generality of class "device"

[Bayesianworld]

I am working with a group developing an ontology for radiation therapy. We are trying to incorporate as many existing classes as possible. OBI_0000968 "device", the definition states: A material entity that is designed to perform a function in a scientific investigation, but is not a reagent."

We would like to use this class but cannot use if if it only covers devices in a "scientific investigation". We are clinically oriented and do not always operate in the investigation mode (though sometimes we do). Is there any interest in making this class more general?

Mark Phillips University of Washington markp@uw.edu

[bpeters42]

Thanks for checking this! The question has come up before, and there are plans to create a more general class within COB ( https://github.com/OBOFoundry/COB). Already within OBI, we are effectively calling things devices that are designed for uses outside of investigations. So the quick answer is yes, and there are several other open issues (https://github.com/obi-ontology/obi/issues/1453 , https://github.com/OBOFoundry/COB/issues/190) .

The longer answer is that we are behind in acting on this, and need to push this forward.

On Fri, Nov 4, 2022 at 7:37 PM Mark PHillips @.***> wrote:

I am working with a group developing an ontology for radiation therapy. We are trying to incorporate as many existing classes as possible. OBI_0000968 "device", the definition states: A material entity that is designed to perform a function in a scientific investigation, but is not a reagent."

We would like to use this class but cannot use if if it only covers devices in a "scientific investigation". We are clinically oriented and do not always operate in the investigation mode (though sometimes we do). Is there any interest in making this class more general?

Mark Phillips University of Washington @.***

— Reply to this email directly, view it on GitHub https://github.com/obi-ontology/obi/issues/1626, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADJX2IWLMROBMLNIYODRASDWGXB6JANCNFSM6AAAAAARXWQHGU . 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

[ddooley]

I'll see if we can make this an item on next OBI call. Seems to me the simplest move is to change OBI "device" to "scientific device" given the definition. And add a new COB term "device".

[DanBerrios]

@Bayesianworld Be sure to check out OBO RBO. If you want to collaborate, let us know... https://github.com/Radiobiology-Informatics-Consortium/RBO

[DanBerrios]

Discussed on OBI Call 11/28:

1. COB should model "general device": a processed material that performs a function
2. Obsolete OBI:device
3. Replace with COB:"general device"
4. Move all current OBI:device classes to replacement COB:general device

@Bayesianworld This is our plan for widening the scope of OBI:device subclasses...will this meet your needs?

[ddooley]

@DanBerrios One thing I missed - I'd thought new COB term was going to literally be called "device", rather than "general device". Seems odd to have an upper level "general [x]" labeled class, and then have more specific labeled classes under it?!

[linikujp]

For your reference, the FDA definition of "device" is codified in section 201 of the FD&C Act.

The term “device” (except when used in paragraph (n) of this section and in sections 301(i), 403(f), 502(c), and 602(c)) means an instrument, apparatus, implement, machine, contrivance, implant, in vitro reagent, or other similar or related article, including any component, part, or accessory, which is— (A) recognized in the official National Formulary, or the United States Pharmacopeia, or any supplement to them, (B) intended for use in the diagnosis of disease or other conditions, or in the cure, mitigation, treatment, or prevention of disease, in man or other animals, or (C) intended to affect the structure or any function of the body of man or other animals, and which does not achieve its primary intended purposes through chemical action within or on the body of man or other animals and which is not dependent upon being metabolized for the achievement of its primary intended purposes. The term “device” does not include software functions excluded pursuant to section 520(o). (https://www.fda.gov/media/154866/download)

The wikipedia may cover a better definition for medical device (https://en.wikipedia.org/wiki/Medical_device)

General device in wikipedia: https://en.wikipedia.org/wiki/Device

[ddooley]

@linikujp it looks like you need a "medical device" or "regulated device" subclass? We need a general "device" parent term that allows for use other than in a medical role. "medical device" etc. can be classed under device.

[linikujp]

@ddooley to clarify, this definition is for your or COB group's reference to define a general device with reasonable solutions within the reasonable domain. For example, an ambulance can be a device used for medical purpose, but not my car. When you create this general device term, how will you treat terms like that? This is just for your consideration. After the general term is created, I will create a medical device as its child.

Besides, I have some general links from the wiki provided in my message...

Thanks, Asiyah

[Bayesianworld]

Glad to see the in-depth discussion. For our part (radiation therapy ontology), we are currently satisfied to put a bunch of different devices directly under the OBI or COB "device" or "general device". These include pretty specific devices such as collimators, patient positioning devices, imaging device and radiation therapy device. Each of these may have more specific children as needed. Is there a need to interpose a "medical device" between a "general device" for these classes?

[linikujp]

Hi Mark, Not necessarily. Are you devices used in a medical facility? If so, we can work together and define the term "medical device". However, note this definition may introduce some nuances, a sequencer used in a biology lab can be a medical device if it is used in a clinical lab. We need to figure out a startegy for that.

On Wed, Nov 30, 2022, 16:51 Mark PHillips @.***> wrote:

Glad to see the in-depth discussion. For our part (radiation therapy ontology), we are currently satisfied to put a bunch of different devices directly under the OBI or COB "device" or "general device". These include pretty specific devices such as collimators, patient positioning devices, imaging device and radiation therapy device. Each of these may have more specific children as needed. Is there a need to interpose a "medical device" between a "general device" for these classes?

— Reply to this email directly, view it on GitHub https://github.com/obi-ontology/obi/issues/1626#issuecomment-1332777521, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACCPEPNLYM7FVIZ5KQMXKW3WK7D5VANCNFSM6AAAAAARXWQHGU . You are receiving this because you were mentioned.Message ID: @.***>

[ddooley]

This was touched on in the OBI call. A thought is that it would be best to group devices by their functionality.

One could also define "regulated device" , "regulated medical device" which is about devices that have regulatory requirements attached to them (legislated licensing or performance specifications). Kind of hard to axiomatize that right now. This could be specialized into the various jurisdictions if desired. But these classes would remain unpopulated.

[ddooley]

I am following up with COB about defining "device" there, and OBI will then import this new term.

[bpeters42]

@ddooley, can you please create a definition etc. to discuss and then make a PR? Happy to also discuss with you offline.

[ddooley]

In this definition we assume devices are made of processed material, not natural artifacts, so we involve artifactual function rather than biological function, but align with a general BFO function sense where functions such as pumping, lifting occur in both contexts. Thus we can compare a biological arm with a robotic arm device.

COB device (because COB label's OBI "processed material" as "processed material entity") can be :

device: A processed material entity which is designed to perform a function.

This covers both "whole" devices like an engine, and components like a bolt. Rather than "is designed to perform a function" we could say "is designed to bear a function". It seems we have to put in the "designed" part to emphasize a devices primary function vs all the other possible dispositions a device may have that may also be useful. E.g. I can use a hammer for a paper weight. We can't now say a naturally formed rock is a hammering device - it doesn't have a hammering function per se.

(A side discussion is whether "tool" is device's more creative, disposition-rather-than-necessarily-function-bearing counterpart. Perhaps a device - or anything - is a tool if used for some disposition other than its function. A rock might be a hammering tool if it has a hammering disposition.)

(There is a psychological sense of device as ploy - "His temper tantrums were just a device for attracting attention." but we'll leave "psychological device" to be defined by behavioural science folks).

@bpeters42 happy to chat about this towards end of week! I'll move this over to a COB pull request after some feedback.

[Bayesianworld]

Referring to the earlier comments about whether it would be useful to include subclasses such as "medical" or "regulated" device. My question is perhaps a bit to fine, but is it better to make a subclass that is based on its intended use or its inherent function. As an example, agricultural and medical devices might both include a pump. In the spirit of making an ontology useful, it would seem that defining first a "pump device" and then further dividing into medical and agricultural devices would add a lot of unnecessary clutter. So I would be inclined to the first approach, but just thought I would check to see if there was a better alternative.

[ddooley]

It seems wise to have a generic functional hierarchy of devices, and place more specific devices under that. It would be nice to have a "medical device" class that could have subclasses inferred under it though by virtue of having some "regulated device role". And maybe some "precomposed" subclasses such as "medical pump" for convenience. Hmm!

[bpeters42]

Damion, I like your definition as proposed. It would be good to capture the other points you are making not as a loose discussion, but rather as examples, elucidations etc. that we can include in the COB term. The logical definition seems obvious from the textual: 'processed material' and 'has characteristic' some function.

I also think we should in parallel spell out how we expect devices to be used. We have the long form with 2 logical statements:

'completely executed planned process' and 'has specified input' some device
'completely executed planned process' and realizes some (function and ('characteristic of' some device))

And we have previously talked about wanting to have a 'utilizes device' relationship. That should be a shortcut that expands to the two above, and is between a planned process and the device. We could have a definition like; 'utilizes device' is a relationship between a planned process and a device, where the device is an input to the planned process and a function of the device is used as part of the planned process.

[ddooley]

Yay!

**label: device** **definition:** A processed material entity which is designed to perform a function. **example of usage:** A whole device like an engine; a component like a bolt. **comment**:

In this definition we assume devices are made of processed material, not natural artifacts, so we involve artifactual function rather than biological function, but align with a general BFO function sense where functions such as pumping, lifting occur in both contexts. Thus we can compare a biological arm with a robotic arm device. 

We say "designed" to emphasize a devices primary function vs all the other possible dispositions a device may have that may also be useful. E.g. one can use a hammer for a paper weight. We can't now say a naturally formed rock is a **hammering device** - it wasn't **designed** to bear a hammering function per se. However, a given rock may still happen to have the disposition to bear a hammering function, and so be used as a tool for that.

Currently "capable of/ capable of part of/ enables" are RO relations attaching devices to processes.

I see how a device could be considered an input to a process; but also how a device - like a robot (an agent) - is a participant in a process. So yes its still a bit wishy-washy there. A human even, is a device for carying out some processes like writing in the sand. Clarifications?

[linikujp]

" A human even, is a device for carying out some processes like writing. "

• A human can't be a device. They are disjoint classes. More accurately, is a human who is a participant of the process. Can we say robot a participant? I think it depends on the definition of participation.

When we say a device is capable of a planned process, which means the device performs its function (can be the designed function or second-use function that the device is not designed for) that is carried out ("realized") by the planned process. -- already covered by Bjoern.

On Thu, Dec 8, 2022 at 11:58 AM Damion Dooley @.***> wrote:

Yay!

label: device definition: A processed material entity which is designed to perform a function. example of usage: A whole device like an engine; a component like a bolt. comment:

In this definition we assume devices are made of processed material, not natural artifacts, so we involve artifactual function rather than biological function, but align with a general BFO function sense where functions such as pumping, lifting occur in both contexts. Thus we can compare a biological arm with a robotic arm device.

We say "designed" to emphasize a devices primary function vs all the other possible dispositions a device may have that may also be useful. E.g. one can use a hammer for a paper weight. We can't now say a naturally formed rock is a hammering device - it wasn't designed to bear a hammering function per se. However, a given rock may still happen to have the disposition to bear a hammering function, and so be used as a tool for that.

Currently "capable of/ capable of part of/ enables" are RO relations attaching devices to processes.

[image: image] https://user-images.githubusercontent.com/4000582/206515435-338ad406-0b47-4135-b9c8-2e8633729f81.png

I see how a device could be considered an input to a process; but also how a device - like a robot (an agent) - is a participant in a process. So yes its still a bit wishy-washy there. A human even, is a device for carying out some processes like writing. Clarifications?

— Reply to this email directly, view it on GitHub https://github.com/obi-ontology/obi/issues/1626#issuecomment-1343018535, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACCPEPJLR4R4GZ5K53PPEPTWMIHR3ANCNFSM6AAAAAARXWQHGU . You are receiving this because you were mentioned.Message ID: @.***>

[ddooley]

About human - my bad, right you are! And I see that robot counts as device.
Given that "capable of/ capable of part of/ enables" can connect devices including robots to processes, do we drop:

'completely executed planned process' and 'has specified input' some device

Also, not all executed planned processes need devices. I can shape dough with my hands, but I am not a device. (I use my hands as tools though.)

Do we also want "has performer" to point to a robot too, per existing example: OBI "has performer" has a mistaken annotation BTW - it should be reshaped into a definition and a comment:

term editor: performer relation covers the need to report on who performed a planned processed. it has to cover processes done by People or Devices (such as a robot controlled by software WF management system)

Or should "has performer" be tweaked to reference only entities that can be considered to bear responsibility/intentionality for seeing a process to completion?

[bpeters42]

has_specified_input is a sub-property of has_participant. So there is no conflict. Humans aren't processed materials, so they are not devices (like Asiyah said as well).

And I don't think we should drop the second statement. We want to list all participants in a planned process. We need both statements because in OWL we can't say that the function being realized is borne by a participants.

Regarding 'capable of' etc., 'utilizes device' is the inverse. I am unclear if there is an existing relationship for that in RO. Note that inverses of shortcut relations are not trivial.

[linikujp]

Besides "performer", another important relation to catch is the person who "operates" the device. As the doctors are critical element to evaluate in a medical procedure, the doctor who operates either a robot or a traditional device in a medical procedure is important to catch. In this case, I'd like to constrain the "operator" as human, so I can later trace back if the different doctor's skills affects the different outcomes of an operation. Driver operates the cars to transport goods or people can be another example here.

Thanks, Asiyah

On Thu, Dec 8, 2022 at 4:26 PM bpeters42 @.***> wrote:

has_specified_input is a sub-property of has_participant. So there is no conflict. Humans aren't processed materials, so they are not devices (like Asiyah said as well).

And I don't think we should drop the second statement. We want to list all participants in a planned process. We need both statements because in OWL we can't say that the function being realized is borne by a participants.

Regarding 'capable of' etc., 'utilizes device' is the inverse. I am unclear if there is an existing relationship for that in RO. Note that inverses of shortcut relations are not trivial.

— Reply to this email directly, view it on GitHub https://github.com/obi-ontology/obi/issues/1626#issuecomment-1343381003, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACCPEPIKQHNAE4ZFPTO5LLDWMJG7BANCNFSM6AAAAAARXWQHGU . You are receiving this because you were mentioned.Message ID: @.***>

[ddooley]

@bpeters42 I think I understand now. RO does have

that connect between materials (including catalysts and devices) and processes (and these OPs have inverses. "utilizes device" has been deprecated once and for as I recall?)

But aside from saying generally that a device can enable a process, we want at an instance level to say that a particular device, or device type, was a participant in a process. If the convention is to say a device is a kind of input, I can handle that, I just hadn't seen that before.

So an updated diagram is [with update "measurement device"]:

Suggestions for "measures" OP label ? And should it have range all characteristics, or just those of material things since assay can only input material things?

[ddooley]

The "measures" discussion has been moved to: https://github.com/oborel/obo-relations/issues/658

[DanBerrios]

COB has now created a more general device https://ontobee.org/ontology/COB?iri=http%3A%2F%2Fpurl.obolibrary.org%2Fobo%2FCOB_0001300

Can we import it as a parent of OBI:device? Would we also want to re-label OBI:device as "investigation device"?

[ddooley]

Sounds good. It will require promoting many subclasses of OBI device up to COB device subclasses since the OBI one has been the catch-all so far. E.g. "apron" and "container" are more general than an "investigation device" would afford.

[DanBerrios]

@ddooley Agreed.

[bpeters42]

I thought we simply obsolete the OBI:device class, and replace it with COB device.

On Thu, Nov 30, 2023 at 2:27 PM Dan Berrios @.***> wrote:

@ddooley https://github.com/ddooley Agreed.

— Reply to this email directly, view it on GitHub https://github.com/obi-ontology/obi/issues/1626#issuecomment-1834659434, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADJX2IVVWPU6FOHZ3CHEFB3YHEB45AVCNFSM6AAAAAARXWQHGWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQMZUGY2TSNBTGQ . You are receiving this because you were mentioned.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

[sebastianduesing]

I believe that the recent progress on adopting "device" as a COB term may be sufficient to close this issue.