RG NTR: transcriptionally active chromatin binding

[gocentral]

Another request from the UK curator pre-jamboree meeting...

In PMID:17965872 Drosophila Nipped-B is shown by ChIP to bind preferentially to transcriptionally active regions of the genome. This contrasts with the orthologous S. cerevisiae protein Scc2 which binds between genes (in transcriptionallly inactive regions).

Could we have new terms along the lines of: transcriptionally active chromatin binding Def: Interacting selectively and non-covalently with chromatin associated with regions of the genome that are being actively transcribed

Does it make sense to also have: transcriptionally inactive chromatin binding Def: Interacting selectively and non-covalently with chromatin associated with regions of the genome that are not being actively transcribed

I think these should be children of chromatin binding rather than chromatin DNA binding because the assay is does not distinguish what component of chromatin is being bound.

Thanks Susan

Reported by: stweedie

Original Ticket: geneontology/ontology-requests/7068

[gocentral]

• assigned_to: nobody --> rfoulger

Original comment by: mah11

[gocentral]

The problem with 'complex' binding terms can be deciding if something is binding a complex or is part of the complex. If a protein is part of the complex I always assume it is not then (usually) binding to it. Note that the definition for chromatin needs to also include RNA.

My feeling about chromatin is that it does not really have an 'end'. It is considered as an organelle without a membrane and my personal interpretation, having spent several years looking down a microscope, is that a protein is either part of the chromatin or not part of the chromatin and that a protein which is considered to bind to chromatin is actually part of chromatin. However, looking at the synonyms for this narrow nuclear membrane vesicle binding to chromatin I do agree that there could be examples where proteins from one 'organelle' enable a connection between it and chromatin. But I don't think the definition describes this limitation.

The majority of human proteins annotated as chromatin binding are IEAs with only 58 manual annotations to this term. But I think the IEA mappings and most of the manual annotations could be changed to chromatin. Note that 64% of human annotations to chromatin binding are also annotated to chromatin.

So I am wondering perhaps the chromatin binding terms requested should actually have been component equivalents rather than functions. eg:

GO:0000785 chromatin >transcriptionally active chromatin >transcriptionally inactive chromatin

Original comment by: RLovering

[gocentral]

PS: I don't think chromatin DNA binding should be a child of chromatin binding, because a protein bound to chromatin DNA is part of chromatin not bound to chromatin.

Original comment by: RLovering

[gocentral]

Hi Becky,

It would be a good idea to consult SGD curators, especially Karen and Eurie, about this one.

cheers, m

Original comment by: mah11

[gocentral]

Ruth, when you say 'part of chromatin', do you mean integral to its formation and maintenance of its structure? Because I would say there is a difference between proteins (such as histones) that are necessary to form chromatin and proteins that associate with chromatin. Maybe we are saying the same thing but expressing it differently.

For the component term suggestions ('transcriptionally active chromatin' and 'transcriptionally inactive chromatin'), how are the ones you are suggesting different from these existing terms:

nuclear euchromatin (GO:0005719) heterochromatin (GO:0000792)

And as a corollary then, shouldn't the MF terms be more along the lines of 'euchromatin binding' and 'heterochromatin binding'?

Original comment by: eurie

[gocentral]

Hi,

Here are my thoughts.

A. Questions from Becky's email:

1. Where does 'chromatin binding' end and 'chromatin' begin?

I do see how this can be a tricky question, but I think researchers working in this area make a distinction between "chromatin proteins" that play a structural role in wrapping the DNA up and other proteins that bind to either the DNA or to these structural chromatin proteins to modulate chromatin structure. For example, I am currently reading about the HMGB domain proteins NHP6A and NHP6B. David Stillman has just written a great review called "Nhp6: A small but powerful effector of chromatin structure in Saccharomyces cerevisiae" on these and he talks about them as a DNA binding protein that modulates chromatin structure, and not as a chromatin protein, even though they are abundantly present with "one Nhp6A molecule for every one to two nucleosomes". So, I would not consider the Nhp6 proteins to be chromatin proteins based on what I have read so far.

