Signaling Meeting Minutes April 2008 -
- 1 March 2008
- 2 30th April 2008
- 3 2nd May 2008
- 4 23rd June
- 5 30th June
- 6 21st July
- 7 8th August 2008
- 8 20th August
- 9 22nd September
- 10 October 2008
- 11 10th November 2008
- 12 14th November
- 13 24th November
- 14 25th november 2008
- 15 19th March 2009
- 16 20th March 2009
- 17 23rd March 2009
- 18 Comments from Alex
- 19 Comments from Sandra
- 20 Discussion on list
- 21 6th April 2009
- 22 4th June 2009
- 23 29th June 2009
- 24 Next step:
Participants: Jennifer Deegan and David Hill
David and Jen met to review Jen's initial structure. David suggested that the general grouping terms could be done along the lines of pre-existing terms in GO. Chemical properties might follow Chebi and location and distance might follow the development model.
Action: Jen will work on this before next meeting.
30th April 2008
Jen and David have agreed to have a 1 hour meeting once a week to get David's input into what Jen has done during the week. Jen will carry out action items in between. One important part is to align this with the regulates work.
One first step is to make a list of the signaling pathways already in the ontology and what they regulate.
Action: Jen is to make a table of signaling pathways and the things they regulate and put them in a table.
2nd May 2008
Participants: Jennifer Deegan, David Hill.
At our previous meeting a couple of weeks ago we realized that we had a big problem with capturing signaling where a signal goes through the blood stream and reaches everywhere in the organism but is not acted upon by all parts of the organism. Physiologists call this long-distance signaling and Jen was keen to capture the fact that the signal goes to all parts of the organism (organismal signaling), whilst David was keen to stick to capturing where the signal comes from and which cells act on the signal (tissue-tissue signaling). Jen discussed this with Tim Deegan who explained the similarities between the signaling in organisms and signaling in computer networks. In computer networking they already have terminology for this, and they are able to describe signals based on both the areas that a signal reaches and the parts of the system that react to the signal. A 'broadcast signal' is one that reaches all points in the system, even though only a selection of parts respond to the signal. A 'directed signal' is one that is sent from one specific place to another, and then there are also 'unicast' and 'multicast' signals.
Jen and David have discussed these ideas and Jen is going to look into making a structure to capture this kind of information, without some more biological words (perhaps systemic signaling?). Jen is taking care of signaling, which is the travel of a signal to the location of action and David is going to deal with signal transduction, which is what cell biologists call the thing that happens when a signal arrives at the cell where it will have its effect. Jen is to look into including all the different kinds of signaling, including light, chemical, mechanical, sound and so on. If she can get the structure down to the level of detail where it will meet up with the signal transductions lever e.g. of retinoid signaling then that will be good as it will set the scene for David to pick up later.
Participants: David Hill, Jennifer Deegan.
We discussed the new signaling terms that had been made. We made some rearrangements.
- Behavioral signaling is now to be included under the top signaling term.
- The top signaling term is to include all parts of signaling from inception to the end of signal transduction (which is the process whereby the signal is received and acted upon.)
- We need to be careful of terms that cover only the transport of a signal and that will not involved any gene products at all.
- We decided to keep the term 'system signaling' as this will involved gene products in ong range nervous system signaling, though it may not in hormone signaling.
- In old terms that say 'X-mediated signaling pathway' the term means the signal transduction part of that where the signal is received and acted upon.
Participants: David Hill, Jennifer Deegan.
Worked on top nodes of signaling.
Action item: David will define what he means by 'at the level of'.
Meeting with David to define the top terms under signaling.
David has defined 'at the level of' such that 'processes occurring at the level of the heart' means 'processes carried out by the heart.
8th August 2008
Some notes for discussion.
Would this be a good standard def for the signal transduction terms?
The series of molecular signals initiated upon sensing of [signal].
Some of the signal transduction terms are oddly defined and need a standard def.
I have put the hormone ones under the new general hormone term.
Participants: Jennifer Deegan, David Hill.
We have checked the definitions and positions of the top signaling terms and they are fine. Now Jennifer will look for all the child terms that will go under these and next meeting we can work out exactly how they should be related. Particular care will be taken over those parts of the graph mentioned in the many open sourceforge items in this area.
Now all terms including the words 'hormone' and 'signal' are under 'systemic hormone signaling'. Some other hormone terms without signal in the text have been put under a temporary grouping term that is under 'systemic hormone signaling'.
Participants: David Hill, Jennifer Deegan.
Lots of 'response to hormone' terms have been moved under 'system hormone signaling'. We started by looking at those. We had a good discussion about standardization of these terms. There are some problems with the 'response to X' and 'detection of X' terms that are quite fundamental and need to be worked out before we can proceed.
