NTRs: behavior and human activity #156
Comments
|
I've been discussing a similar issue with Chris Stoeckert on the Apollo-SV
listserv: ApolloDev/apollo-sv#199
We're happy to model human activity in OMRSE, perhaps as some subtype of
planned process.
…On Tue, Mar 23, 2021 at 4:01 PM Chris Mungall ***@***.***> wrote:
We should add behavior and link to NBO. We should follow NBO and make this
a neurologically influenced behavior and not behavior in the very general
sense.
We also need to think about where human activities like walking, smoking,
going to the gym, etc fit in. I don't think they are in scope for NBO as
these are not things a neurobiologist studies, and I don't think we should
even put these in OBO as they are not biological. But NCIT has a decent
starting vocabulary - see obo-behavior/behavior-ontology#109
<obo-behavior/behavior-ontology#109>
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<#156>, or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJR55XTEK3IUJTVLIAKS5DTFDXTDANCNFSM4ZV3722Q>
.
|
👍
Lots of need for these for GWAS and whole genome sequencing association data. They need a reliable home. Some potential issues with boundaries here. "Going for a walk" may be a planned process, but walking itself is a behaviour controlled by particular circuits. And how do I record the behavioral results of messing with decision making circuits in the brain of an animal (lots of this type of working happening in Drosophila right now). Do we make a distinction between human planning and animal decision making? |
|
Just upping this as it came up in @hoganwr's ICBO talk. I think people are using GO hevavior to categorize processes that are not genetically determined. Defining behavior is hard. See: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2760923/ I suggest the following as a pragmatic approach:
Here the two classes would be class-disjoint. They could be linked by parthood relation at the domain ontology level. E.g with @dosumis' example, the NCIT:going-for-a-walk activity may have a has-part to GO:walking/locomotory behavior. COB would be neutral about when this would be done. This approach can be criticised in that it can be really hard to know the boundaries. There will always be debates as to the extent of genetic influence on behavior, and things vary by species. We may end up with a bit of concept duplication and artificial labels to distinguish terms in the two hierarchies. But it's very pragmatic. If there is evidence for the involvement in genes then the concept can be added to GO, if there are annotations. Otherwise keep it outside. There is definitely much to be improved. I think the labels for GO behaviors should probably be modified to make their restricted sense clear. But my proposal here can go ahead without being blocked on this. I don't know what that means for NBO, but it's inactive, and not really a COB problem. We just need a solution for people who need "behaviors" that don't belong in GO. |
|
+1
I support this approach.
Matt, Clint, and I have done some work defining “human activity” but it has
already been criticized on at least one issue thread (for some ontology) as
being too species specific.
I think it needs to be more general as you say. This would be very helpful
in a number of areas for OMRSE and other ontologies we’re working on.
…On Tue, Sep 27, 2022 at 2:59 PM Chris Mungall ***@***.***> wrote:
Just upping this as it came up in @hoganwr <https://github.com/hoganwr>'s
ICBO talk. I think people are using GO hevavior to categorize processes
that are not genetically determined.
Defining behavior is hard. See:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2760923/
I suggest the following as a pragmatic approach:
- COB: process
- COB(new): behavioral action or activity of an organism
- going for a walk
- taking a class
- aggressive actions or activities
- COB/GO: biological process: *genetically programmed*
- GO(+COB-new?):behavior *genetically programmed*
- GO:learning or memory
- GO:locomotory behavior
- GO:aggressive behavior
Here the two classes would be class-disjoint. They could be linked by
parthood relation at the domain ontology level. E.g with @dosumis
<https://github.com/dosumis>' example, the NCIT:going-for-a-walk activity
may have a has-part to GO:walking/locomotory behavior. COB would be neutral
about when this would be done.
This approach can be criticised in that it can be really hard to know the
boundaries. There will always be debates as to the extent of genetic
influence on behavior, and things vary by species. We may end up with a bit
of concept duplication and artificial labels to distinguish terms in the
two hierarchies.
But it's very pragmatic. If there is evidence for the involvement in genes
then the concept can be added to GO, if there are annotations. Otherwise
keep it outside.
There is definitely much to be improved. I think the labels for GO
behaviors should probably be modified to make their restricted sense clear.
But my proposal here can go ahead without being blocked on this.
I don't know what that means for NBO, but it's inactive, and not really a
COB problem. We just need a solution for people who need "behaviors" that
don't belong in GO.