2. Are the existing 'chromatin binding' terms valid?

I think that the term "chromatin binding" is valid in the sense that this activity is thought to exist. However, I don't think you can annotate to "chromatin binding", or any more specific child term, on the basis of an in vivo ChIP experiment. If the ChIP was done from cells, you can say that protein x of interest localized a particular type of chromatin, but you can't say that protein x binds chromatin because you have no idea if it is direct or not. In the rare case of an in vitro set up with chromatin reconstituted from histones and DNA, you might be able to use a ChIP experiment to annotate to "chromatin binding". There are probably some other in vitro assays looking at binding to reconstituted chromatin that would also allow annotation to "chromatin binding".

3. If I create coponent terms for: GO:0000785 chromatin >transcriptionally active chromatin >transcriptionally inactive chromatin do you think we need 'transcriptionally inactive/active chromatin binding' terms too?

First of all, I agree with Eurie that I don't see how these terms are different from the existing "heterochromatin" and "euchromatin" terms, so we should use the existing terms and add new synonyms if needed.

I think the component terms for the types of chromatin and the function terms for binding to types of chromatin are complementary, but not completely overlapping or redundant. However, as I stated above, if the experiment is an in vivo ChIP (which the cited Drosophila example seems to be), I think that you can only annotate to the component terms, e.g. "euchromatin" perhaps using the "colocalizes" qualifier, but NOT to the corresponding function term, e.g. "euchromatin binding", because you don't know what the Drosophila Nipped-B protein was binding to in order to localize to transcriptionally active euchromatin.

B. Other comments from reading the thread

4. If the function terms:
   • euchromatin binding (transcriptionally active chromatin binding)
   • heterochromatin binding (transcriptionally inactive chromatin binding)

are created, then I agree with Susan that they should be a child of "chromatin binding" because as defined, these term definitions merely specify binding to "chromatin", not to either the DNA or the protein components of chromatin.

5. I disagree that "a protein bound to chromatin DNA is part of chromatin not bound to chromatin". As I stated in #1, I'm currently reading about the NHP6 proteins in cerevisiae and they are shown to bind DNA, and are thought to perform this function in the context of chromatin, but are not themselves considered to be chromatin proteins.

6. I have never heard of RNA being a component of chromatin. Could you provide explanation and perhaps also a reference for including RNA in the definition of chromatin?

-Karen

Original comment by: krchristie

[gocentral]

Thanks very much for your comments Eurie and Karen!

Suggested actions so far:

ADD IN SYNONYMS TO EXISTING COMPONENT TERMS:

nuclear euchromatin ; GO:0005719 NEW narrower synonym: transcriptionally active chromatin (according to our current definition, nuclear chromatin covers both transcriptionally active and inactive forms)

heterochromatin ; GO:0000792 narrower synonym: transcriptionally inactive chromatin

CREATE NEW FUNCTION TERMS:

nuclear euchromatin binding ; GO:NEW narrower synonym: transcriptionally active chromatin binding Interacting selectively and non-covalently with the dispersed, less dense form of chromatin in the interphase nucleus. source: ISBN:0198506732

heterochromatin binding ; GO:NEW narrower synonym: transcriptionally inactive chromatin binding Interacting selectively and non-covalently with the compact and highly condensed form of chromatin.

chromatin binding ; GO:0003682 --%nuclear euchromatin binding ; GO:NEW --%heterochromatin binding ; GO:NEW

Does this work for everyone?

I'll wait to hear about the RNA bit before updating the chromatin defs.

Thanks, Becky

Original comment by: rebeccafoulger

[gocentral]

My concern is that euchromatin and heterochromatin are defined based on a cytogenetic assay and although there is some correlation between euchromatin and activity and heterochromatin and inactivity this is far from exact - there are transcriptionally active genes in heterochromatin and our definition of euchromatin acknowledges the fact that it isn't all active. So these terms don't really reflect what is being assayed and I don't think I would be keen to use them for this annotation even if you add the narrow synonyms. I suppose I feel that the heterochromatin/euchromatin distinction based on affinity to dyes may be becoming a bit irrelevant to the types of experiments that are done now and that a distinction based on active v inactive is more biologically relevant./interesting. But I can also see that it isn't good expand the ontology unnecessarily so I will withdraw this request and annotate to colocalizes_with chromatin if you don't agree with this. I don't mind if the new binding terms are added or not since I take Karen's point that the ChiP expt isn't vailid for binding anyway.

