heterochromatin domain assembly, vs heterochromatin assembly

[gocentral]

I hadn't seen heterochromatin domain assembly before

The aggregation, arrangement and bonding together of a set of components to form a heterochromatin domain.

how does this differ from the parent heterochromatin assembly (i.e isn't couldn't all heterochromatin be referred to as a heterochromatin domain?)

Reported by: ValWood

Original Ticket: geneontology/ontology-requests/11065

[gocentral]

A "heterochromatin domain", aka "HOOD", is actually a specific type of heterochromatic region, despite its unspecific-sounding name. Details in the referenced CC term (GO:1990343) and the refs cited for its def.

Might be an idea to copy some of the CC def as a gloss for the BP term.

Original comment by: mah11

[gocentral]

Hmm, maybe HOOD should be in the term name as "heterochromatin domain " is used more generally. Also, other chromatin requires RNAi and histone H3 methylated on lysine 9 (H3K9me) (centromeric heterochromtin for instance), so is the only difference the region where this is formed type of heterochromtin?

Although this is odd, because it was previously reported that the mechanism operating at these regions (stress genes and LTRS) was co-transcriptional (although dependent on RNAi) http://curation.pombase.org/pombe/curs/e64102699ced97dc/ro/ which I think would not require heterochromatin formation? Did they detect any heterochromtin (I notice this is an IGI annotation), and I can't find anything annotated to heterochromatin domain using the query builder.

v

Original comment by: ValWood

[gocentral]

• labels: --> Other term-related request
• Description has changed:

Diff:

--- old
+++ new
@@ -1,4 +1,3 @@
-

 I hadn't seen 
 heterochromatin domain assembly before

• assigned_to: Paola Roncaglia
• Group: None --> PomBase

Original comment by: paolaroncaglia

[gocentral]

I have no objection to changing the term name.

They say they saw heterochromatin, and I don't know enough to have grounds to doubt them.

If I understand correctly, the distinguishing features of HOODs are (a) dynamic formation (as opposed to centromeric, telomeric, etc.) and (b) requiring RNAi (there's another type of dynamically formed heterochromatin, "heterochromatin islands", that allegedly doesn't require RNAi). If the H3K9me bit is general it can be removed from the defs.

Original comment by: mah11

[gocentral]

I'll have a look at both papers, and I might ask Mark Buehler how much these two processes have in common.

These sound like they could be the same process, but the annotations from the 2 papers currently put them in totally different parts of the graph http://www.ebi.ac.uk/QuickGO/GMultiTerm#a=64%2408CQ7GZB&tab=chart&c=

This term http://www.ebi.ac.uk/QuickGO/GTerm?id=GO:0070924 would place under both 'chromatin aseembly by RNAi' and 'heterochromatin formation'

at present I'm still not clear what the differentia is for 'heterochromatin domain assembly' from normal heterochromtin assembly, apart from the location where it occurs.....

Original comment by: ValWood

[gocentral]

I seem to recall there being a distinction based on when hc formation occurs, as well as where -- HOOD assembly is dynamic and associated with particular events or conditions (tho I'd have to look up the specifics). Same might go for hc islands.

Original comment by: mah11

[gocentral]

Discussed at SF jamboree. Val says they need to do more reading and decide course of action. Set to pending while we wait for their resolution.

Original comment by: paolaroncaglia

[gocentral]

Based on this (which cites some of the authors of the paper which generated the "heterochromatin domain" term with a different usage),

PMID:20211136 In fission yeast, domains of heterochromatin are found at telomeres, the silent mating-type locus, and on pericentromeric repeats (reviewed in Bu ̈hler and Moazed, 2007; Grewal and Jia, 2007).

possibly the heterochromatin domain term is OK, but if it exists it should also be a parent of all of the other heterochromatin regions which are RNAi dependent, need to check if there is any other criteria here.

For instance are they located at protein coding regions?

(this is a note for me....I still haven't got around to comparing the 2 papers yet)

Original comment by: ValWood

[gocentral]

note to self: the difference between the Grewal and Buhler papers could be establishment/maintenance differences...

In fission yeast, domains of heterochromatin are found at telo- meres, the silent mating-type locus, and on pericentromeric repeats (reviewed in Bu ̈hler and Moazed, 2007; Grewal and Jia, 2007). This heterochromatin is characterized by histone H3 lysine 9 methylation (H3K9me), mediated by the sole H3K9 methyltransferase, Clr4. H3K9me creates binding sites for the chromodomain proteins Swi6, Chp1, Chp2, and Clr4 (Bannister et al., 2001; Sadaie et al., 2004; Zhang et al., 2008). Several histone deacetylases (HDACs), Clr3, Clr6, and Sir2, are also required to facilitate H3K9 methylation (Grewal et al., 1998; Nakayama et al., 2001; Shankaranarayana et al., 2003). At centromeres, RNAi promotes H3K9 methylation on centromeric outer repeat sequences (Motamedi et al., 2004; Verdel et al., 2004; Volpe et al., 2002). It is possible to distinguish between establishment of H3K9me, which is fully dependent on RNAi, and its subsequent maintenance, which is only partially RNAi dependent (Sadaie et al., 2004). RNAi also targets mating-type locus and telomeric elements with homology to centromere outer repeats. However, here, alternative pathways act redun- dantly with RNAi to recruit chromatin modifiers, so that RNAi is required for establishment but not for maintenance of H3K9me at these loci (Hansen et al., 2006; Jia et al., 2004; Kanoh et al., 2005; Kim et al., 2004).

