Difference between revisions of "Seattle working sessions Sept. 8-9, 2006"

From GO Wiki
Jump to: navigation, search
(Rules we tried to adhere to)
Line 31: Line 31:
6. Sometimes a parent can have only an is_a relationship but its is_a children can have part of relationships to other parents.  
6. Sometimes a parent can have only an is_a relationship but its is_a children can have part of relationships to other parents.  
'''Example''':  sex determination,  mating type determination (part of '''?''') , female sex determination (part of multicellular organismal development).
'''Example''':  sex determination is_a children,  mating type determination (part of cell differentiation) , female sex determination (part of multicellular organismal development).
7. A term can be a part of organismal development and is_a cell development.  
7. A term can be a part of organismal development and is_a cell development.  

Revision as of 11:24, 3 November 2006

David Hill, Jane Lomax, Jen Clark, Tanya Berardini


1. Ensure that each term in the BP ontology has a path to the root that passes solely through is_a parent-child relationships.

2. Clarify the difference between BP terms that represent collections of processes and BP terms that represent a specific, entire process. Right now, there are terms of the form “xx process” that are either one or the other.

3. Refine definitions, term-term relationships as they are encountered.

4. Check for redundant relationships using OBO-Edit reasoner plug-in and resolve them.

5. Calculate number of is_a orphans before and after (should be zero) editing process.

Rules we tried to adhere to

1. Metabolic process is defined as a collection of processes. Therefore, is_a children of metabolic process can be specific metabolic processes.

2. Cell division terms should be children of the related cell proliferation terms.

Example: Endothelial cell division is part of endothelial cell proliferation.

COMMENT: Double-check this in light of the long discussions about cell division vs. cell proliferation that we've had in the past. X cell division part_of X cell proliferation is probably fine where X cell is some kind of cell in a multicellular organism tissue, but there has been resistance to making the relationship between the generic parents. See SF 1219649 and this thread in the email archive for many gory details. (MAH)

3. A term should not be an is_a child of two top level granularity terms but it can be an is_a of one and a part_of the other.

Example: A term can be an is_a cellular and part of organismal but not an is_a of both cellular and organismal.

4. Define each specific process as having a discrete beginning and end.

5. regulation of xx process and is_a vs. part_of relationship to xx process: It can be an is_a child if the parent is a collection of processes in which any instance of the regulation of a process is an instance of the parent. If the parent process is a specific process and an instance of the regulation is not a complete instance of the parent but rather contributes to an instance of the parent's whole then the regulation is a part_of that process. (Verbatim quote from DH.)

6. Sometimes a parent can have only an is_a relationship but its is_a children can have part of relationships to other parents.

Example: sex determination is_a children, mating type determination (part of cell differentiation) , female sex determination (part of multicellular organismal development).

7. A term can be a part of organismal development and is_a cell development.

Example: dopaminergic neuron development is_a cell development and part_of multicellular organismal development

8. For BP process terms that refers to a collection of processes, use this string “xxx biological process” or "xxx process"

Example: metabolism becomes ‘metabolic biological process’ and 'developmental process'. The distinction is still vague because all types represent collections, but the idea is that more generic processes will be distinguished in this way.

Potential top level terms under biological process

Molecular process: rejected because that's what the function ontology represents

Subcellular process: rejected because cellular is going to include the level of the cell and below

Cellular process: accepted (cellular includes single celled organisms)

Multi-cellular process: considering but transformed into organismal (multi-cellular organismal process

Tissue process: will probably included under multi-cellular organismal

Organ process: under development only

Organismal process: accepted but called multi-cellular organismal process (see above) , for our purposes, single celled organisms = cells. Single-celled organisms will be included in this group.This gets around the nasty confusion when we refer to organismal.

Multi-organism: accepted

Population: too far, someone should do it but not us (Jane in a former life)

Notes from the discussions

1. We did not create separate terms for 'development' and 'developmental process' because their definitions would essentially be the same. If we want to use development to refer to the development of XXX then we will create that specific XXX development term. We don't need a generic term for binning purposes only.

Example: Cell development is_a developmental process.

2. We need to be very clear about what's happening at the level of the cell and what is happening at the level of the tissue.

Examples: meristem vs. meristem cell

Cell migration vs. cell motility

Cell division vs. cell proliferation

3. Cell development and cell developmental process are different things.

4. Many terms that were is_a organismal development were made part_of organismal development and is_a developmental process. This is not a hard and fast rule because some of these children terms need to go to deeper terms than organismal development.

5. Problems arise from single cell organisms = organisms. Therefore, we’ve gone with single cell organisms = cell.

6. We don't want to have cell types of single celled organisms.

7. We touched areas pertaining to the following Sourceforge items:

   a. Hatching – can now make different types: blastocyst hatching, avian hatching (would be a behavior)
   b. somatic embryogenesis (fixed embryonic development definition)
   c. meristem/stem cell population maintenance

8. Cell division was not made an is_a child of cellular reproduction.

Proposed Ontology changes (some implemented in the working draft)

1. New terms added: (note: ids are temporary and will change upon final implementation)

GO:0048856 multicellular organismal process

GO:0048857 developmental process

GO:0048858 meristem cell maintenance

GO:0048859 multicellular organism reproduction

GO:0048860 cellular reproduction

GO:0048861 shoot system development

GO:0048862 root system development

GO:0048863 stem cell differentiation

GO:0048864 stem cell development

GO:0048865 stem cell fate commitment

GO:0048866 stem cell fate specification

GO:0048867 stem cell fate determination

GO:0048868 pollen tube development

GO:0048869 cellular developmental process

GO:0048870 cell motility involved in cell locomotion

GO:0048871 multicellular organismal homeostasis

GO:0048872 homeostasis of number of cells

GO:0048873 homeostasis of number of cells within a tissue

GO:0048874 homeostasis of number of cells in a free-living population

GO:0048875 chemical homeostasis within a tissue

GO:0048876 retinal chemical homeostasis

GO:0048877 homeostasis of number of retinal cells

GO:0048878 chemical homeostasis

2. Terms merged

merged organismal physiological process ; GO:0050874 into multicellular organismal process ; GO:0048856

merged reproductive physiological process ; GO:0050876 into reproduction ; GO:0000003

merged physiological process ; GO:0007582 into biological_process ; GO:0008150

merged cellular physiological process ; GO:0050875 into cellular process ; GO:0009987

3. First iteration of Homeostasis – renamed as homeostatic (biological) process

   --Homeostatic process  (related synonym:  homeostasis; set of homeostatic processes ??)
   ----[i]organismal homeostasis
   ----[i]cell homeostasis
   ----[i]cell number homeostasis
   ------[i]cell number homeostasis within a tissue
   ------[i]free living cell number homeostasis
   ----[i]tissue homeostasis
   ------[p]cell number homeostasis in a tissue
   ------[?]tissue metabolic homeostasis

4. ‘development of anatomical structure’

    --developmental process
    ----[i]development of anatomical structure (GO:new!)- apparently,  this term is still under discussion (Doug, Cindy Smith)
    ------[i] nerve development
    ------[i] reproductive structure development
    ------[i] tube development

5. May want to make a term 'cell division of a single celled organism' and make that a synonym of cellular reproduction.

6. May want 'multicellular organismal sex determination' and 'single cell sex determination' for grouping. Possibly also 'multicellular organismal sex development'.

7 . May need stem cell development. Look at all part of children and see if they can be grouped under a single is_a parent. –Done: stem cell differentiation and related children added. (see terms added above)

Open Questions

1. TO LIST: Should reproductive structure development be a part_of reproduction or not? If not, we should add onto the definition of reproduction that this does not include the development of the reproduction structures. Will reproduction start with embryogenesis and development of the reproductive structure? Does the reproductive process begin before or after the formation (maturation?) of the reproductive structures? (Similarly is development of the eye a part of visual perception?)

COMMENT: This is clearly madness. Reproductive structures are necessary for reproduction, and hence their (correct) development is necessary for reproduction. There is obviously a relationship but I think it's not "part_of". Would you say "sight" is part_of "reproduction" when the lights are on? OK, I am being silly - but what I am trying to say is that not every relationship can (or should) be captured with is_a and part_of. - BEN

COMMENT: I'm beginning to agree that these should be split even if a mutant phenotype can't necessarily distinguish between the two. Perhaps at some point the connection can be made via annotations.-DPH

2. Do we need regulatory process in addition to regulation?

COMMENT: I think regulation of biological process will cover these is_a paths. THe regulation terms make the graph complicated with respect to multiple is_a parentages. Perhaps the intorduction of the regulates relationship will help this. For now we didn't deal with regulation terms. -DPH

3 Do we need a higher level term: cell population process?

COMMENT: Currently terms like 'cell migration' and 'cell proliferation' refer to populations of cells. Currently cellular process can include processes that occur in more than one cell but at the cellular level, for example 'cell signaling'. I think we are o.k. because when we talk about these population issues, we are almost always talking about a homogeneous group of cells. -DPH

4. Tissue homeostasis relationship to organismal homeostasis: part of vs. is_a? Should it just be is_a homeostasis and its children should be children of organismal homeostasis?

COMMENT: I thought we had resolved this in the draft graph. Tissue homeostasis referred to the steady state level of cells in a tissue, but other homeostasis terms refered to chemical levels. I think I remember us splitting this out. -DPH

5. Should intracellular signaling cascade, or its parent signal transduction, be is_a cellular physiological process? See SF 1558514.