Conference calls Biomaterial 2008
Contents |
Biomaterial Conference Calls
December 5th 2008
I/ creation of defined class organization bearing role regulatory role as regulatory agency regulatory agency: The agency or government department that has responsibility over the legislation (acts and regulations) for a given sector of the government. source: biobasics.gc.ca/english/View.asp
II/ artifact object: the biom branch will update restrictions and revisit
III/ Identify 10 "core" terms
- material entity
- organism
- organization
- chemical entities in solution (mixtures)
- environmental matter
- macroscopic part of organism
- molecular entities
- artifact object
- EOO
note: protein shouldn't be outside molecular entities we need a round of clean up, including metadata
agreement on having roles reagent and specimen role on a case by case basis (e.g. Alan's cat not specimen)
Review of those terms 1. organism - Homo sapiens alternative term natural person, alt term human person is subject to confusion, e.g. as a legal term "This is a common offshore legal term meaning a living breathing person as opposed to an unnatural person such as a corporation", thus we chose to have natural person as alternative term - organism should include dead ones as well
2. organization
We added an instance organization Affymetrix to serve as model/template for other branches. For example the instrument branch needs to represent manufacturers. Instances will be added in a separate file - Affymetrix has been added in the Biomaterial branch file just as an example how we proceed. Role branch will add a manufacturer role Relation branch: there is a relation is_manufactured_by, do we need the inverse is_manufacturer_of? do we need a more generic one is_supplier and then use roles?
November 7th 2008
Present: JF,BP,PRS.
-i. concern over 'children under molecular entities / organism part
We came to the conclusion that this is essentially a tool related issue. OBI Biomaterial need to document the use of those classes and make it clear that OBI relies on external sources, there OBI users should not expect to find all terms in OBI.
-ii. generic terms and species specific terms:
We will stick to getting terms from FMA for the moment. Again, we need to make clear in the documentation that we are aware of this . Solving this is outside OBI scope.
-iii / iv/ branch reorganisation: Main Goal = make sure that 'entity of organismal origin' is classified as a child of a'artefact object' Secondary Goal = work on setting restrictions on potential 'entity of organism origin' so that they are picked up bu the classifer
MG: entity of organismal origin is now a defined class with the following restrictions: -derives from some (organism or 'organism part' or cell*) (cell needs to be added, delayed due to pb with protege hanging during classification) -is_ouput_of some ('material transformation') (used to be 'protocol application')
SG: bronchio alveoloar lavage no longer asserted under EOO restrictions set as derives from some (organism part) (should have been lung but did not have time to import) is_output_of some lavage
Question:
Asserting under defined class or leaving the classifier to do the job ? To be further discussed with all.
October 17th 2008
I/ status of relations
OBO rel stalled at implementation 1. class relations and instance level relations -> diff namespace 2. move bfo specific relations into bfo proper 3. third class of relations that are non primitive = cross products relations 4. reflexivity of part of: reconsider? 5. set of proposed relations
example of current issues:
- results_from between quality and process
- process located_in independent continuant seems not valid
Proposed: relation meeting addressing outstanding issues
- 2 steps process
- OBI review
set up OBI needed relations, test them.
- in coordination with Chris Mungall, RO
move them to OBO relations
AI: MC contact liju for relation call, make list of times and add to dev
II/ Review of Alan's proposal for concentration
note: we agreed on trying to use this proposal, as it is equivalent to Bjoern's idea but more explicit for example for units - do we want to have monadic and relational qualities (could be imported from pato) - roles: solute and solvent = issues when not clear which is what eg ethanol/water - independant continuant: portion + molecular entities molecular entites are "items" portion of liquid is portion of substance which has part only h20 -defined classes
- issue where to put units? on quality side we standardize with a single unit, we may have tool converting from annotation to our standard (note: no unit conversion/comparison possible, not even with owl2) tool = little language -> annotation property -> script translating into standard set of units (same model than for alt term tag (might be problematic as need to know masses for all components though) AR chose molarity has example here
- current status decision to have tris buffer now and maybe later if needed tris solution query: what is tris made of? no good answer yet -> classification ok, but not easy to get at what components are (eg.: what is the solute in tris buffer would be ugly in sparql) option: add other property to get those parts? caveats: all things will have ranges of concentration (owl2 feature) open question about scripting issue how we aks for components and concentrations
AI: AR will put concentration example into obi file
AI: BP will provide a short textual description
October 3rd 2008
focus on representing solutions.
Proposed working example: Tris solution has components tris and water. We will define a tris solution, then assign concentration to tris and water, and finally say that the tris solution is bearing the role buffer when the concentrations are in an acceptable range.
September 26th 2008
1. protein complexes
PRO doesn't cover complexes, but GO does.
protein complex mireot-ed from GO http://amigo.geneontology.org/cgi-bin/amigo/term-details.cgi?term=GO:0043234
definition: Any macromolecular complex composed of two or more polypeptide subunits, which may or may not be identical. Protein complexes may have other associated non-protein prosthetic groups, such as nucleotides, metal ions or carbohydrate groups.
2. mixture
more precisely: solutions, gas, alloy, emulsion, suspension
solution can be split into multiple portions that are also solutions with the same components: Portion of molecular matter - ensemble of molecules
has_granular_part relation:
ex portion of water has_granular_part H2O molecule (and not water has_granular_part water) => has_granular_part: idea of composition, there also need to be a large number of H2O molecule, idea of identity to the whole is necessary but not sufficient of has_granular_part
Definition: Granular parts are the constitutent parts of Pluralities. They are of the same sort. Generally pluralities have a high number of grains. The loss of a grain has no effect on the sortality of the identity of the plurality
Example tris solution: has_granular_part H2O, has_granular_par tris: we import H2O and tris from chebi and create tris solution in OBI
Problem: how do we deal with concentration?
