- 1 EBI/MGI/TAIR Conference call 10-18-06
- 2 To-do list
- 3 EBI/MGI Conference call 10-20-06
- 4 EBI/MGI Conference call 10-23-06
- 5 EBI/MGI Conference call 10-25-06
- 6 EBI/MGI Conference call 10-26-06
- 7 EBI/MGI/TAIR Conference call 10-26-06
- 8 EBI/MGI Conference call 10-27-06
- 9 EBI/MGI Conference call 10-30-06
- 10 EBI/MGI Conference call 10/31/06
- 11 Berkeley/Editors/MGI Conference call 11-08-06
- 12 EBI/MGI/TAIR Conference call 11-15-06
- 13 EBI/MGI/TAIR Conference call 12-7-06
EBI/MGI/TAIR Conference call 10-18-06
Jane, Jen, David, Tanya (WebEx + Skype) (1 hr working time)
Tested Skype high-speed conferencing. Not good.
Jen/Jane have found homes for about 200 orphan terms. (about 6 hrs working time)
Jane has list of outstanding issues.
We worked through the first part of this list together. Discussions below and to-do items added to list.
1. Big hairy issue of regulation - Does the regulation graph need to imitate the is_a graph in the actual graph?
Approach 1: recreate isa path. Means creating a regulates relationship.
Approach 2: ask Chris M. since it's still in the graph as regulation of xxx, may not
want to recapitulate it. Just have the top level terms.
All processes are in and of themselves processes and those processes regulate processes. I.e. phosphorylation regulates transcription. If possible, we should have the most specific terms possible, i.e. regulation of transcription by phosphorylation.
In new CNS there is an example of this and that would be??. In most cases, we don't have enough info to be this specific. Quite alot of places where we don't have enough info. Jane found loads of things that were just straight processes and not 'regulation of xxx' ... leads to true path problems please finish this thought.
Once we create 'regulates' relationship, define by saying it doesn't violate the true path. correct?
Needs to be all some, not some some relationship. (Jane, confirm and expand please.)
For now, made all regulation of xxx, direct children of regulation of biological process. In the round robin phase, make regulation of a multicellular organismal/cellular/metabolic process, regulation of a molecular function (maybe?) as high level terms.
2. Some CNS terms are still is_a incomplete.
22017: neuroblast division in the pallium
David: is_a stem cell division (anything that's a xxxblast is a stem cell) Future: rearrange things under stem cell
Jane: there are a bunch of neuro* (neuromast) terms
David: not terms from the CNS meeting. the xx cell differentiation terms are all children of cell differentiation.
3. When in doubt, looking at the term's definition may help determine what the (missing) is_a parent should be.
1. Sort out maturation and all children that end in maturation, whether they fit into 'developmental maturation' or 'cell maturation'.
2. Sort out 'cuticle tanning' and parentage.
3. The model definition for reproduction is in GO:0048860. Use this one to formulate all the child definitions.
5. Look at all of the terms that have an is_a relationship to developmental process and group them. Need to create appropriate grouping terms. Come up with Aristotelian definition for the grouping terms.
6. Define root and shoot system development.
7. Check FMA to see how organ and system are defined and related.
8. Ask Eurie to look at whether mating type differentiation child of sex differentiation or not.
9. Check if some single celled stuff falls under multicellular. Use GO term finder and yeast annotations?
10. Differentiate secretion by what is secreted. Leave localization terms as a mixture of cell and organismal processes. Amelia will sort out. Doug also has input. Current placement could be correct.
11. Split transport into cellular and organismal transport. The current transport term is effectively "cellular transport". We've talked about renaming it and adding "organismal transport". Also see SF 1548649 on secretion and transport.
12. Have Doug look at 'cell motility during locomotion' term, definition and placement in graph.
13. Sort out 'locomotion', 'movement', and 'motility'. A gnarly problem indeed.
14. Ask Harold about whether urea cycle metabolism is always organismal metabolism.
15. Check if cell growth has growth and cellular process as parents.
16. Ask Michelle Gwinn about pigmentation. Can we make pigmentation a multicellular organismal process?
COMMENT: (from Michelle via SF item 1562205) Many bacteria certainly produce pigments. There are tons and tons of pigmented bacteria out there in every color of the rainbow. An obvious example are the photosynthetic bacteria. Two examples where some kind of function has been associated with pigment are 1. in Staph the caratenoid pigments appear to play a role in virulence and 2. in Pseudomonase the pigment pyocyanin has a role in iron metabolism (which may also link to virulence). So pigmentation definately happens in proks. - MGG
17. The 'regulation of xxx' terms will need to be dealt with and cleaned up in the future.
18. Since we merged 'physiological process' and 'biological process', we should probably also merge 'regulation of biological process' and 'regulation of physiological process'.
