 |
|
| FriendFeed is the easiest way to share online. Learn more » |
Anatomy workshop notes Snomed, Kent Spackman
steschu liked this
A little logic puzzle: Want reasoning over partonomy in regional subdivision of integument. Tried to do this with separate part retationships - as FMA does - using separate part relations (e.g.- consitutional_part). Formalization is very interesting.
- dosumis
I am not quite sure which are the benefits of distinguishing between regional part and constitutional part
- steschu
What about that species of fish that has an eye on each side of its head when it's young and as it grows older one of the eyes migrates across to the other side?
- Colin Batchelor
Also, to take a human example, I knew someone at university with situs inversus.
- Colin Batchelor
Kent's ontology is here: http://tinyurl.com/3n6nkz8
- dosumis
@steschu wrote: "I am not quite sure which are the benefits of distinguishing between regional part and constitutional part ". I suspect that some of these FMA part relations are doing useful work. For example, I think it may be useful to distinguish a tissue or organ from its innervation or blood supply. FMA uses systemic part for this. I think it's probably be better to use something like contained_in.
- dosumis
Developing an Application Ontology for Biomedical Resource Annotation and Retrieval: Challenges and Lessons Learned - Carlo Torniai
Meena Kharatmal liked this
context of eagle-i project. use ontology to control UI, display and search. very central role of ontology
- Melanie
layered architecture (application modules for specific app, eagle-i core ontology, mireot files).1. app module for UI: annotation file and anno def file. 'data mode exclude': classes to not display in UI for example
- Melanie
Interesting conclusion: need for community views - 'slims' (?)
- Kerstin Forsberg
from iPhone
Annotating Experimental Records Using Ontologies - Alexander García
research questions: lab notebook and ELN? semantic annotation of lab notebook?
- Melanie
Interesting overlap/synergy with what eagle-i and is doing form the perspective of resources. They should get in touch with biodata folks.
- Carlo Torniai
An Advanced Strategy for Integration of Biological Measurement Data - Hiroshi Masuya
as an expansion of PATO. Model of a datum as an informational entity. Stevens scales of measurements.
- Melanie
Hozo ontology editor - http://www.hozo.jp/
- Melanie
Allows to say "abnormally light in elephant' is larger than "abnormally heavy in ant" -> context
- Melanie
Where GO is Going and What it Means for Ontology Extension - Catia Pesquita
extension, refinement or enrichment
- Melanie
I am actually quite skeptical about her guidelines. A class with single subclass doesn't imply it shoud be merged. Nor does the fact of having many instances imply you should subclass. She cited http://dx.doi.org/10...
- Melanie
I think a class with a single subclass is an ontology smell. Not necessarily wrong, but probably a sign that something is up somewhere. Possibly with our present state of knowledge about the subject. (cf. http://en.wikipedia.org/wiki...)
- Colin Batchelor
Single subclasses can also simply be a reflection of the completeness of the ontology - we know other subclasses exist but there hasn't been a effort to add them yet.
- Jane Lomax
Revising the Cell Ontology - Terrence F. Meehan
complicated pipeline to convert OBO to OWL and mireot in terms then reconvert to OBO and filter out mireot. Seems it would be much more straightforward and robust to develop in OWL?
- Melanie
question about publishing unreasoned CL.owl version as well
- Melanie
Melanie: one step at a time..
- Chris Mungall
This is what is currently being used for the pipeline: http://code.google.com/p...
- Chris Mungall
Thanks for the link! Do you know if there is a strong desire to keep OBO, or if developers would be happy either way? Meant to ask Terry but didn't get the chance.
- Melanie
Dispositions and Processes in the Emotion Ontology - Janna Hastings
emotion occurrent: John is angry, doesn't have to know it
- Melanie
emotion disposition (to undergo emotion occurrent if right circumstances): John has been angry with Mary for 3 weeks
- Melanie
emotional personality trait: John is an angry person. Characteristic of a person involving a predisposition
- Melanie
MC: is 'cognitive representation' a term submission for BFO?
- Melanie
Not clear why appraisal process is a necessary part of emotion occurent
- steschu
I think a cognitive representation is just a sort of quality. There won't be a corresponding GDC because you can't transfer it.
- Colin Batchelor
Colin, not only qualities - perhaps processes - think dram
- Alan Ruttenberg
@Alan - yeah, meant to say "just things that are already in BFO like qualities" - didn't mean to exclude processes or dispositions!