We started discussing this, but didn't conclude. The discussion is to be resumed during next biomaterial call.
Few initial notes:
AR: initial attempt https://wiki.cbil.upenn.edu/obiwiki/images/5/57/Concentration.pdf - some issues, like for example if no zinc in sample what does the concentration quality inheres in?
BP: it would be nice to have an easy construct like Material has_granular_part (H20, 25%)
BP: Tris buffer has_granular part H20
Tris buffer has granulart part Tris
Tris buffer has component (all H20 molecules in buffer) (where all H2O molecules in buffer = scattered portion of water)
scattered poriton of water has_concentration (25% by mass)
AR: from tris has_quality concentration_of_water (value 25%), one can imply that tris has_granular_part molecule of H2O
====>>>> we would like to end up specifying element and concentration and get the rest
(molecule and is_granular_part_of solution) has_concentration X
for example: (H2O and is_granular part of tris buffer) has concentration 25%
Note: we probably need to define a range of concentration, 1-25% to have for example the role buffer. so we need to define the tris solution based on its granular parts, and the buffer role that it plays if the components are in an acceptable range of concentration.
September 19th 2008
- mireot
-- AR will try and improve scripts
Sometimes neurocommons is notup-to-date with latest ontologies versions, and causes extra information not to be loaded properly, requiring manual checks
-- JM will check on status with protege people, to be done during OWLED workshop: we will present the mireot mechanism and this will probably foster interactions with developers present
- mixture
-- mixture are composed of different elements, will have properties as well.
-- concentration will be defined or not, people will have choice to specify them or not
-- start with material entity and has_part A, has_part B...
future possible: if part B in such proportion has properties 'buffer"
AI for PRS: liaise with CHEBI people, check if this seems like an acceptable solution and if they can work out a template together.
- function branch?
-- questions about how to edit that branch: who is responsible?
e.g., obi_function term seems to be a placeholder and we might consider deleting it.
- JM tiny clean-up of branch, on AR's latest email model
-- to be examined during next call
- next steps for the biomaterial branch?
-- PRS: digested DNA -> connect adjectives to the PA that cause them to appear
2 options:
1. biomaterial has_quality irradiated and is_output_from PA irradiation
2. irradiated results_from PA irradiation
"results_from" is a new relation that would need to be created, and would allow linking directly a protocol to its effects (e.g. irradiation PA -> irradiated) without need for a bearer (version 2)
version 1 is currently possible, as all required elements are in the ontology, and in this case we would have a link between the bearer and the "historical_role"
PRS will start going through to the file, implementing solution 1.
PRS will drop line to Chris Mungall for relation "results_from" and inclusion in ro.
- defined classes
currently in buckets at top (_defined protocol application, _defined_material)
proposal from AR to have mini buckets positioned as needed by editor (e.g. I need a biomaterial defined class, I add a placeholder "_defined_classes" when editing the file, and this get removed by script when release
divergent opinions on this, MC fears people will add mini-buckets with various syntaxes, and current solution seems satisfactory.
- by next release (october), replace DENRIE with IAO in OBI
September 12th 2008
Proposed changes from AR:
1. peptide, currently OBI_0100083 -> CHEBI:16670 agreed.
2. enzyme = defined class (protein or has_part some protein) and has_function some GO:0003824 (catalytic activity)
Note from BP that there are RNA enzymes as well. Agreed on: (protein or rna) or has_part (protein or rna) and has_function some GO:0003824 (catalytic activity)
We will import molecular function with parent function We will import GO:0003824 with parent molecular function
We will also import CHEBI:23367 molecular entities, with parent material entity.
Amino acid CHEBI:33709 will be imported as well and will be a child of molecular entities (we will fix the label to be singular in our import, Alan to fix code to respect this. Original one will be kept as alternative term. For terms where we have obi:preferred term we will store that in external.owl and this will take precedence over the rest)
3. DNA Polymerase = defined class enzyme and has_function some GO:0034061 (DNA polymerase activity) 4. Reverse transcriptase = defined class enzyme and has_function some GO:0003964 (RNA-directed DNA polymerase activity)
GO:0003824 catalytic activity to be imported, which child GO:0034061 DNA polymerase activity
AI for BP: check on pro and definition or protein, does it cover complexes?
5. recombinant plasmid = plasmid and output of some <appropriate material
transformation>
recombinant plasmid is already output of recombinant plasmid construction process. We need the term "plasmid" to move forward.
We will create our own defined class for plasmid: plasmid = DNA and has_quality circular and has_function (is_realized_as some gene expression) GO:0010467 circular is to be imported from SO:0001235: "A region containing at least one unique origin of replication and a unique termination site."
6. environmental matter (formerly known as environmental sample) rename "environmental matter" and move rest to editor note. agreed.
7. anatomical entity currently OBI_0100015 -> CARO:0000000 we decided against this import: current definition in CARO reads:"Biological entity that is either an individual member of a biological species or constitutes the structural organization of an individual member of a biological species" This would cover individual organisms as well and is not desired. This class will be renamed to "organism part" with definition "An organism part is a part of organism's anatomy or substance arising from an organism from which the biomaterial was derived, excludes cells. E.g. tissue, organ, system, sperm, blood or body location (arm)."
8. cell compartment currently OBI_0100054 ->? membrane-bounded organelle GO:0043227 agreed. Current parent cell component which was imported from CARO will be replaced by GO import of GO:0005575
AR will experiment with declaring the CARO and GO classes as equivalent.
Postponed:
- mireot system: how to make it easier for people to use?
- mixtures (cf email from Bjoern: http://groups.google.com/group/obi-biomaterial/browse_thread/thread/eb81a4e80556f8e4?hl=en)