19. Test obomerge (done - Midori)
20. Feature request for obomerge: warn (instead of clobber) if terms in the different changed files share an ID but have different term names, as a way of avoiding ID clashes.
21. Update go/numbers/GO.numbers file (Midori)
22. Assign a separate ID range for each content meeting (Midori; David as backup)
23. Make part_of children of cell cycle is_a cell cycle process (some done at meeting)
24. Add children to 'protein complex assembly', e.g. RNP assembly
25. Look at children of protein complex assembly to see which can go under cellular protein complex assembly
26. Create grouping terms under anatomical structure development to organize children
27. Ditto for cellular process, other high-level terms that wind up with lots of children
27a. Add 'generation of signal' (is_a cellular process; part_of signaling)
27b. Add 'cytokinetic process'
28. Delete 'physiology' from term names (done for some, but not all, sep. regulation terms)
29. Resolve 'GO:' and 'NOM:' IDs -- make sure 0022400 and 0022401 are merged
30. Remove any dodgy instances of '(de)adenylyl' -- may have to change some back to '(de)adenyl' (Midori)
31. Rename all X metabolism, X biosynthesis, and X catabolism terms to X metabolic process, X biosynthetic process, and X catabolic process, respectively (warn the world first!)
32. Sort out synapse-related processes (e.g. synaptic specificity question below); get Erika to help
33. Add synonyms containing 'assembly' to 'complex formation' terms
34. Improve defs for pattern formation terms (David)
35. Ask Erika to check the relationships and definition for determination of muscle attachment site (GO:0016204)
36. Add assembly and disassembly of [unclear whether we wanted 'macromolecular structure' or 'macromolecular complex']
36a. Add cellular structure assembly
37. Create cross-products between 'cellular structure disassembly' and cellular component terms [presumably can do same for cellular structure assembly ...]
37a. Make sure existing structure assembly & disassembly terms are consistent with cellular component ontology (term names and relationships)
38. Add 'extracellular matrix disassembly'
39. Check 'a process' vs. 'the process' in defs throughout; collection-of-processes term defs should have 'a process'
40. Check DNA topological change (GO:0006265) def and relationship to unwinding (Midori)
COMMENT: Based on ISBN 0935702490 Nucleic Acids: Structures, Properties, and Functions. VA Bloomfield, DM Crothers, I Tirocco Jr. Chapter 10: Supercoiled DNA
- (cites Cozzarelli et al; Cozzarelli is one of the Big Deals in DNA topology, so if the Bloomfield et al book reflects C et al even remotely accurately, it's good. An even better source would be DNA Topology and Its Biological Effects, NR Cozzarelli & JC Wang, Eds; CSHL Press 1990; ... if we can find it, but it's out of print)
- Anyway, as David recalled correctly,
Lk = Tw + Wr
- Of these, only the linking number is a topological property; twist and writhe are geometrical. DNA molecules with the same linking number are topologically equivalent, regardless of the twist and writhe values. So the mention of linking number in the GO:0006265 def is correct. I'm inclined to delete the part of the def that mentions helicases, because helicases don't change the linking number of a DNA molecule. A true change in topology requires breaking one or both strands, as topoisomerases do.
- DNA unwinding by helicases causes "local" changes in twist and writhe such that the linking number, and therefore the topology, of the entire molecule or segment remains unchanged; topoisomerases then act to relieve tension caused by the twist/writhe changes.
- The local effects on geometry are thus very much biologically relevant, though I'm not sure how we might represent them in the process ontology. I am very reluctant to do it by changing the definition of 'DNA topological change' in a way that makes it inaccurate.
- I'm also not much the wiser as to how to represent the relationship between unwinding (i.e., strand separation) and topological change. They are related, but one isn't a type or part of the other either way round. They're coupled in vivo because topology has to change to accommodate geometrical changes and allow unwinding.
- Relevant quotes from Chapter 10:
- "The linking number is a topological quantity, which remains unchanged so long as the DNA backbone remains continuous. In contrast, Tw and Wr are geometrical quantities, which may change as the shape of the DNA changes through bending, twisting, or kinking."
- "Molecules that differ only in a topological quantity such as Lk are called topoisomers."
- [ Topological change occurs during DNA replication ] "... to separate the two single strands of DNA from one another during semiconservative replication, Lk must be reduced to zero."
41. Ask Becky to check the relationships and definition for epidermal growth factor ligand processing (GO:0007174)
42. Carefully craft email on the subject of development vs. developmental process vs. multicellular organismal development; send to GO list (David)
43. When we add the 'regulates' relationship, run Obol to create 'regulation of process X regulates process X' relationships (Chris + curators)
44. add def of AP site to base-excision terms
45. define intussusceptive angiogenesis in 'regulation of cell adhesion during intussusceptive angiogenesis
46. Make standard definitions for 'xx cycle phase' and 'xx cycle process'. Add this information to the biological process standard definitions documentation.
47. Need to review menstruation during round robin phase. Are any of the children of menstrual cycle process also children of menstrual cycle phase? (David will do this.) luteolysis - YES growth ones possibly are phases - David will look up and make sure. Current placement is not wrong. Could be improved.
48. Ask Chandra what the best is_a parent for GO:0000745 'nuclear exchange during conjugation without cellular fusion' should be. Do we need a 'conjugative process'?
EBI/MGI Conference call 10-20-06
Jane, Jen, David (Skype) (1.5 hr working time)
- We need to add some intermediate terms under regulation of cell migration to describe cell guidance by e.g. neurotransmitters
- Ask Karen what splice site recognition terms are a type of - is it mediate by proteins or just RNA binding?
- Can we have a new term 'genome maintenance' under chromosome org and biogenesis? To house some of the DNA repair terms.
- Ask Eurie to check DNA repair terms
- Ask Midori to look at DNA unwinding terms (not topological? Do we need a generic DNA unwinding term?)
EBI/MGI Conference call 10-23-06
Jane, David (Skype) (1.0 hr working time)
New Rule: The maiteneance of a cell is_a homeostasis. Actually in almost all cases I can find, it refers to the maintenance of the tissue the cell is found in. For example, meristem, amniosera, muscle etc.
- Ask Harold for a definition of GO:0006490, oligosaccharide-lipid intrermediate assembly. Also ask advice on is_a parentage. (PENDING 10/23/2006)
- Check other cell maitenance type terms and see how they are currently arranged in the graph. In some cases, maintenance means keeping a cell from differentiating, such as "stem cell maintenance". In some cases it means control of cell proliferation. Perhaps we need to tweak the def of homeostasis.
- Ask Eurie about the 'nucelotide excision repair terms':GO:0000717, GO:0006295, GO:0006296, GO:0006293.
- Ask Karen about terms like 'branch site recognition'. Are they really processes?
EBI/MGI Conference call 10-25-06
Jane, David, Jen (Skype) (1.5 hr working time)
Made new term for organismal attachment. Needed a home for 'puparia attachment' and many of the PAMGO terms that describe attachment of an organism to a host will fit under this term as well.
Glycoprotein catabolism is broken down in two ways. One by what is being broken down and one by the actual chemical reaction.
- Ask Midori about prospores. What is the membrane and are the 2 prospore terms the same?
- Ask Michele if provirus excision is a DNA recombination. Ask her to look at all of the viral terms.
- Rearrange adhesion terms appropriately under the new term multicellular organism adhesion to substrate.
EBI/MGI Conference call 10-26-06
Jane, David, Jen (Skype) (3.5 hr working time)
Every time we use organismal, we need to be explicit about multicellular or single-celled organism
Made biological adhesion as direct child of biological process to get around the single-celled multicellular issues.
Merge restriction of R8 fate into R8 fate commitment
Rearrange ribosome biogenesis and assembly. Merge biogenesis into biogenesis and assembly. Enumerate all of the parts of this process and make the correct is_a parents for those parts. Harold will look at this during round robin.
- Is viral reproduction an interaction between organisms? (Jane)
- Define adhesion terms (Jen)
- Is hibernation and estivation sleep? The question we ask is might there be gene products that are involved in hibernation whose orthologs are involved in estivation or sleep. We thought they probably were.
- Make removal of nonhomologous ends a DNA endolucleolytic process. Check with Eurie about this.
- Make rhabdomere a type of organelle in cellular component
- Jennifer to check whether flower development should be under reproductive structure development.
- Check order of "biogenesis" and "assembly" in terms.