—
Reply to this email directly, view it on GitHub
<#156 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJR55UCF3KGCS2OTBXMUO3WAM7Z5ANCNFSM4ZV3722Q>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
I am all in favor of giving these processes a home in COB, but would like
to see the connection to 'planned process' being worked out well. It seems
that with organism behavior, there typically is an 'objective', even if
there isn't a full scale plan, so that needs to be made compatible.
- Bjoern
On Tue, Sep 27, 2022 at 12:22 PM Bill Hogan ***@***.***>
wrote:
… +1
I support this approach.
Matt, Clint, and I have done some work defining “human activity” but it has
already been criticized on at least one issue thread (for some ontology) as
being too species specific.
I think it needs to be more general as you say. This would be very helpful
in a number of areas for OMRSE and other ontologies we’re working on.
On Tue, Sep 27, 2022 at 2:59 PM Chris Mungall ***@***.***>
wrote:
> Just upping this as it came up in @hoganwr <https://github.com/hoganwr
>'s
> ICBO talk. I think people are using GO hevavior to categorize processes
> that are not genetically determined.
>
> Defining behavior is hard. See:
> https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2760923/
>
> I suggest the following as a pragmatic approach:
>
> - COB: process
> - COB(new): behavioral action or activity of an organism
> - going for a walk
> - taking a class
> - aggressive actions or activities
> - COB/GO: biological process: *genetically programmed*
> - GO(+COB-new?):behavior *genetically programmed*
> - GO:learning or memory
> - GO:locomotory behavior
> - GO:aggressive behavior
>
> Here the two classes would be class-disjoint. They could be linked by
> parthood relation at the domain ontology level. E.g with @dosumis
> <https://github.com/dosumis>' example, the NCIT:going-for-a-walk
activity
> may have a has-part to GO:walking/locomotory behavior. COB would be
neutral
> about when this would be done.
>
> This approach can be criticised in that it can be really hard to know the
> boundaries. There will always be debates as to the extent of genetic
> influence on behavior, and things vary by species. We may end up with a
bit
> of concept duplication and artificial labels to distinguish terms in the
> two hierarchies.
>
> But it's very pragmatic. If there is evidence for the involvement in
genes
> then the concept can be added to GO, if there are annotations. Otherwise
> keep it outside.
>
> There is definitely much to be improved. I think the labels for GO
> behaviors should probably be modified to make their restricted sense
clear.
> But my proposal here can go ahead without being blocked on this.
>
> I don't know what that means for NBO, but it's inactive, and not really a
> COB problem. We just need a solution for people who need "behaviors" that
> don't belong in GO.
>
> —
> Reply to this email directly, view it on GitHub
> <#156 (comment)>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/AAJR55UCF3KGCS2OTBXMUO3WAM7Z5ANCNFSM4ZV3722Q
>
> .
> You are receiving this because you were mentioned.Message ID:
> ***@***.***>
>
—
Reply to this email directly, view it on GitHub
<#156 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADJX2ISDFQQYRBQFAWERGKDWANCRDANCNFSM4ZV3722Q>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
--
Bjoern Peters
Professor
La Jolla Institute for Immunology
9420 Athena Circle
La Jolla, CA 92037, USA
Tel: 858/752-6914
Fax: 858/752-6987
http://www.liai.org/pages/faculty-peters
|
|
Agree. They are not the same. Certain compulsive behaviors are not
programmed nor are they planned. They are closely related and teasing out
the distinctions will be important.
…On Tue, Sep 27, 2022 at 3:50 PM bpeters42 ***@***.***> wrote:
I am all in favor of giving these processes a home in COB, but would like
to see the connection to 'planned process' being worked out well. It seems
that with organism behavior, there typically is an 'objective', even if
there isn't a full scale plan, so that needs to be made compatible.
- Bjoern
On Tue, Sep 27, 2022 at 12:22 PM Bill Hogan ***@***.***>
wrote:
> +1
>
> I support this approach.
>
> Matt, Clint, and I have done some work defining “human activity” but it
has
> already been criticized on at least one issue thread (for some ontology)
as
> being too species specific.
>
> I think it needs to be more general as you say. This would be very
helpful
> in a number of areas for OMRSE and other ontologies we’re working on.