- Colin Batchelor
OBO Foundry panel - BFO session - Barry Smith
BS: "if you have 3 apples in a box you do not have 4 first-class entities, 3 apples and the number 3" MC: I disagree, you have 3 apples and somebody counting them and saying "there are 3 of them". This is the principle behind having and assay and measurement datum
- Melanie
I agree from the perspective of the assay design. from a general point of you 3 apples in a box are entities in reality that 'exist' even if you are not measuring them.
- Carlo Torniai
Towards a Semantic Web Application: Ontology-Driven Ortholog Clustering Analysis - Yu Lin
in CAO (COG analysis ontology): new relation 'denoted_by' not inverse of iao:denotes. Unclear as why this is the case to me?
- Melanie
Supports ontoCOG, under the Ontobat umbrella (that's a lot of onto- ;))
- Melanie
denoted_by sounds dodgy to me---Cambridgy.
- Colin Batchelor
pointer to CAO?
- Alan Ruttenberg
Stefan asked the same question re denoted_by, and she was explaining (I hope I got it right) that she is trying to link a material entity to a GO ID, but as the same GO ID may denote several mat entities it didn't work.
- Melanie
Maybe confusion between class2class relations with instance2instance relations
- steschu
HeartCyc: A Cardiac Cycle Process Ontology Based in the Ontology of Physics for Biology - Daniel L. Cook
Ontology of Physics for Biology (OPB) http://sig.biostr.washington.edu/project...
- Kerstin Forsberg
from iPhone
dynamical properties lead to structural relation changes, e.g. valve in heart open or close is adjacent or not to other structure
- Melanie
Rapid Development of an Ontology of Coriell Cell Lines - Chao Pang (presented by James Malone)
"OBI is a high-quality artefact" :)
- Melanie
motivation: add Coriell cell line info in BioSample db at EBI
- Melanie
'cell line' is_model_for 'disease'
- Melanie
Coriell SQL dump -> excel -> mapping. Mapper tool at ebi.ac.uk/efo/tools based on Levenshtein distance +manual check. Using Rector's normalization (flat hierarchy, inference)
- Melanie
need money, people and tools to build very big ontologies
- Melanie
Automating generation of textual class definitions from OWL to English Robert Stevens, et. al. http://dx.crossref.org/10...
- Kerstin Forsberg
from iPhone
Talk: OWL-DL Ontology for Classification of Lipids: Hong-Sang Low
Meena Kharatmal, Frank and Melanie liked this
Semantic integration is necessary for lipid research, but no single unified, consistent and universally accepted lipid classification system is there.
- Robert Hoehndorf
Objective: formalize lipid nomenclature and classification using OWL-DL.
- Robert Hoehndorf
Based on BFO as top-level, but mainly reuse the Independent continuant category.
- Robert Hoehndorf
Chebi not suitable: not all relationships from ChEBI are relevant, lack of formal consistency
- Robert Hoehndorf
Compound has-part Organic_group AND Organic_group part-of Compound
- Robert Hoehndorf
34 additional properties of the has-part-X kind to specify number restrictions which do not work with transitive (has-part) relations
- Robert Hoehndorf
"Added additional properties to enable cardinality constraints" - sounds like not QCR aware - probably should be fixed for OWL 2 version to use fewer properties. Re transitive issue, they still don't need so many properties.
- Alan Ruttenberg
Ah - future works mentions QCR :)
- Alan Ruttenberg
So, the discussion is exactly about QCRs together with transitive properties.
- Robert Hoehndorf
Wondering if anyone has a good definition of "functional group"
- Alan Ruttenberg
@alan: As a very very rough first draft, functional groups are those special parts of molecules that ground (or maybe their structures, which are qualities, ground) dispositions. At least that's what I argued in my tutorial.
- Colin Batchelor
It would have to be some very special dispositions: the disposition to be attracted by gravity would not belong there. Do you have a way of saying which dispositions are grounded there?
- Robert Hoehndorf
@leechuck "some very special dispositions" - yes, those which are realized as processes of bond breaking or bond formation, roughly. (Sorry, missed this at the time.)
- Colin Batchelor
http://www.biomedcentral.com/1471-21... - Prototype Semantic Infrastructure for Automated Small Molecule Classification and Annotation in Lipidomics. Just came out
- Melanie
Attending the EHR tutorial by Werner Ceusters and Bill Hogan
Kerstin Forsberg and Mark Jensen liked this
Difference between term ('Beethoven'), concept ('the famous composer') and referent (the person himself)
- Melanie
concepts such as 'drapetomania' (http://en.wikipedia.org/wiki...) don't have a referent -> avoid using concepts
- Melanie
Distinction between data and what data are about, L1: first order reality (entities such as body), L2 representations (clinicians beliefs about L1) inaccessible (in our mind), L3 clinician explaining things to students is also representation
- Melanie
2nd part: problems with mainstream EHRs
- Melanie
WC: "better have a bad one that none at all"
- Melanie
diagnosis: L2 (in physician's head), disorder L1 in patient's baody
- Melanie
EHR models: patient centered, or encounter focused
- Melanie
Problem: in EHR, use of codes does not provide context. E.g. 3 records of fracture on one patient, are they the same?
