IGI. But the issue is, typically we use IMP for the single mutant and in this case there is no room to use IMP when one gene product is studied. Are all transfections over expressions? May be not! It is also the cultural difference in how these experiments are done and perceived in different systems (rat vs mouse vs worm). If we use IDA for this term then we have to use col-16 to indicate necdin as being present in the assay and that is not optimal.
- if we decide to use IGI we need a more granular version of it (using ECO). But where would we place it in the ECO? If we place it under IGI then it becomes a flavor of IGI.
- Rama: I would like to consult other curation projects: e.g. bioGRID to see how they treat co-transfections. We should be consistent in how we analyze this data.
- if people come across more of these kinds of experiments please send me (rama) an email and we will discuss at the upcoming GOC meeting
- RNA pol II binding and RNA pol II Txn factor complex are not redundant. Make separate annotations.
- Txn factors should not be annotated to chromatin. Ontology editors will add a comment to the term saying so. Rama will make a SF item.
- Txn repressor and Txn co-repressor. Cystin is a repressor of necdin and this is covered by 43433. Co-repressors bind to repressors. This is a bit complicated because at the time transcription terms were overhauled we decided to make function terms only when the mechanism or 'how' something happens. At that time the process of repressor was defined and not the function.
- Karen mentioned that she will use the cell type ontology term that is biologically relevant. This is true for not just cell type but for capturing enzyme substrates as well. The cell line ontology are not in the cell type ontology. There was some discussion about use of primary cell lines vs model cell lines used to study certain aspects. Karen will discuss with other MGI curators and present her findings in a future call.
- no time-did not discuss
Please read the summary. We will discuss at the next call.