(i.e I'm not sure if the heterochromatin should be distinguished at the CC level by HOW it is formed as there are redundant pathways for formation of the same stucture...this should be captured by process terms)

Original comment by: ValWood

[pgaudet]

Merge ?

[ValWood]

I'm not sure.

Heterochromatin island is defined

Definition | A region of heterochromatin that is formed dynamically in response to environmental signals by a process that does not require RNAi, and is enriched in histone H3 methylated on lysine 9 (H3K9me).

All heterochromatin is H3K9 dependent. Not all heterochromatin is formation is RNAi independent.

So, based on the def of "Heterochromatin island" heterochromatin island assembly does not seem different from "heterochromatin assembly".

However, I suspect there is something more about "Heterochromatin island" If it is required we need a better differentia for heterochromatin island.

@mah11 curated these - is there anything else that will differentiate?

[ValWood]

While we decide if we need to do anything with

heterochromatin island assembly (GO:1902794) while this term is under review can we place it under heterochromatin assembly. It is not there currently.

[mah11]

I don't know much beyond I put in earlier comments on this ticket (here, here, and here.

I used PMIDs 23151475 and 24210919 for the terms in question. I'm neutral on whether to keep distinct CC terms (without them, ~10 FYPO terms lose logical defs; no big deal), but the assembly processes do seem to differ -- the PMID:24210919 intro says

"S. pombe contains three distinct types of heterochromatin, which differ in their dependency on trans-acting factors."

and then goes on to provide some details.

[ValWood]

@pgaudet can we keep for now.? Once we have purged the silencing stuff, I will ask the heterochromatin community to review so that we have everything classified correctly by the different processes (mechanisms). I think this (mechanism) is really more important than the location, although sometimes we are using one as a proxy for the other.

Possibly these domains are more related to constitutive heterochromatin assembly. Once the other changes are in place it will be easier to decide about this. Maybe we will rename the term, I m not sure.

This is how our annotation looks now ( I fixed a few more today so it will be better tomorrow). https://www.pombase.org/term/GO:0031507

This also needs the remaining subtelomeric heterochromatin terms and the heterochromatin island term under this branch. Once this is done I can ask for community input.

I can also look how the gene products in the individual processes intersect so I can look for any inconsistencies.

At that point I can open new tickets. So I would move this under heterochromatin assembly and close this ticket .

[ValWood]

Oh we still need to deal with the fact that "heterochromatin assembly" and "heterochromaitn island assembly" are defined identically

The problem is that "heterochromatin island" does not have sufficent differentia from heterochromatin. Heterochromatin also fits this definition. @mah any suggestion how to differentiate?

[mah11]

@mah any suggestion how to differentiate?

I just mentioned in my previous comment that relevant info can be found in PMID:24210919 and the papers cited in its introduction (and maybe there'll be more in any papers that cite it in turn; I haven't looked into that).

[ValWood]

so PMID:24210919 says

The first type corresponds to major H3K9me peaks at centromeres, telomeres, and the mating-type locus. RNAi proteins Argonaute (Ago1) and Dicer (Dcr1) and RNA-dependent RNA polymerase (Rdp1) target transcripts produced by dg/dh repeats in these regions to generate siRNAs that facilitate loading of the Clr4/Suv39h methyltransferase (Reyes-Turcu and Grewal, 2012). The second type includes small blocks of heterochromatin islands, which include meiotic genes and other loci (Cam et al., 2005, Zofall et al., 2012). Islands are akin to facultative heterochromatin and are dynamically regulated in response to environmental signals (Zofall et al., 2012). The assembly of heterochromatin islands at meiotic genes requires Red1 and Rrp6, but not RNAi (Hiriart et al., 2012, Tashiro et al., 2013, Zofall et al., 2012). The third type of heterochromatin consists of domains called HOODs (heterochromatin domains), which are detected under specific growth conditions and are dynamically regulated (Yamanaka et al., 2013). HOODs preferentially assemble at sexual differentiation genes and retrotransposons and require RNAi as well as elimination factors (Yamanaka et al., 2013). Red1 and its partners Pla1 and Pab2 are cofactors for both RNAi and the exosome, which act in parallel to degrade transcripts from loci within these domains (Yamanaka et al., 2013).

Which implies subtelomeric heterochromatin is also RNAi dependent.

islands form at meiotic genes in mitotic cells (facultative)

Suggest This def A region of facultative RNAi-independent heterochromatin that is formed dynamically at specific loci in response to environmental signals.

Then the "heterochromatin island assembly" term will also make sense.

[pgaudet]

@ValWood also asked to move subtelomeric heterochromatin assembly and heterochromatun island assembly under heterochromatin assembly

[pgaudet]

Fixed defintion of 'heterochromatin island' as follows: 'A region of facultative heterochromatin formed dynamically at specific loci in response to environmental signals, independently of RNAi.'

[ValWood]

also heterochromatin domain assembly under heterochromatin assembly!

[pgaudet]

heterochromatin domain assembly under heterochromatin assembly!

Yes, done.