- Melanie
MC note: Confusion between OBO Foundry and BioPortal? (WC showing OBOFoundry site and claiming back quality of resources, then switching to a screenshot of bioportal and saying same about Snomed and MedDRA)
- Melanie
BFO and RT as a unifying solution
- Melanie
Apparently (per WC and BS) http://bit.ly/ymBYE (guide to creating your first ontology) proves that France doesn't exist. Need to read and check on that.
- Melanie
Problem: we know some disorder is involved, not which one specifically. Goal of RT is to be explicit about what we are talking about: use UID (unique identifier) for each thing
- Melanie
Paper "how to deal with mistakes in the RT" (maybe that one? http://www.sciencedirect.com/science...)
- Melanie
Astonishingly, that guide does say, and I quote: "Therefore, we define all wine regions as classes. Protégé-2000 allows users to specify some classes as Abstract, signifying that the class cannot have any direct instances. In our case, all region classes are abstract." Seems a very odd way of doing parthood relations, realism or no realism.
- Colin Batchelor
IUI: instances unique IDs assigned to each occurrence of disorder -allows us to know that 2 diagnoses are referring to the same disorder (disorder being the L1 thing happening in reality)
- Melanie
RT represents each instance with timestamp in case of continuants using UIDs
- Melanie
Speaker switch: Bill taking the stand. Representing time in RT. Program for this session is at http://icbo.buffalo.edu/2011...
- Melanie
Limits of current handling in time in BFO: no such statement as temporal intervals (eg my life), no mechanism to handle open-ended intervals (my life is not over)
- Melanie
bfo:temporal_region, connected and scattered as subclasses. definitions of occurring_at, first_instant, last instant in http://genomebiology.com/2005... (Relations in biomedical ontologies) Expansion to temporal intervals explained in http://krr.meraka.org.za/~aow201...
- Melanie
TSMI: a CEN/TC251 standard for time specific problems in healthcare informatics and telematics http://www.sciencedirect.com/science...
- Melanie
Bill suggests enhanced temporal relation ontology: 'start_later_than' (subrelations such as during, co-ends, since...) and 'start_at_same_instant_ as etc
- Melanie
BS: will hand out proposal for BFO2.0 tomorrow. BFO not designed to be restricted list of preferred terms, but instead logically robust basis in terms of which you can define extensiosn (such as enhanced temporal region) -> no need for it to be part of BFO, can be an extension
- Melanie
Didn't all this get sorted out with the Allen paper donkeys' years ago?
- Colin Batchelor
Bill mentioned this was the basis for his work - I had missed the reference, thanks!
- Melanie
'IUI always" filler for occurrents to fulfill template requirements instead of creating a new one -> workaround to comply with system, i.e. extension to RT
- Melanie
Use Case 1: Diagnoses and ICD-9-CM (BH)
- Melanie
No time left for remaining use cases - Werner will post his and Bill's slides on his website - I quote "google Ceusters" :)
- Melanie
Promise to answer emails addressed to him regarding adverse event paper
- Melanie
Illustration of Referent Tracking (RT) in #EHR demographic data http://ow.ly/1v2owJ by William Hogan #ICBO #ICBO2011
- Kerstin Forsberg
from iPhone
Anatomy workshop notes Plant ontology - Laurel Cooper
Anatomy workshop notes. Wasila Dahdul - phenoscape
Highlight of needs for synchronization across different anatomy ontologies and tools to do that
- Carlo Torniai
synchronization requirements: Problem, different ontologies use different strategies for referencing terms from other ontologies. Some use xrefs. Some import/Mireot. Need to standardise this better and make better tools
- dosumis
Anatomy workshop notes FMA - Alan Ruttenberg, taming the beast.
Melanie and Mark Jensen liked this
Problems - artifacts from frames cause problems for translating into OWL. Hidden assuptions.
- dosumis
currently focused on getting part_of correct.
- dosumis
Also big problems that part relations are missing in current OWL translations (e.g.- on bioportal) - this is an artifact coming from translating from frames.
- dosumis
Other ways to read this feed:
