GO:0071625 vocalization behavior, additional child term?

[gocentral]

Hi

looking at the ontology structure and definitions of these two terms: GO:0042297 vocal learning
GO:0071625 vocalization behavior

I think that there should be the following relationship added, plus a new term added:

GO:0071625 vocalization behavior

GO:0042297 vocal learning
new GO term innate vocalization behavior Definition: The innate behavior of an organism to produce sounds by a mechanism involving its respiratory system

I don't have a usage case for the innate vocalization behavior term, however the definition for vocal learning: 'A behavioral process whose outcome is a relatively long-lasting adaptive behavioral change whereby an organism modifies innate vocalizations to imitate or create new sounds', I think implies that innate vocalization behavior exists and it would make it easier to understand what is meant by GO:0071625 vocalization behavior, definition: The behavior in which an organism produces sounds by a mechanism involving its respiratory system, if the innate term was present as a sibling to vocal learning.

Thanks

Ruth

Reported by: RLovering

Original Ticket: geneontology/ontology-requests/10565

[gocentral]

• assigned_to: David Osumi-Sutherland

Original comment by: jl242

[gocentral]

• status: open --> closed-fixed

Original comment by: dosumis

[gocentral]

Added terms for innate and learned vocalisation behavior.

Original comment by: dosumis

[gocentral]

Hi David

I think you have misinterpreted my request.

I think the new GO term GO:0098583 learned vocalization behavior is the same as GO:0042297 vocal learning, I apologise for not making it clearer. I was looking for

GO:0071625 vocalization behavior is a child >GO:0042297 vocal learning

Perhaps this is a problem with this new SF look that > is changed to a straight line, and then as I hadn't added is_a it was not clear what I was after?

So if you agree that GO:0098583 learned vocalization behavior is the same as GO:0042297 vocal learning please delete the GO:0098583 learned vocalization behavior.

If you feel these are different then please create some sort of link between the two.

Although I guess for deaf people the parent term for GO:0042297 vocal learning: GO:0031223 auditory behavior is not correct. and perhaps a better parent term woudl be GO:0007638 mechanosensory behavior Behavior that is dependent upon the sensation of a mechanical stimulus.

Thanks

Ruth

Original comment by: RLovering

[gocentral]

Hi Ruth,

The two terms seem distinct to me - the process of learning and the process of performing something learned are different. I'm guessing you'd pay to hear a singer perform a piece of music, but not to hear them learning it. There is a causal relationship between the two, but one is not a subclass of the other. Do you have any suggestions for tweaks to the definition or comments that would make this clearer?

I agree that vocal learning is not necessarily an auditory behaviour (defined as a behavioural response to sound) - although I suspect it almost always it. It is certainly possible for humans to learn to control vocalisations from sources other hearing (e.g. I might learn a piece of music by reading it). I've removed this classification.

(BTW - for formatting on the new SourceForge - please see: https://sourceforge.net/p/geneontology/ontology-requests/markdown_syntax - link should also be a button on your left on this page. Quite straightforward. Using special characters without being aware of this can lead to confusion.)

Cheers, David

Original comment by: dosumis

[gocentral]

One more thought - as cases of vocal learning as currently defined are probably rare, perhaps we should narrow the definition to specify that the learning itself must be based on some auditory stimulus ?

Original comment by: dosumis

[gocentral]

Hi David before I think about this, an urgent update is to edit the term name: GO:0098582 innate localization behavior should have a v for vocalization

Ruth

Original comment by: RLovering

[gocentral]

Ughhh, how did I not spot that? Now fixed. Should go live tonight.

Original comment by: dosumis

[gocentral]

Hi David

I agree that - the process of learning and the process of performing something learned are different. But are they different wrt the genes required. Would you have a gene that would affect your ability to perform something learned but not affect your ability to learn something? And if there was such a gene how would you be able to test that it affected the performance and not the learning? Conversely, if you have a gene defect that prevents your brain from being able to learn to speak (or perform learned vocalizations) then you will not be able to demonstrate learned vocalization behaviour. Consequently, it order to have a 'learned vocalization behavior' I think you have to spend some time learning. So I would expect GO:0042297 vocal learning to have a part_of relationship to learned vocalization behaviour.

GO:0098583 learned vocalization behavior

Part_of GO:0042297 vocal learning

If we look at the visual learning ontology (definition: Any process in an organism in which a change in behavior of an individual occurs in response to repeated exposure to a visual cue) then I think that GO:0008306 associative learning would be a good is_a parent term for GO:0042297 vocal learning.

