Ethology module: Upper level changes by @DitchingIt

[matentzn]

Addressing #184, these changes update five of the highest levels of the NBO's process branch and add three new ones to enable future reaggregation of a number of its sub-branches. Two labels are recast, definitions follow a standard NBO style, and fixed sources are given. Annotation properties closely follow OBO Foundry recommendations.

This PR was opened by me, but performed by @DitchingIt

[DitchingIt]

Unfortunately the process removes links to child classes, unless those children have their own line in the template defining their parents.

[DitchingIt]

@matentzn Looks good. Do I need to do anything here?

[matentzn]

No the next step is to get a review from @pmidford and/or @dosumis

[pmidford]

I'm fine with the new classes inserted below NBO_0000313, despite the slight increase in complexity. I agree that splitting parental and reproductive behavior is controversial, but combining them might break existing inferences, so let's stick with that for now.

Line 1220 - an extra space at the start of the synonym string
Line 1865 - same
Maybe add 'playing behavior' (US spelling) as a synonym
Line 2033 - What's the mess at the end of the definition; guessing it's a TODO that needs attention.

I'll leave this open for @dosumis to review. Maybe @aclark-binghamton-edu should look at the fear and reproductive behavior changes as well.
Line 2033

[DitchingIt]

@pmidford Thanks:

1. The extra spaces seem to be to do with how the ROBOT system handles a carriage return after a pipe: I have removed the carriage return.
2. Playing behavior is listed as an exact synonym in line 1866: it is a recently introduced special synonym type (OMO:0003008 previous name).
3. The line 2033 mess is not in the original. When the modified ROBOT sheet (point 1 above) is assimilated, it should overwrite the garbled stuff but if not it can be done manually before merging.
4. @aclark-binghamton-edu has been monitoring the process at Ethology module - Upper levels #184 since August and I have taken into account the concrete changes suggested. Whilst welcome to comment further here, I have suggested given time constraints that further concerns be picked up as new issues in future. The Fear/Anxiety question needs untangling one day for sure (e.g. why have a superclass which simply combines the labels of its only two children? But this ROBOT sheet is not the place to make obsoletions). What I am trying to do at the moment is to improve what is currently in NBO.

Here is the modified ROBOT sheet and I suggest that merging the next PR be the next step (barring odd things like item 3 above) if there are no more comments added in the next week.

[sbello]

Taking a look at this and I have a few initial thoughts/questions, in no particular order:

1. Is it correct to use 'external threat' for all fear related behavior? Fear responses can be triggered by things that aren't actual threats but only perceived threats. Does that fall within the definition?
2. Are the terms placed under survival behavior ALWAYS related to survival? If not these should not be placed under this term.
3. I have in my notes for MP that we can only have 1 comment for a term as the presence of multiple comments is not allowed in the OBO format version of the file. Assuming this is still true the multiple comments need to be combined into a single comment.
4. You have changed many definitions. Please consider carefully if the meaning of the term has been changed in such a way that prior usages of the term would be rendered incorrect. In the ontology the relationship between the meaning and ID should always remain stable. If your change of the meaning/definition would change the meaning and usage of the term then you should obsolete the old term and create a new term. You should do this even if the new term you create has the same label as the old term. This avoids drift in meaning that would result in prior usage of the term becoming incorrect following the change in definition. I generally try to consider if the change to the term could make prior annotations incorrect before modifying the definition.

[dosumis]

I'll leave this open for @dosumis

I think @sbello is a more suitable reviewer given her experience in developing and applying MP to mouse behavioural phenotypes. Happy to help if more specific issues need discussion.

[dosumis]

@sbello 's points 1 and 2 remind me of earlier discussions stemming from ABO which attempted to separate more strictly observation from interpretation. The suggested compromise was that we allow for defined terms combining observation and interpretation - so we might have some more explicit compound terms like "fear related behaviour in response to external threat" composed of separate terms for 'fear related behaviour' and 'response to external threat'.

[DitchingIt]

Thanks @sbello for your really useful comments. I would welcome more in future as @dosumis suggests. I can only take a moment now so I will just answer your first point.

Is it correct to use 'external threat' for all fear related behavior? Fear responses can be triggered by things that aren't actual threats but only perceived threats. Does that fall within the definition?

• I would locate your example as an internal source and relate it to anxiety rather than fear.
• I tried to distill the minimum agreed elements from the debate in the source, and 'external' seemed important.
• Anyone wishing to do a better job of untangling fear/anxiety is welcome to do it; I'm just trying to align the components I've termed the Ethology Module with the new style sheet: I'm not even planning to tackle 'anxiety '

More later and again thanks.