- Protein complex assembly and RNP assembly and biogenesis. How do they relate?
- GO:0030423 is this a complex assembly? Ask Midori about this
- Jen needs to change the IDs from NOM to GO.
EBI/MGI/TAIR Conference call 10-26-06
Jane, David, Jen, Tanya (Skype and WebEx) (0.5 hr working time)
Brought everyone to speed on where we were and how far we had to go. Tried WebEx again to share OBO-Edit. Merged 'sarcomere alignement' with 'sarcomere organization' and removed it as a child of 'muscle contraction'.
Decided that it helps a lot to be able to see the editing as it happens as a check of what is going on.
EBI/MGI Conference call 10-27-06
Jane, David, Jen (Skype) (2.5 hr working time)
- Look at the relationship between subcellular anatomical structures and cell organization and biogenesis children. Can we align better with cellular component. Is ...organization and biogenesis the development of a subcellular structure.
- Ask Val about barrier septum site selction mechanism. Is it an actin-based process.
- merge sex determination, female germ-line determination (undefined) and female germ-line sex determination (defined). Make male terms parellel.
- Do we need a term to describe how an organism interprets its genomic status.
- Ask Michael to look at the sex determination part of the graph.
- Clean up the protein to membrane docking terms to be sure they have the correct parent.
- Strand displacement, strand invasion. Do we need a DNA geometric change? (Midori: GO:0032392 'DNA geometric change' added rev. 5.22)
- Are synergid death and synergid degeneration the same thing? [Yes. They are the same. Aka synergid cell death.TZB 11.1.06]
EBI/MGI Conference call 10-30-06
Jane, David, Jen (Skype) (0.75 hr working time)
No to do list today, just continued with placing terms under respective is_a parents.
EBI/MGI Conference call 10/31/06
Jane, David, Jen (Skype) (1.0 hr working time)
- Added vesicle targeting to establishment of vesicle localization. Is there a better place for this?
- Do we need a term to describe membrane invagination/evagination?
Berkeley/Editors/MGI Conference call 11-08-06
Chris, David, Jane, Midori (Skype+VNC) (1.0 hr working time)
We discussed logical definitions (cross-products) for Regulation terms; see:
Chris gave a demo of using the oboedit reasoner operating on an enhanced version of the gene_ontology_edit.obo file. The idea is to show show how DAG placement of regulation terms can be automatically managed using oboedit; this can be seen as a prelude to cross products with external ontologies.
ACTION: We decided that it would be a good idea to do this as part of the IsaComplete process.
DONE: This has now been done, and this should be updated daily in:
this is the IsaComplete file automatically enhanced with logical definitions for all regulation terms. This is updated daily.
Follow up call?
EBI/MGI/TAIR Conference call 11-15-06
Jane, David, Jen, Tanya, Midori (Skype and WebEx) (1.0 hr working time)
Note: edit IsaComplete.obo until further notice.
We'll switch to editing IsaComplete-xp.obo soon, but must communicate with Chris first.
recap of Chris' IsaComplete-xp.obo
looked at some sample terms to make sure we understand what's going on
Ask Chris/John about redundant links in graph viewer -- due to is_a/part_of lines saying same as intersection_of lines?
bug report? graph viewer seems to be missing parents Jane to ask John what we should see in graph viewer; is OBO-Edit displaying things wrong?
Most cases with redundant relationships mean we'll move the term to a child.
Must have a meeting with Chris and John ASAP.
Here are some screenshots of what happened with our testing:
When we first clicked on a term with an implied link shown as a blue squiggle the graphview looked like this:
Then we moved the term so that the blue squiggle was real in the ontology and so that the higher up, now redundant, parentage was gone.
We thought that looked a bit odd so we saved and reloaded and then it looked like this:
EBI/MGI/TAIR Conference call 12-7-06
Jane, David, Jen, Tanya, Midori (Skype and Google Spreadsheet) (0.75 hr working time)
- Discussed progress on the to do list. Basically we decided that most of the issues with is_a complete were done and now we have started working on issues that exist in both the live ontology as well as in is_a complete.
ACTION: Do the merge and have the new graph go live before Christmas.
ACTION: Everyone is going to have a look at a node on the graph that is 'XXX process' and be sure it only has is_a children. In these cases the defintion should begin 'A YY process by which' or something very similar. Other terms that describe specific processes with parts will be called 'ZZZ' and can have both parts and subtypes.