silencing terms will need new logical definitions

[ValWood]

All of the GO "chromatin silencing' terms will be merged into correspondeing "heterochromatin assembly" terms.

Therefore, andy logical desf which use them will need to be defined differnetly (i'm not sure how but mostly probably "negattive regulation of gene expression" and "heterochromatin"

The complete list is here: https://docs.google.com/spreadsheets/d/1a-iWPesVFSTpSgRIx9p8D5D5BLNjzMGI8wHVmqvWu5Q/edit#gid=0

[mah11]

ontology checks done upon release would flag anything like this

[ValWood]

Ah OK, I wondered if it did. Do merges generally create a lot of rejigging?

[mah11]

The linked spreadsheet doesn't have anything about merging "chromatin silencing' terms into "heterochromatin assembly" terms. Is it the correct one? (Whether it is or not, it's not easy to follow what the actual proposed changes are in it, but if it's the wrong one that's not a problem for FYPO.)

[mah11]

Do merges generally create a lot of rejigging?

Fortunately, GO merges that affect FYPO have been rare enough that I don't have a lot of experience to draw on. My sense is that it depends on the terms involved as well as how many FYPO terms are affected.

Aside: Actually doing term merges is about to become a much bigger PITA, when we switch to using Protege.

[mah11]

Or has that spreadsheet got some weird filtering imposed? I just noticed that the line number column jumps from 2 to 28, then 35, 41, 49 ... best get it fixed if you want anyone to be able to use it.

[mah11]

In any case, it'll be far easier to find the affected terms, and see how to update FYPO, after the GO work is done.

[ValWood]

yeah I think that spreadsheet has some weird filtering. I think it's only for Pascale to work from. I thought I would give you a heads up, but it sounds as though it won't make so much difference

[mah11]

There's now a relevant GO ticket - https://github.com/geneontology/go-ontology/issues/19112

[mah11]

@ValWood Question: Do you want to change the phenotype terms to "heterochromatin assembly" or do you want to leave the FYPO term names unchanged, and then use "chromatin silencing" phenotypes to infer a role in or effect on GO BP heterochromatin assembly?

If the FYPO terms should still use "chromatin silencing" then maybe they can't use GO BP logical defs ... I'd have to have a think.

[mah11]

affected GO:0006343 FYPO:0007223 GO:0030702 FYPO:0000147 GO:0030702 FYPO:0002360 GO:0030702 FYPO:0002834 GO:0030702 FYPO:0003411 GO:0030702 FYPO:0004541 GO:0030702 FYPO:0005071 GO:0030702 FYPO:0006995 GO:0030466 FYPO:0000156 GO:0030466 FYPO:0002336 GO:0030466 FYPO:0002827 GO:0030466 FYPO:0004376

[ValWood]

The phenotypes should still. be "silencing". So the logical defs will now be incorrect. Go won't have the concept of 'silencing' (phenotype) I hope it isn't too many :(

[mah11]

OK, that does at least make it an easy fix - ditch the LDs. It's only the dozen terms in my previous comment. I'll implement when the GO work is actually done.

[ValWood]

I thought there were bzillions of silencing terms. I think there are more GO term changes on the way...

[mah11]

/me goes and counts ...

Ah, 46 terms with "silencing" in the name. For the list above I only looked at the IDs Pascale put in the GO ticket (19112), so presumably more will be on the way. Not much more work to deal with the lot all at once (I'll still wait for the first wave of changes to hit GO, though).

There are a handful of FYPO terms that say "gene silencing" and refer to GO:0016458. Do you know whether anything will happen to that one?

[ValWood]

I think GO:0016458 is safe as it is post-transcriptional. Well at least right now. There may be rejigging but I expect it will be more minor.

[colinlog]

Would something along the lines of an arbitary split whereby;

'Silencing can only be a BP and it has amongst its children 'Heterochromatin assembly / maintenance' that are MF, or at the very least BPs with proteins that have known relevant MF such as histone chaperone, histone modification, modified histone biding, etc...'

work-out in the logical definitions, phenotypes and other places where the anticipated simplification of the heterochromatin branches of GO is anticipated to perhaps be problematic?