>
> On Tue, Sep 27, 2022 at 2:59 PM Chris Mungall ***@***.***>
> wrote:
>
> > Just upping this as it came up in @hoganwr <https://github.com/hoganwr
> >'s
> > ICBO talk. I think people are using GO hevavior to categorize processes
> > that are not genetically determined.
> >
> > Defining behavior is hard. See:
> > https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2760923/
> >
> > I suggest the following as a pragmatic approach:
> >
> > - COB: process
> > - COB(new): behavioral action or activity of an organism
> > - going for a walk
> > - taking a class
> > - aggressive actions or activities
> > - COB/GO: biological process: *genetically programmed*
> > - GO(+COB-new?):behavior *genetically programmed*
> > - GO:learning or memory
> > - GO:locomotory behavior
> > - GO:aggressive behavior
> >
> > Here the two classes would be class-disjoint. They could be linked by
> > parthood relation at the domain ontology level. E.g with @dosumis
> > <https://github.com/dosumis>' example, the NCIT:going-for-a-walk
> activity
> > may have a has-part to GO:walking/locomotory behavior. COB would be
> neutral
> > about when this would be done.
> >
> > This approach can be criticised in that it can be really hard to know
the
> > boundaries. There will always be debates as to the extent of genetic
> > influence on behavior, and things vary by species. We may end up with a
> bit
> > of concept duplication and artificial labels to distinguish terms in
the
> > two hierarchies.
> >
> > But it's very pragmatic. If there is evidence for the involvement in
> genes
> > then the concept can be added to GO, if there are annotations.
Otherwise
> > keep it outside.
> >
> > There is definitely much to be improved. I think the labels for GO
> > behaviors should probably be modified to make their restricted sense
> clear.
> > But my proposal here can go ahead without being blocked on this.
> >
> > I don't know what that means for NBO, but it's inactive, and not
really a
> > COB problem. We just need a solution for people who need "behaviors"
that
> > don't belong in GO.
> >
> > —
> > Reply to this email directly, view it on GitHub
> > <#156 (comment)
>,
> > or unsubscribe
> > <
>
https://github.com/notifications/unsubscribe-auth/AAJR55UCF3KGCS2OTBXMUO3WAM7Z5ANCNFSM4ZV3722Q
> >
> > .
> > You are receiving this because you were mentioned.Message ID:
> > ***@***.***>
> >
>
> —
> Reply to this email directly, view it on GitHub
> <#156 (comment)>,
> or unsubscribe
> <
https://github.com/notifications/unsubscribe-auth/ADJX2ISDFQQYRBQFAWERGKDWANCRDANCNFSM4ZV3722Q
>
> .
> You are receiving this because you are subscribed to this thread.Message
> ID: ***@***.***>
>
--
Bjoern Peters
Professor
La Jolla Institute for Immunology
9420 Athena Circle
<https://www.google.com/maps/search/9420+Athena+Circle+%0D%0D%0ALa+Jolla,+CA+92037,+USA?entry=gmail&source=g>
La Jolla, CA 92037, USA
<https://www.google.com/maps/search/9420+Athena+Circle+%0D%0D%0ALa+Jolla,+CA+92037,+USA?entry=gmail&source=g>
Tel: 858/752-6914
Fax: 858/752-6987
http://www.liai.org/pages/faculty-peters
—
Reply to this email directly, view it on GitHub
<#156 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAJR55V3YS34ER5HBB3BQQDWANFWTANCNFSM4ZV3722Q>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
It is very naive to think we can easily make simple distinctions between genetically programmed / reflexive vs learned behaviour. Even the simplest behaviours (classic reflexes) can be modified by experience. Even complex behaviours rely on evolved circuits structured according to genetically determined developmental programs. Naming a term "behaviour - genetically programmed" will cause lots of well justified push back from biologists. (Actually - that's an understatement - it will make GO look ridiculous to most potential users working in neurobiology/behavior). I think we can split out human, culturally-located(?) behaviour terms ("going to the gym", "smoking", "taking a class", "getting married"). But I'm very wary of doing anything more than that. If GO wants to reject a wider range of terms than this, then these terms will need another home, and if they don't want these terms to live under GO process in a different namespace (e.g. NBO), then we need a different grouping term. |
I'm not convinced. For the OBI use case - distinguishing types of process in an experiment - 'planned process' totally makes sense. In the context of annotating behaviours in, say, a GWAS experiment, it does not.