[DitchingIt]

COMMENT 2: Are the terms placed under survival behavior ALWAYS related to survival? If not these should not be placed under this term.

I am indebted to @sbello for taking the bull by the horns. My instinct, without good ontological experience, is to aggregate inclusively rather than exclusively. I now see that I have made a mistake, and I'm glad to have it pointed out early enough to take account of it in future. As for 'survival behavior', if it can't be a simple aggregator, this isn't the place to introduce it as a new term. I agree with @dosumis about using combined terms, but I don't have the time or energy to put 'survival behavior' out on its own limb and connect it back in where appropriate, although it is something to put on a list for later. For now, I will remove it from the structure and simply connect its children directly to 'behavior process'. I won't be able to do this for a few days so don't hold your breath.

I'll look at comments 3 and 4 tomorrow.

Thanks

[DitchingIt]

COMMENT 3: I have in my notes for MP that we can only have 1 comment for a term as the presence of multiple comments is not allowed in the OBO format version of the file. Assuming this is still true the multiple comments need to be combined into a single comment.

1. @sbello This is a good opportunity to explore a structural question about NBO: NBO no longer has set ownership, let alone institutional support. As a community ontology with semi-open editing, I felt it was important to have something analogous to a wiki talk page for future editors (including myself) to refer to regarding some questions around the construction or future of certain entries. If possible it needed to not be visible via the simplest viewers (to avoid confusion for basic users), but accessible without pass codes to anyone with basic editing skills (like me) using common editing tools (like Protégé or the GitHub desktop). In constructing the ROBOT template, I was advised that using A rdfs:comment would allow visible notes and IAO:0000116 'invisible' notes.
2. I noticed that 'Prior Definition' also uses A rdfs:comment and will combine it with the 'Notes' in the ROBOT template; thanks for the observation.

[DitchingIt]

COMMENT 4

I had taken the issue of preserving meaning when changing definitions or labels to heart, but thanks for the reminder. For example the original way that NBO framed reproduction excluded brooding and parenting. I have preserved this despite many references including them. This is something to keep an eye on in future.

[DitchingIt]

I have uploaded an updated Google sheet with the following changes:

1. A rdfs:comment 'Prior Definition' is combined with 'Notes'
2. 'Survival behavior' has been removed and all its former children are now redefined and children of 'behavior process', except:
3. 'Fear-related behavior' is not yet promoted pending restructuring of the whole of NBO:0000018 'fear/anxiety related behavior'

[DitchingIt]

Thanks for the latest commit @matentzn. Unfortunately it retains the old links to NBO:0070005 'survival behavior' which was removed; specifically from NBO:0000091, NBO:0070007, NBO:0070008, NBO:0000338, NBO:0000011 and NBO:0070006.

[matentzn]

Are you certain? The pipeline

1. Takes NBO from master branch
2. Removes all classes you are editing in the spreadsheet except for their child relations
3. Converts the spreadsheet to OWL and merges into NBO

[DitchingIt]

@matentzn Yesterday Protege opened for me with 'survival behavior' gone but all the links to it still there. Today I restarted it and behold, the links are gone as well!

All good from my end. @pmidford @sbello Any further reviews?

[pmidford]

I'm ready to post a review and let this out, but what are these failing qc checks? Are they cause for concern?

[DitchingIt]

@matentzn @pmidford @sbello I have uploaded a slightly edited ROBOT template (to the same location as before) with a number of redundant comments removed, and added a subset to more easily search for the parts of NBO that this effort is editing or adding.

I don't know if the qc failure is related to the construction of this pipeline? It was our first time and Nico has included 'never merge' in the title???

[DitchingIt]

@matentzn When you have time, the pull request is titled 'DRAFT (never merge)...' and the QC checks have from the beginning said 'Error: Process completed with exit code 2'. @pmidford and I are wary of completing our reviews but would like to merge the changes. Your thoughts would be gratefully received.

[matentzn]

The checks have passed now (I made a change to you template in the "subset" column) but this PR needs a positive approval from @pmidford to be merged in!

[pmidford]

OK, absent any further comments from @sbello and since other issues have been addressed, I approve this merge.

[matentzn]

@DitchingIt which issues does this address, can you provide a three sentence high level summary of the changes, and tell me a more appropriate title for the pull request?

[DitchingIt]

@DitchingIt which issues does this address, can you provide a three sentence high level summary of the changes, and tell me a more appropriate title for the pull request?

Ethology module: Upper level changes

Addressing #184, these changes update five of the highest levels of the NBO's process branch and add three new ones to enable future reaggregation of a number of its sub-branches. Two labels are recast, definitions follow a standard NBO style, and fixed sources are given. Annotation properties closely follow OBO Foundry recommendations.