[mah11]

@colinlog If I understand correctly, Val's point is that silencing is a phenotype, and therefore belongs in ontologies like FYPO rather than in GO. It's not really a matter of splitting between MF and BP in GO.

[colinlog]

'Silencing defect' or probably 'suppression/enhancement of position effect variegation caused by translocations close to centromeres or telomeres' is then a set of phenotypes (the famous polycomb mutants and all the beautiful biology they underlie), as would be some homeotic transformation phenotypes and some diseases such as cancer progression driven by aberrant chromatin silencing processes that cause tumour suppressor epigenetic inactivation. But the process silencing is synonymous to a number of GO processes amongst which heterochromatin assembly and maintenance, no?

[ValWood]

@colinlog But we agreed on the last that transcription call that BP 'chromatin silencing' as currently in GO? i.e./ Repression of transcription by altering the structure of chromatin, e.g. by conversion of large regions of DNA into an inaccessible state often called heterochromatin. is equivalent to BP 'heterochromatin assembly'

This is how it has been used so I believe the merges are safe? If not you should alert @pgaudet now as the merges are in progress (today hopefully)

NOTE that this is the FYPO tracker. You have hopped into a new species ;) We are describing fission yeast phenotypes here (PomBase makes detailed phenotype annotation to attach phenotypes with specific genotypes and conditions)

This effort is orthogonal to GO. But we use GO normal processes in FYPO logical definitions. In FYPO we will still have 'chromatin silencing' phenotypes to describe assays where a reporter gene is inserted into constitutive heterochromatin and derepression is observed. But for GO we do not need to worry about this ticket. ;)

[colinlog]

@Pascale Gaudet no problem with our decision. I just got triggered by the remark thatthe word silencing would fully disappear. This sounds crazy. but it is a matter of synonyms, ot of refactoring, which is very sensible and should not be stopped!!!

On Tue, Mar 24, 2020 at 12:21 PM Val Wood notifications@github.com wrote:

@colinlog https://github.com/colinlog But we agreed on the last that transcription call that BP 'chromatin silencing' as currently in GO? i.e./ Repression of transcription by altering the structure of chromatin, e.g. by conversion of large regions of DNA into an inaccessible state often called heterochromatin. is equivalent to BP 'heterochromatin assembly'

This is how it has been used so I believe the merges are safe? If not you should alert @pgaudet https://github.com/pgaudet now as the merges are in progress (today hopefully)

NOTE that this is the FYPO tracker. You have hopped into a new species ;) We are describing fission yeast phenotypes here (PomBase makes detailed phenotype annotation to attach phenotypes with specific genotypes and conditions)

This effort is orthogonal to GO. But we use GO normal processes in FYPO logical definitions. In FYPO we will still have 'chromatin silencing' phenotypes to describe assays where a reporter gene is inserted into constitutive heterochromatin and derepression is observed. But for GO we do not need to worry about this ticket. ;)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/pombase/fypo/issues/3635#issuecomment-603181847, or unsubscribe https://github.com/notifications/unsubscribe-auth/ALZVLKACNGWRFBU4MXNVZNLRJCJSTANCNFSM4LHGGKGQ .

[pgaudet]

@colinlog I think 'silencing' should remain as a synonym. This also avoids creating terms that we have previously obsoleted.

[mah11]

I've done the bulk of the work, so I'll close this and open one new ticket for one outstanding bit (which is waiting for work on the relevant GO term). New tickets for anything else you spot too, please.

edit file: 1406da4f08419332f5edf314d3b32c69f3d7202f release: e78a27623827f37533246f6b61131aff7b85e8ed

[ValWood]

Great. I have been avoiding sending out CC sessions/reminder for chromatin related stuff because I think the big GO issues are one (of a few) reasons that this group does not curate much. I hope these changes (mainly to GO ) will help to make the curation much easier/clearer.