material separation and purification need disambiguation

[obi-bot]

I was going over some of the classes recently added and spotted the following:

Class: purification no formal differentiation with material separation, which lead to the question: Is the distinction needed? There are several editor notes pointing to discussions point we had 2 years back. I vaguely recall lengthy discussions about this as it as also related to the notion of 'conferred quality' (i.e quality acquired following a process)

Consequence: DNA extraction is not classified under purification but under material separation simply because the objective is set to "material separation objective".

One fix is to set the right objective (e.g purification objective) but it might be better to have constraints on the materials involved in the actual process (which could/would avoid having to duplicate the hierarchy between processes and objectives as a by product). Essentially, it is about focusing of what makes the materials different before and after the process they undergo.

So I had a look one the side of the qualities and possibly came up with the restriction that should added at least to material separation class:

obo:has_specified_output some (snap:material_entity and (ro:has_part some (snap:material_entity and (roproposed:bearer_of some 'PATO:biomaterial purity'))))

or as an alternative (even though 'PATO increased concentration' could be seen as a contentious point as it is about biological phenotype not planned processes)

obo:has_specified_output some (snap:material_entity and (ro:has_part some (snap:material_entity and (roproposed:bearer_of some 'increased concentration(PATO_0001162)))) )

Hermit reasoner classifies correctly when setting this find of restrictions.

Then, I went over the "material component separation" class (OBI_0600014) and its inferred subclasses in order to check which children would satisfy this condition (or a similar one conveying the same idea). I noticed that class 'material portioning' was classified as a child of it and this is *incorrect* and these should be disjoint. The textual definitions are in conflict but the formal definitions do not allow discrimination. The main reason 'material portioning' get classified the way it is has to do with the difference between objectives and potentially associated classes, which again do not reflect the qualities or properties of the output materials in the formal definitions.

Question: Is making "purification" a synonym of 'material_separation' acceptable? Which are those 'material component separations' which are not purifications? Currently the only i could find was 'material portioning' but again I feel there is a difference between 'making portions of material' and 'making separation of material component' (which implicitly meant separation of material constituants) and this is already acknowledged in the textual definitions.

Again, not sure if the distinction is needed unless someone can formally define the distinction between separation and purification. Is it that purification should be restricted only to a specific entity (e.g purification of protein kinase C isoform a vs separation of protein fraction) so the restriction could look like "has_grain only entity X" but this would be hard to satisfy as nothing is ever completely pure.

Reported by: proccaserra

Original Ticket: obi/obi-terms/608

[obi-bot]

done during 20120227 call. Philippe

Original comment by: proccaserra

[obi-bot]

• status: open --> closed

Original comment by: proccaserra

[obi-bot]

• status: closed --> open

Original comment by: proccaserra

[obi-bot]

re-opened following Matthew Brush's request

Original comment by: proccaserra

[obi-bot]

Thanks Phillipe. Pardon the following long posts, but I had some thoughts on this issue after it was discussed on the Feb 27 call.

First - I with regard to Philippe's comment that a 'material portioning' should not be a type of 'material component separation' in the second half of his post: To me, OBI has it right in dividing 'material component separations' into those whose outputs are of similar composition to each other (these have 'portioning' objectives), and those whose outputs are not (these have 'separation into different composition' objectives). Perhaps one way to make it more clear that 'material portioning' is included under the scope of 'material component separations', would be to clarify the wording on the 'material component separation' textual definition (currently reads "a material processing in which components of an input material become segregated in space), to read more like the definition for the 'material separation objective' class, which seems more general and clear to me (". . . transform a material entity into spatially separated components"). Then the distinctions outlined above between portionings and purifications as children of component separations should be more apparent.

Second, I want to be sure that the solution implemented during Feb 27 call, which involved creating a 'purified material', satisfied Phillips concerns. It was my understanding that the point of creating a 'purified material' class was to use it in a logical definition of 'purification', so that we more formally capture the distinction from a 'material component separation'. However, I have updated my OBI SVN and do not see any such axiom on the 'purification' class. 'Purified material' itself, however, is logically defined as being the specified output of a 'purification'. . . so is this sufficient to address Phillipe's concerns?

Original comment by: mbrush

[obi-bot]

Philippe, can you weigh in if you were satisfied with what was implemented, or if any further updates along the line of what Matthew wrote should be done? I thought we were done with this item, and I am not sure if/what specifically needs to be done now to close the ticket.

Original comment by: bpeters42

[obi-bot]

Hi Matthew. Thanks for the input as it sheds interesting light on the issue. One could be simply the ambiguity of the class name "material component separation", which to me necessarily entails that the process results in the constitutive parts of a given material entity being split into somehow atomic part i.e the components. To me this is incompatible with the notion of portioning, which entails that the composition is not changed (proportions of parts). We need to distinguish those material processing which leaving the material composition unchanged ( a portioning) from those which result in alteration of the composition. Currently material portioning is classified under material component separation based on formal definition which clashes with the class name. So this point needs addressing as it is in my eyes incorrect.

This leads to the second point. If we consider a true 'material component separation' which results in a output material with a different composition compared to the input material, then I feel that material purification is just a subtype at best or just an equivalent. What I somehow insisted on having captured in the formal definition of these processing were statements about the material properties (composition). It would also force to go beyond the mere duplication between processes and objective, which obviously in classification but can not detect ambiguities as the one we are discussing.

best

Philippe

Original comment by: proccaserra