Thanks for the input. Susan

Original comment by: stweedie

[gocentral]

Hi Karen and Eurie,

Thank you for providing such clear explanations about this. I am afraid my view of chromatin is from 20 years ago as a microscopist which has coloured my view of the scientific advances that have been achieved since then.

In answer to Karen's Point 1, I completely agree that proteins can be defined as binding chromatin structural proteins and that ChIP experiments do not provide this information. But the current GO definition of chromatin: The ordered and organized complex of DNA and protein that forms the chromosome. does not limit chromatin to just DNA and the structural proteins that bind it. Indeed, several online definitions of chromatin include RNA, and are as broad as the GO term definition and are along the lines of my microscope view of chromatin: all of the molecules which make up the chromosome.

I am happy for the definition to be modified to follow Karen's suggestion that chromatin proteins 'play a structural role in wrapping the DNA up and other proteins that bind to either the DNA or to these structural chromatin proteins to modulate chromatin structure'. However, I would have thought that this definition would include 'Nhp6: A small but powerful effector of chromatin structure'. The Nhp6 review does not use the term chromatin protein at all, but this doesn't mean he doesn't consider any of the proteins he discusses to be part of chromatin. In addition Karen's definition currently could include transcription factors which I don't think is the intention, so maybe we could decide what proteins are going to be included in 'chromatin' and agree on a definition which makes the limitations very clear.

Personally I would prefer to see this definition include chromatin structure modifiers and I would also question whether the term defined by 'DNA and the structural proteins which bind directly to the DNA' would be better described as 'chromatin fibre'.

In answer to Karen's Point 2, depending on the definition of chromatin, I can see that ChIP expts cannot distinguish between chromatin binding and being part of chromatin.

In answer to Karen's Point 3, In response to the comment about heterochromatin=transcriptionally inactive chromatin, I am relieved to be able to find that the genes transcribed in heterochromatin are still considered as such, please see PMID: 17080025 for comments on the existence of transcription in heterochromatic regions. 'heterochromatic genes have been found in Saccharomyces cerevisiae, Schizosaccharomyces pombe, Oryza sativa, Drosophila and Arabidopsis thaliana, as well as in humans. The presence of expressed genes in heterochromatin seems paradoxical because they appear to function in an environment that has been considered incompatible with gene expression'. Therefore 'transcriptionally active chromatin' and 'inactive' is a more accurate statement than euchromatin or heterochromatin.

In answer to Karen's Point 4, yes I agree.

In answer to Karen's Point 5, I am interested to see how chromatin will be defined to meet the criteria listed in this comment.

In answer to Karen's point 6, http://wordnetweb.princeton.edu/perl/webwn?s=chromatin S: (n) chromatin, chromatin granule (the readily stainable substance of a cell nucleus consisting of DNA and RNA and various proteins; during mitotic division it condenses into chromosomes),; http://en.wiktionary.org/wiki/chromatin, (biology) A complex of DNA, RNA and proteins within the cell nucleus out of which chromosomes condense during cell division;

Ruth

Original comment by: RLovering

[gocentral]

If the newer phrases "transcriptionally active chromatin" and "transcriptionally inactive chromatin" older are not synonymous with the older "euchromatin" and "heterochromatin" terms then it may be appropriate to create new terms for these. I thought Eurie brought up a good question and it was consistent with my knowledge of those terms, but it sounds like there is not a complete correspondance between the two sets of phrases. I'm not against creating new terms when there's a need for them, but I agree with Eurie that it's worth looking at existing terms and making sure that the needed entity doesn't already exist, though perhaps with a different name.

-Karen

Original comment by: krchristie

[gocentral]

If the newer phrases "transcriptionally active chromatin" and "transcriptionally inactive chromatin" older are not synonymous with the older "euchromatin" and "heterochromatin" terms then it may be appropriate to create new terms for these. I thought Eurie brought up a good question and it was consistent with my knowledge of those terms, but it sounds like there is not a complete correspondance between the two sets of phrases. I'm not against creating new terms when there's a need for them, but I agree with Eurie that it's worth looking at existing terms and making sure that the needed entity doesn't already exist, though perhaps with a different name.