Other species have culture too . Maybe: 'culturally-located activity' as a sibling term to behaviour? |
|
The concept of objectives, plans and actions is very well defined far
beyond OBI. Calling something "culturally-located(?) behaviour term" sounds
super ridiculous to me, but clearly we are coming this from different
perspectives. It would be useful if you could define what you mean; the OBI
definitions are out there to review, and are supposed to be part of COB
specifically because they have a broader scope than investigations. I am
sure they can be improved, but not everything has to start and end in GO.
- Bjoern
…On Wed, Sep 28, 2022 at 1:32 AM David Osumi-Sutherland < ***@***.***> wrote:
would like to see the connection to 'planned process' being worked out
well. It seems that with organism behavior, there typically is an
'objective', even if
there isn't a full scale plan, so that needs to be made compatible.
I'm not convinced. For the OBI use case - distinguishing types of process
in an experiment - 'planned process' totally makes sense. In the context of
annotating behaviours in, say, a GWAS experiment, it does not.
Matt, Clint, and I have done some work defining “human activity” but it has
already been criticised on at least one issue thread (for some ontology) as
being too species specific.
Other species have culture too maybe:
'culturally-located activity' as a sibling term to behaviour?
—
Reply to this email directly, view it on GitHub
<#156 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADJX2IRSHE7ZQ4DBQ6ZGZ33WAP7BRANCNFSM4ZV3722Q>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Bjoern Peters
Professor
La Jolla Institute for Immunology
9420 Athena Circle
La Jolla, CA 92037, USA
Tel: 858/752-6914
Fax: 858/752-6987
http://www.liai.org/pages/faculty-peters
|
|
I would stay away from using |
|
My response was careless: I was intending to only refer to the 'human
activities' part that Chris had outlined (human activities like walking,
smoking, going to the gym, ...) and Bill responded to as suggesting they
could go under planned process. I agree that there might be more broader
parent terms for behaviors, genetically determined or not, and I don't have
strong feelings how they would be modeled as long as we don't build a
parallel overlapping set of terms to 'planned process' which has served us
very well.
- Bjoern
…On Thu, Sep 29, 2022 at 7:21 AM Bill Duncan ***@***.***> wrote:
I would stay away from using objective specification as a general
differentia for defining behavior. You wade deeper into the murky idea that
concretizations can be realizables (I still don't think that make sense).
Plus, it seems (to me) to be asserting that every behavior an organism
engages in has to have objective specification related to it, regardless
of how loosely defined the objective specification may be. I am skeptical
that frogs, flies, amoeba, and hydra have an objective specification
related to every behavior they engage in.
—
Reply to this email directly, view it on GitHub
<#156 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADJX2IXQ7CIFHLWVOIWZIHLWAWQWHANCNFSM4ZV3722Q>
.
You are receiving this because you commented.Message ID:
***@***.***>
--
Bjoern Peters
Professor
La Jolla Institute for Immunology
9420 Athena Circle
La Jolla, CA 92037, USA
Tel: 858/752-6914
Fax: 858/752-6987
http://www.liai.org/pages/faculty-peters
|
GO behavior: "The internally coordinated responses (actions or inactions) of animals (individuals or groups) to internal or external stimuli, via a mechanism that involves nervous system activity" is currently a subclass of GO biological process, "A biological process represents a specific objective that the organism is genetically programmed to achieve." I agree that this is problematic on a few levels but from the perspective of this issue, that is the definitions in GO. If we want to discuss changing definitions in GO, this should be done on the GO issue tracker. |
|
So it seems like right now we have something like:
And I believe we are discussing variants such as this:
to avoid that all 'behaviors inherit the axioms from GO: biological process, which imply genetic determination? I would be very much in favor of that. |
|
OBI and others are putting assays, calibration, data transformation etc. under planned process - and robots or devices may be executing them, so "planned process" would be peer to organismal process. Maybe a new term "organismal planned process" could be peer to "biological process" though. |
We should add behavior and link to NBO. We should follow NBO and make this a neurologically influenced behavior and not behavior in the very general sense.
We also need to think about where human activities like walking, smoking, going to the gym, etc fit in. I don't think they are in scope for NBO as these are not things a neurobiologist studies, and I don't think we should even put these in OBO as they are not biological. But NCIT has a decent starting vocabulary - see obo-behavior/behavior-ontology#109
behavior is also mentioned in #42
The text was updated successfully, but these errors were encountered: