Material entities and their representations

[trypuz]

IMO, the current version of the IDMP ontology has a problem with making a clear distinction between material entities and their representations, which at some point can be a source of confusion. Some concepts are overloaded, e.g., a structure which can be 1) an arrangement and organization of interrelated elements in a material object or 2) a representation of the collection of atoms. In consequence, e.g., the molecular weight applies to substance and representation of substances. Is there any justification for this modeling choice?

[ElisaKendall]

@trypuz Hi Robert, It's nice to hear from you. For our initial use case, we attempted to represent what was in the ISO 11238 standard, which itself is somewhat ambiguous (thus the need for an ontology), but the only use in terms of our use case we had for structure was to relate the molecular structure to Chemantix in this first iteration. The scientific distinctions between a substance and a molecule are not clear in the IDMP standards, and in most cases the standard ignores that - not relevant for the labeling that IDMP is focused on. I suspect that we will end up doing more work to represent the scientific concepts properly, and add viewpoints with respect to the regulatory view vs. the scientific view as we tackle more use cases, but we are not there yet :).

Please feel free to keep reminding us, and please also contribute to the use cases that would tease this out further. Note that I will continue to hold our Monday calls on definitions even though we are slowing down a bit before kicking off the next phase of work next month. Quentin Darrasse has the call in details (I think he is the main participant from Roche), but I can add you to those calls/send you the webex invitation if you like. They are on Mondays at 7 am pacific time, 4 pm CET.

[trypuz]

Hi Elisa! AFAIC, the distinction between a substance and a molecule is clear. BTW, BFO would be really helpful here ;-) IMO, what's challenging is a model/ontology that not only reflects what the link between those two is but also is handy in real applications. In the context of IDMP, it is unlikely that a molecule (class) will be used to classify real molecules, whereas it is almost certain that a substance (class) will be used to classify portions of a "real" substance. And when it comes to the substance itself, in most cases, only information about the types of molecules that compose a substance will be enough. This means that we have to be at the same time smart and precise in making a clear distinction between classes that classify real/material entities and those that will be just information content entities (using the terminology of IAO). And the worst that may happen is to have classes in ontology (like "substance" in the current version of the ontology) that, on one occasion, classify material objects and, on another occasion, representations of material objects. I'll talk to Roche's RTS team colleagues and see if we can join your weekly meeting. Best!

[ElisaKendall]

Hi Robert, I think as we move forward with the ontology, especially with new use cases that may require more definition of substances than is present in the IDMP standards, we will end up teasing more of this out. At present, the focus has mainly been on answering the first set of competency questions, which we can now do. But the next series of use cases will likely get more complicated, and so will help us refine some of these details. We will be adding more science, in other words, and as we do so the distinctions will become clearer. The IDMP standards themselves overload all kinds of things, and have a very relational orientation that we need to address from a modeling perspective - such that the various pharma companies can align their data with the resulting ontology but without compromising the meaning as the ontology evolves. I am not sure how far Roche is in aligning data with the competency questions yet, but look forward to hearing how you get on with that.

[fabien-roche]

Hi Elisa.
I work with Robert in Roche. I have several points about "Medicinal product". The first one is: Based on the IDMP definition: "A substance or combination of substances that is intended to treat, prevent or diagnose a disease, or to restore, correct or modify physiological functions by exerting a pharmacological, immunological or metabolic action " it seems that medications that are not substances such as cells or virus (e.g. in cell or virus therapies) are not considered as medicinal product. What are they then according to IDMP or ISO?
The second point is: what do you mean by "combination of substances"?
The third point is about the term "substance" and is more a comment. As you mention above ISO does not have a scientific view and does distinguish between substances and molecules yet. The pharma industry is a highly based on science and we currently need this scientific view in the ISO standard. Moreover, the descriptions of substances, single substance are quite cryptic. A lay language description would help a lot.
Many thanks in advance for addressing my points.
BTW, Robert and I would like to attend the Monday meetings. Would you mind inviting us? Many thanks.

[ElisaKendall]

@fabien-roche @trypuz Hi Fabien, nice to meet you!
In the ISO 11615 and in the ontology, we have the following definitions:

1. For pharmaceutical product - 'qualitative and quantitative composition of a medicinal product in the dose form approved for administration in line with the regulated product information' (from 11615, clause 3.1.60)
2. For medicinal product - 'pharmaceutical product or combination of pharmaceutical products that may be administered to human beings (or animals) for treating or preventing disease, with the aim/purpose of making a medical diagnosis or to restore, correct or modify physiological functions' (from 11615, clause 3.1.50, and from 11238, clause 3.42)
3. There is a note in the standard that we have included in the ontology on the medicinal product class that is not the actual definition, but a note quoted from the standard, that does say "A medicinal product may contain in the packaging one or more manufactured items and one or more pharmaceutical products. In certain regions, a medicinal product may also be defined as any substance or combination of substances which may be used to make a medical diagnosis.", which includes the phrase you mention, but that's not the primary definition of medicinal product. That note can be found in "Note 1 to entry", clause 3.1.50, in ISO 11615, but it specifically says that it is in the context of some region.

I think the broad brush definition does cover cells or virus therapies based on what I have learned from other pharma team members including GSK's vaccine group, but we could talk about that on one of our Monday calls. I would also think that a "combination of substances" might be a mixture, or might be more general than that as a substance itself, but again I would defer to one of the SMEs on our team to confirm.

With respect to additional scientific information, in our next iteration and based on the competency questions outlined by the SMEs, we may be adding more science, including a proper definition of molecule, but we haven't integrated that yet, and we are still working out how to bridge the regulatory / labeling / science perspectives. We met with FDA at the Pistoia Alliance conference, and they do recognize that there are some things that need to be addressed in the standard, for example, and that they are interested in our feedback as we build out the ontology. Fortunately, in addition to FDA, a number of the folks who have contributed to the ontology to date are also working on the ISO standards, at least one of which is currently under review from an ISO perspective. We and they would be happy to have your feedback to make this better. Adding "lay language" could be done in explanatory notes in the ontology, even if those notes are not in the core standards, but I would want concurrence across pharma companies for that. Then we may be able to influence the next version of the ISO standard, but no promises on that front since I'm not on the committee :).

I need your direct email addresses (I might have Robert's somewhere, but better if you resend both to me) to add you to the meeting. Quentin Darrasse has been participating on behalf of Roche, so I'll copy him on the invitation to you both. Please email me directly at [email protected], and I will be delighted to do so. Note that for the moment, that meeting has involved SMEs primarily rather than ontologists per se. The ontology meeting is on Wednesdays - I will ask Heiner to invite you to that meeting, which might be more relevant for you.

[fabien-roche]

Thanks Elisa. Email sent. Based on our experience, this is crucial that SMEs and ontologists actively and directly talk each other. Very often SMEs do not see the scientific or ontological issues, and ontologists do not always understand the SMEs' points of view. My 2c suggestion,

[fabien-roche]

Hi Elisa,
The ISO definitions are even more confusing to me because 1. they are circular between medicinal product and pharmaceutical product, 2. its seems that there is a clash in the notion of parthood between these 2 concepts. I may miss something. However, my naive/fresh view highlights that something needs to be improved in those definitions.
Those definitions may cover my use cases. Let's confirm it after they are clarified.

BTW the definition I provided above was from EMEA (https://www.ema.europa.eu/en/glossary/medicinal-product) . I am very surprised that there is not alignment between ISO/IDMP and EMEA on that matter yet.

Another point is that the ISO definitions investigational medicinal product and medicinal product do not seem to me consistent/aligned. On that matter, the EMEA definition is more consistent: (A substance or combination of substances) that is intended to treat, prevent or diagnose a disease, or to restore, correct or modify physiological functions by exerting a pharmacological, immunological or metabolic action.

Lastly, I would like to clarify my words "definitions in lay language". We have a good practice principle in our team when creating/reviewing a concept: the definition and the preferred label of a concept must be self-explaining without using jargon and expert's knowledge. Notes are usually not available or read by users. Thus, we avoid to explain a definition in notes. However, providing examples in notes can be useful.

[fabien-roche]

Hi Elisa, Our email addresses are cced. Thanks for your feedback. I will think of them tomorrow. Have a nice evening. Best Regards, Fabien *** Please consider using our Webform for any kinds of requests to the RTS team *** https://webforms.roche.com/RTSSC *--* *Fabien RICHARD* Scientific Knowledge Engineer RTS Data Harmonization Services Scientific Solution Engineering & Architecture (S2EA IV) Roche Pharma Research and Early Development Informatics Roche Innovation Center Basel F. Hoffmann-La Roche AG Building 92-06.NBH01 CH-4070 Basel, Switzerland Office: +41 61 68 73074 Mobile: +41 79 81 31 847 ***@***.*** ***@***.***>

…

On Thu, Nov 10, 2022 at 5:53 PM Elisa Kendall ***@***.***> wrote: @fabien-roche <https://github.com/fabien-roche> @trypuz <https://github.com/trypuz> Hi Fabien, nice to meet you! In the ISO 11615 and in the ontology, we have the following definitions: 1. For pharmaceutical product - 'qualitative and quantitative composition of a medicinal product in the dose form approved for administration in line with the regulated product information' (from 11615, clause 3.1.60) 2. For medicinal product - 'pharmaceutical product or combination of pharmaceutical products that may be administered to human beings (or animals) for treating or preventing disease, with the aim/purpose of making a medical diagnosis or to restore, correct or modify physiological functions' (from 11615, clause 3.1.50, and from 11238, clause 3.42) 3. There is a note in the standard that we have included in the ontology on the medicinal product class that is not the actual definition, but a note quoted from the standard, that does say "A medicinal product may contain in the packaging one or more manufactured items and one or more pharmaceutical products. In certain regions, a medicinal product may also be defined as any substance or combination of substances which may be used to make a medical diagnosis.", which includes the phrase you mention, but that's not the primary definition of medicinal product. That note can be found in "Note 1 to entry", clause 3.1.50, in ISO 11615, but it specifically says that it is in the context of some region. I think the broad brush definition does cover cells or virus therapies based on what I have learned from other pharma team members including GSK's vaccine group, but we could talk about that on one of our Monday calls. I would also think that a "combination of substances" might be a mixture, or might be more general than that as a substance itself, but again I would defer to one of the SMEs on our team to confirm. With respect to additional scientific information, in our next iteration and based on the competency questions outlined by the SMEs, we may be adding more science, including a proper definition of molecule, but we haven't integrated that yet, and we are still working out how to bridge the regulatory / labeling / science perspectives. We met with FDA at the Pistoia Alliance conference, and they do recognize that there are some things that need to be addressed in the standard, for example, and that they are interested in our feedback as we build out the ontology. Fortunately, in addition to FDA, a number of the folks who have contributed to the ontology to date are also working on the ISO standards, at least one of which is currently under review from an ISO perspective. We and they would be happy to have your feedback to make this better. Adding "lay language" could be done in explanatory notes in the ontology, even if those notes are not in the core standards, but I would want concurrence across pharma companies for that. Then we may be able to influence the next version of the ISO standard, but no promises on that front since I'm not on the committee :). I need your direct email addresses (I might have Robert's somewhere, but better if you resend both to me) to add you to the meeting. Quentin Darrasse has been participating on behalf of Roche, so I'll copy him on the invitation to you both. Please email me directly at ***@***.***, and I will be delighted to do so. Note that for the moment, that meeting has involved SMEs primarily rather than ontologists per se. The ontology meeting is on Wednesdays - I will ask Heiner to invite you to that meeting, which might be more relevant for you. — Reply to this email directly, view it on GitHub <#219 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/A4DXKPLGDGLU7CXS6QMDEFTWHUSCLANCNFSM6AAAAAAQ5J3724> . You are receiving this because you were mentioned.Message ID: ***@***.***>