In addition you will see that GO:0007632 visual behavior is an is_a parent to GO:0008542 visual learning.

Looking at the definitions for these 2 terms GO:0042297 vocal learning: A behavioral process whose outcome is a relatively long-lasting adaptive behavioral change whereby an organism modifies innate vocalizations to imitate or create new sounds.

This definition does not include any reference to the learning process, perhaps it should be: A behavioral process whose outcome is a relatively long-lasting adaptive behavioral change, whereby an organism modifies innate vocalizations to imitate or create new sounds, following the recognition of a changed outcome in response to a specific vocalization.

GO:0098583 learned vocalization behavior: A vocalization behavior that is the result of learning. I think this definition breaks all GO rules with respect to just rearranging the words in the term to create a definition. Possible improvement: A vocal behavioral process that has been learned through recognition of a changed outcome in response to a specific vocalization.

Although I am still not completely convinced that these 2 are separate terms. The current definitions for these terms are too vague and so do not distinguish between them. Also I think that the expanded the definition for vocal learning could be applied to the learned vocalization behavior. If I had to choose which term to keep I would keep 'learned vocalization behavior'

Similarly the innate vocalization behavior term is rather badly defined (A vocalisation behavior that is innate), better might be: a vocalisation behavior that does not require cognition or consciousness to perform.

Also perhaps a new parent 'innate behaviour' term is needed: definition: http://en.wikipedia.org/wiki/Instinct Any behavior which does not require cognition or consciousness to perform.

Ruth

Original comment by: RLovering

[gocentral]

Hi Ruth,

It's taking me some time to think through and research responses to all your points here, so I've decided to break it down into separate chunks. Here's the first chunk, dealing with the distinctness of the two terms discussed and their relationship to each other.

---

On 8 Apr 2014, at 16:09, "Ruth" lovering@users.sf.net wrote:

I agree that - the process of learning and the process of performing something learned are different. But are they different wrt the genes required.

I would have thought so. We certainly shouldn't assume that there are not, for example, genes are involved in recall of memories but not in the initial learning process.

Would you have a gene that would affect your ability to perform something learned but not affect your ability to learn something? And if there was such a gene how would you be able to test that it affected the performance and not the learning?

Loss of function tests of this are possible with a conditional mutant that you can control via the environment (drug administration, heat shock, optogenetics). With that you could test the effects of knocking out or reducing function only during learning or only during recall. I'm pretty sure such experiments have already been carried out for conditioning experiments in Drosophila (can try to dig them up if you like). You might also see circumstantial evidence for a roll in recall based on activity of a gene product during recall but not learning or vice-versa.

Conversely, if you have a gene defect that prevents your brain from being able to learn to speak (or perform learned vocalizations) then you will not be able to demonstrate learned vocalization behaviour.

True. There is a causal relationship.

Consequently, it order to have a 'learned vocalization behavior' I think you have to spend some time learning.

Yep.

So I would expect GO:0042297 vocal learning to have a part_of relationship to learned vocalization behaviour.

GO:0098583 learned vocalization behavior Part_of GO:0042297 vocal learning

A causal relationship and partonomy are not the same thing. Unfortunately I don't think we yet have a suitable causal relationship. Perhaps the recently discussed 'causally related to' could work, but it feels a bit weak.

ACTION ITEM: Discuss possible causal relationships to use in cases like this with wider editor's group. (I will add this to the agenda for tomorrow's meeting).

Original comment by: dosumis

[gocentral]

On reflection, I can see that, while the role of a gene in vocal learning vs learned vocal vocal behavior will be distinguishable under circumstances, in many case they will not be. So, I think it makes sense to make a grouping term, even though this leads to potential maintenance issues. There is precedent for this with the term 'learning or memory': "The acquisition and processing of information and/or the storage and retrieval of this information over time."

So I've added the following grouping term:

+[Term] +id: GO:0098598 +name: learned vocalization behavior or vocal learning +namespace: biological_process +def: "Vocalisation behavior that is the result of learning, or the process by which new vocalizations are learned." [GOC:dos, PMID:16418265, PMID:17035521] +comment: This grouping term is necessary because, in the absence of conditional mutations, it is not possible to use phenotypic evidence to distinguish an effect on vocal learning from an effect on learned vocalisation behavior. +is_a: GO:0007611 ! learning or memory +created_by: davidos +creation_date: 2014-04-10T11:34:33Z

Original comment by: dosumis

[gocentral]

• status: closed-fixed --> open

Original comment by: dosumis

[gocentral]

On 8 Apr 2014, at 16:09, "Ruth" lovering@users.sf.net wrote:

I think that GO:0008306 associative learning would be a good is_a parent term for GO:0042297 vocal learning.

