Re: BiologicalEntity - BiologicalType

Hi,


Yes, I'm afraid that Study is a core level for my use cases.

I've suggested ISA hierarchy as a reference because it's the result of many months of discussions and agreements into WP7, it's quite stablished and there are many documentation and tools related to it. But I'm not a biologist neither I work on a lab, so I'm afraid I am not very experienced in the subject ...


Greetings,

Carlos

________________________________
From: Leyla Garcia <ljgarcia@ebi.ac.uk>
Sent: 26 June 2017 13:15:05
To: Carlos Horro (EI); Olga Ximena Giraldo
Cc: S-A Sansone; Enckevort, DJ van (medgen); public-bioschemas@w3.org
Subject: Re: BiologicalEntity - BiologicalType

Hi Carlos, David,

>From Olga's email, lab protocols are fine without a 'study' property or type.
@Carlos, @David. Do you still need it for your specific cases? If not, then so far there would be no need to take this discussion further. If yes, I would propose Carlos to lead a group to work on it.

Regards,


On 26/06/2017 13:04, Carlos Horro (EI) wrote:

Hi all,


I would like to mention that I'm collaborating with Elixir WP7 (plants use case), and, in this use case, we also had to coordinate/map many different models (7 if I'm not wrong) among them.

A good compromise solution we reached was to use ISA hierarchy as a reference for all of us:

http://isa-tools.org/format/specification/

Specifications | ISA tools<http://isa-tools.org/format/specification/>
isa-tools.org
The ISA Abstract Model, originally developed as a tabular format (ISA-Tab) since 2007, has been developed with several international collaborators and in synergy with ...

I think its 4 main entities (Investigation, Study, Assay and Sample) can fit quite well with most necessities, of course being aware that in each case there will be unique problems to be solved individually (ie: distinct-level entities with same name, or same-level entities with different name, and others).


Greetings,

Carlos


________________________________
From: Olga Ximena Giraldo <oxgiraldo@gmail.com><mailto:oxgiraldo@gmail.com>
Sent: 26 June 2017 12:15:26
To: Leyla Garcia
Cc: S-A Sansone; Enckevort, DJ van (medgen); Carlos Horro (EI); public-bioschemas@w3.org<mailto:public-bioschemas@w3.org>
Subject: Re: BiologicalEntity - BiologicalType

Hi all.

I would stay away from the discussion investigation/study/assay and consequently protocol. I agree with Susana, "the unit and sub-units of work are clearly named differently in different community, or worse the same name has different meaning."  I would also stay away from the "aggregation" idea because it implies that you know the boundaries of whatever is being aggregated.

Extending or adapting a Web-scale vocabulary like schema.org<http://schema.org/> that shall be applicable for all kinds of research settings  and thousands of applications is quite a different thing than designing a conceptual data  model (ontology) in an academic committee-fashion.  We should be aware that every new element in schema.org<http://schema.org/> adds complexity and cost to the users of schema.org<http://schema.org/>, in terms of searching the documentation, maintaining examples and validators, etc. Therefore, bioschemas should be simple and stay away from pretending to be an ontology.

>From my research, looking into more than 1000 protocols I have seen that the information they all share is summarized in: sample, instrument, reagent and objective (SIRO). These elements are common across all protocols I have seen. For experimental protocols, investigation information is not so relevant. The elements that I consider important for protocols are the SIRO elements plus applicability and provenance.

Let me clarify something a protocol is not a study, strictly speaking. Actually, studies and investigations don't necessarily have a detailed  description of steps, as in a workflow of operations/actions, whereas protocols do describe step by step the how to do something.

Best,

Olga


On Mon, Jun 26, 2017 at 12:11 PM, Leyla Garcia <ljgarcia@ebi.ac.uk<mailto:ljgarcia@ebi.ac.uk>> wrote:
Hi Olga, Carlos, and David,

There is a MedicalStudy, we could define a BiologicalStudy that can aggregates studies. So, if you think an investigation is bigger than a study, you just use BiologicalStudy and link it to other 'sub-studies'. This could work for experiment (small study) and lab protocols (even smaller study). So, there are ways to specify this in a light way which is what is needed by now in Bioschemas.

Now, the question is, how useful a BiologicalStudy would be? Particularly for the phenotype and lab protocol case (and David's as will but I am afraid I do not know which entity he is working on). What are the advantages of having a link to the study in your cases? Is it findability, summarization, completeness? Or any other use case that you have identified?