-Karen

Original comment by: krchristie

[gocentral]

If the newer phrases "transcriptionally active chromatin" and "transcriptionally inactive chromatin" older are not synonymous with the older "euchromatin" and "heterochromatin" terms then it may be appropriate to create new terms for these. I thought Eurie brought up a good question and it was consistent with my knowledge of those terms, but it sounds like there is not a complete correspondance between the two sets of phrases. I'm not against creating new terms when there's a need for them, but I agree with Eurie that it's worth looking at existing terms and making sure that the needed entity doesn't already exist, though perhaps with a different name.

-Karen

Original comment by: krchristie

[gocentral]

If the newer phrases "transcriptionally active chromatin" and "transcriptionally inactive chromatin" older are not synonymous with the older "euchromatin" and "heterochromatin" terms then it may be appropriate to create new terms for these. I thought Eurie brought up a good question and it was consistent with my knowledge of those terms, but it sounds like there is not a complete correspondance between the two sets of phrases. I'm not against creating new terms when there's a need for them, but I agree with Eurie that it's worth looking at existing terms and making sure that the needed entity doesn't already exist, though perhaps with a different name.

-Karen

Original comment by: krchristie

[gocentral]

If the newer phrases "transcriptionally active chromatin" and "transcriptionally inactive chromatin" older are not synonymous with the older "euchromatin" and "heterochromatin" terms then it may be appropriate to create new terms for these. I thought Eurie brought up a good question and it was consistent with my knowledge of those terms, but it sounds like there is not a complete correspondance between the two sets of phrases. I'm not against creating new terms when there's a need for them, but I agree with Eurie that it's worth looking at existing terms and making sure that the needed entity doesn't already exist, though perhaps with a different name.

-Karen

Original comment by: krchristie

[gocentral]

"euchromatin" and "heterochromatin" both have "transcriptionally active chromatin" and "transcriptionally inactive chromatin" areas, therefore these would be great new terms to create, I can't see how they would relate to "euchromatin" and "heterochromatin" unless they were "transcriptionally active euchromatin" etc. however some assays may not be able to distinguish whether the "transcriptionally active chromatin" was in a euchromatic or heterochromatic region, so the parent terms "transcriptionally active chromatin" and "transcriptionally inactive chromatin" would also be needed and the more specific child terms maybe unnecessary.

Ruth

Original comment by: RLovering

[gocentral]

So.... if everyone is agreed, I'll create new COMPONENT terms:

chromatin ; GO:0000785 --%transcriptionally active chromatin ; GO:NEW --%transcriptionally silent chromatin ; GO:NEW

transcriptionally active chromatin ; GO:NEW The ordered and organized complex of DNA and protein that forms regions of the chromosome that are being actively transcribed.

transcriptionally silent chromatin ; GO:NEW synonym=transcriptionally inactive chromatin The ordered and organized complex of DNA and protein that forms regions of the chromosome that are not being actively transcribed.

I won't create any new chromatin-binding FUNCTION terms, but will leave the existing ones in the ontology for when an assay truly shows this.

If we find we need them, I can create terms later for: transcriptionally active heterochromatin transcriptionally active euchromatin transcriptionally silent heterochromatin transcriptionally silent euchromatin

I've created a new SF item about tweaking the definition of chromatin because it affects a number of entries (SF: 2958739 ).

https://sourceforge.net/tracker/index.php?func=detail&aid=2958739&group\_id=36855&atid=440764

thanks, Becky

Original comment by: rebeccafoulger

[gocentral]

I'm happy with that. Thanks everyone. Susan

Original comment by: stweedie

[gocentral]

Sounds good to me.

-Karen

Original comment by: krchristie

[gocentral]

Added 2 new component terms:

transcriptionally active chromatin ; GO:0035327 transcriptionally silent chromatin ; GO:0035328

Thanks, Becky

Original comment by: rebeccafoulger

[gocentral]

• status: open --> closed-accepted

Original comment by: rebeccafoulger