Here's the definition:

associative learning: "Learning by associating a stimulus (the cause) with a particular outcome (the effect)."

This is generally used for conditioning paradigms: animals learns to associate a stimulus with a reward or some second noxious stimulus. In this case I don't think there is necessarily a particular outcome associated with the stimulus. The animal is learning to imitate the stimulus. This is consistent with the refs on the term 'vocal learning':

http://www.pnas.org/content/103/4/1088.long http://www.ncbi.nlm.nih.gov/pubmed/17035521

I've added the following two terms (defined based on Wikipedia so far, but I'm planning to track down some decent primary refs):

+[Term] +id: GO:0098596 +name: imitative learning +namespace: biological_process +def: "Learning in which new behaviors are acquired through imitation." [GOC:dos] +is_a: GO:0098597 ! observational learning +created_by: davidos +creation_date: 2014-04-10T11:00:39Z + +[Term] +id: GO:0098597 +name: observational learning +namespace: biological_process +def: "Learning that occurs through observing the behavior of others." [GOC:dos] +comment: Observation here is used in a broad sense to include perception of the behavior of others via any form or combination of forms of sensory perception (visual, auditory etc). Observational learning is broader than imitative learning in that it does not require a duplication of the behavior exhibited by the model. +is_a: GO:0007612 ! learning +created_by: davidos +creation_date: 2014-04-10T11:07:10Z

I've also changed the definition of vocal learning slightly to limit it to learning by hearing and imitating a sounds:

id: GO:0042297 name: vocal learning namespace: biological_process -def: "A behavioral process whose outcome is a relatively long-lasting adaptive behavioral change whereby an organism modifies innate vocalizations to imitate or create new sounds." [PMID:16418265, PMID:17035521] +def: "A behavioral process whose outcome is a relatively long-lasting behavioral change whereby an organism modifies innate vocalizations to imitate sounds produced by others." [PMID:16418265, PMID:17035521] +comment: Examples include language learning by human infants and song learning in zebra finches. +is_a: GO:0098596 ! imitative learning +is_a: GO:0098598 ! learned vocalisation behavior or vocal learning

Original comment by: dosumis

[gocentral]

Final chunk (!)

On 8 Apr 2014, at 16:09, "Ruth" lovering@users.sf.net wrote:

Similarly the innate vocalization behavior term is rather badly defined (A vocalisation behavior that is innate), better might be: a vocalisation behavior that does not require cognition or consciousness to perform.

Also perhaps a new parent 'innate behaviour' term is needed: definition: http://en.wikipedia.org/wiki/Instinct Any behavior which does not require cognition or consciousness to perform.

Both of these suggestions make quite strong statements about the meaning of the term 'innate'. In particular, bringing up 'cognition' and 'conciousness' takes us into difficult and potentially controversial territory. It seems much simpler to me to have the definition hinge on whether learning is required or not. I've added:

+[Term] +id: GO:0098599 +name: innate behavior +namespace: biological_process +def: "Any behavior that does not need to be learned in order to occur." [GOC:dos] +is_a: GO:0044708 ! single-organism behavior +created_by: davidos +creation_date: 2014-04-10T11:53:49Z

id: GO:0098582 name: innate vocalization behavior namespace: biological_process -def: "A vocalisation behavior that is innate." [GO:dos, GO:RL] +def: "A vocalisation behavior that is innate, i.e. that does not need to be learned in order to occur." [GO:dos, GO:RL] is_a: GO:0071625 ! vocalization behavior +is_a: GO:0098599 ! innate behavior created_by: davidos creation_date: 2014-01-24T12:47:01Z

Please let me know if these changes look good and if so I'll commit.

Ta, David

Original comment by: dosumis

[gocentral]

Hi David

I think these all look good. Please remember that I have to generate reports for the charities that I am funded by, as this was started when I was funded only by BHF please ensure that GOC:BHF is included in all terms I have contributed too, so that when I search in AmiGO I can generate stats on the number of terms generated following requests from this project.

Also your dbxrefs are a bit variable above: ranging from GOC:dos, GO:dos, GO:RL. Please change GO:RL to GOC:rl so that I can also search for these terms.

Thanks

Ruth

Original comment by: RLovering

[gocentral]

• status: open --> closed-fixed

Original comment by: dosumis

[gocentral]

Fixed. See diff here for details: http://viewvc.geneontology.org/viewvc/GO-SVN/trunk/ontology/editors/gene_ontology_write.obo?r1=16940&r2=16945

Original comment by: dosumis