Cheers,


On 26/06/2017 10:56, S-A Sansone wrote:
Dear All,

this discussion about Investigation, Study, Project, Experiment, Dataset is a long-standing one, unfortunately. I have witnesses this in the last 15 years working with many standards developing communities, and also during the MIBBI checklists harmonization project (some of you may remember). Unfortunately the unit and sub-units of work are clearly named differently in different community, or worse the same name has different meaning.

Can we go back to the original point: what is that we aim to 'discover' in this case?

Thanks,

Susanna


On 26/06/2017 06:24, Enckevort, DJ van (medgen) wrote:
Hi,
In MIABIS we also defined Study, which should also be reflected in OMIABIS or OBIB. Wouldn’t that be the more appropriate option?

With kind regards,

David van Enckevort

Op 23 jun. 2017, om 19:03 heeft ljgarcia <ljgarcia@ebi.ac.uk<mailto:ljgarcia@ebi.ac.uk>> het volgende geschreven:

hi Carlos,

It does not seem to be much about investigations or studies except for http://health-lifesci.schema.org/MedicalStudy.
Olga Giraldo is working on a schema for Lab Protocols which are experiments. Experiments can be part of studies, studies can be part of investigations. Or you can just see an investigation as an study (as I understand was Rodrigo's proposal).

I would suggest you to talk about it with Olga, maybe you can come up with something in that regard. Please keep me in the loop, I would like to participate in that too if we see that it fits and it works for Bioschemas.

@Alasdair, Rafael, Carole. If we decide to model these investigations/studies required by phenotypes, we might need an extension to deliver the specification as it would be something new. Also, Lab Protocols have reagents which are chemicals which, at the moment, fit into Biological Entity. I am not sure we are covering all the reagent needs from a Lab Protocol perspective so Biological Entity is likely to change. I know this is not ideal with the deadline on the 30th coming, but we just had a meeting with Olga to help her moving her specification to Bioschemas templates.

Regards,

On 2017-06-23 12:03, ljgarcia wrote:
Hi all,
I would propose the check schema.org<http://schema.org> to see whether there is something
there that can be used for study or investigation.
Regards,
On 2017-06-22 12:53, Lopez, Rodrigo wrote:
Can I propose the term 'study' to replace 'investigation'? But still,
'study' is not a biological type but rather a 'collection' or
'aggregation' of biological types?
Kind regards,
R:)
On 22/06/2017 12:41, ljgarcia wrote:
Hi,
I would say "investigation" is not a biological type. You have there, for instance, people participating in the investigation, starting day, grant, and so on that just does not fit withing BiologicalEntity. Is there something more appropriate in schema.org<http://schema.org>?
Regards,
On 2017-06-21 12:00, Carlos Horro (EI) wrote:
Hi,
We have defined BiologicalEntity - Phenotypes use cases, and we would
be interested on one case about searching for an investigation (ie. by
its name or description) and obtaining information about it ,
organisations and others. For mapping this case to BioSchemas,
BiologicalType would have to support something like 'Investigation',
which it's not currently supported... would it be OK? do I include it
into the biologicalType description?
I think the question it would be similar with other use cases we need,
as Trials, Cultivars or Traits...
Greetings,
Carlos
--
============================================================
Rodrigo Lopez Serrano,
Head of Web Production,
European Bioinformatics Institute (EMBL-EBI),
European Molecular Biology Laboratory,
Wellcome Trust Genome Campus,
South Building,
Hinxton, Cambridge, CB10 1SD
United Kingdom
------------------------------------------------------------
ORCID: 0000-0003-1256-7306
http://scholar.google.co.uk/citations?user=7fhGnVEAAAAJ&hl=en
============================================================
Love data? You can now search over 1 billion biological data
records in one go using EBI Search at https://www.ebi.ac.uk
============================================================





--
Olga Ximena Giraldo Pasmin


PhD Student
Ontology Engineering Group
Escuela Técnica Superior de Ingenieros Informáticos
Universidad Politécnica de Madrid
Campus de Montegancedo, sn
Boadilla del Monte, 28660, Spain

Orcid ID: orcid.org/0000-0003-2978-8922<http://orcid.org/0000-0003-2978-8922>

Twiter: @olgaxgiraldo

Skype:olgaximenagiraldo

Website: http://oxgiraldo.wordpress.com