We started with this:
[i]systemic hormone signaling ---[p]response to hormone stimulus ------[i]detection of hormone stimulus ------[i]response to auxin stimulus
1) The signal is sent. 2) The signal travels to its destination. 3) The signal arrives at its destination in such a way that the receiving organism or cell is now aware of the presence of the signal. This may mean that a signal protein has bound to a receptor, or an electromagnetic (light) wave has been absorbed by a photoreceptor, or there may be some other mechanism. 4) The point of first contact usually sends a further message via some other mechanism, down a chain to the place or process where the signal will be acted upon. 5) The signal message arrives at its final destination, which is the place where changes need to be made to cause an appropriate response in the organism. 6) The organism responds.
A real example
Light approaching a plant shoot from the side causes bending.
1) Light exits the sun or a light bulb. 2) The light travels through space and hits one side of a plant. 3) The light is absorbed by a photoreceptor at the tip of the plant, on only one side of the plant. 4) A message is sent from the photoreceptor via auxin to a place a centimetre down from the tip of the shoot, on the other side of the shoot from where the light was absorbed. 5) The auxin signal arrives at its destination and is registered in the region that will respond. 6) Cells on the side of the shoot away from the light that have received the auxin signal elongate to make the shoot bend toward the light.
Some language questions
The question that we were considering was to do with 'response to signal', 'detection of signal', and 'reception of signal'.
Q/ Does response to signal only include step 6 or does it include everything between step 3 and 6?
Q/ Is detection of the signal a type of or part of the response, or even something separate? In other areas of the graph detection has been made a type of response, in that steps 3 and 4 together can be considered to be a response in their own right.
Q/ Does step 3 constitute the entirety of detection, or does detection require steps 3 and 4.
Should we say 'response to signal?'
We got rid of the term 'response to signal'. (It has never been live.)
Example - ethylene and ferns
Fern gametophytes grow taller and thinner when they are planted closer together. They do not detect proximity by the amount of shade falling on them, as they are not so close together as to be shaded. Instead they are thought to detect it by the concentration of ethylene in the surrounding environment. Ethylene is produced by all the individuals and so builds up more in areas of high population density.
To describe this in GO we could make the following terms:
[i]response to population density ---[p]hormone signaling involved in population density ------[p]ethylene signal transmission involved in response to population density ------[p]ethylene signal reception involved in response to population density
[i]hormone signaling ---[p]hormone signaling involved in population density
Things to note here include the idea that the elongation of the ferns is not really a response to ethylene, but a response to population density. So we should not have 'response to hormone' or 'response to signal' terms. Instead we just have 'response to population density'.
As stop-gap we have decided to rename the 'response to signal' terms to 'reception of a signal' where 'reception' is intended to mean the sum of 'detection' and 'reaction'. In other words it represents everything between steps 3 and 6 in the first example. We may need to come back and change the word 'reception' but the new def is fine so we will make the change and retrofit the name.
Action item: Implement the new reception name and standard def on the signaling terms.
New standard definition for 'reception of a signal' terms:
Def: The process whereby the cell or organism changes its state or activity as a result of a signal. A signal is an entity that conveys information.
Action item: Put in the second sentence for the 'reception of hormone' defs. 'Abcissic acid is a hormone that does this and has that structure.'
We also need to make slides for the consortium meeting to show the decisions we have made about the top signaling terms, and the questions that we have raised.
Prepared the work we have done to be presented at the GO Consortium Meeting for approval to implement. Conclusions
10th November 2008
Participants: Jennifer Deegan, David Hill.
Started to apply aristotelian definitions to is_a children of signal transduction.
to do items:
x signaling pathway both spellings
x signal transduction both spelling
ask tair about GO:0010019
OBO-Edit bug: cannot supresses warnings of newlines. Popup window is in the way of seeing the text we are trying to edit.
space between title and def is getting bigger and bigger but later goes away.
add synonyms with 'mediated'.
look back at GO:0009401
fix 'effecting' so it says 'effects'.
check defs of signal transduction terms all the way down.
check terms to see that they have 'response to' parents
Resolved more defs.
GO:0010017 add def for red and far red light.
I have to look into the TOR signaling pathway and find out if it it beginning, end or en route.
David to look into stress activated kinase stuff.
Fixed the term names to be signal transduction with synonym of pathway. Fixed effects/effecting. This is for terms right under signal transduction.
25th november 2008
Email from Alex Diehl.
1) I have particular problems with the definition of "creation of signal" which also bring up some concerns about the general nature of signaling. The definition from the PowerPoint slide reads: The process whereby an entity that conveys information is produced. Following creation, the entity will be transmitted, received, and interpreted by a cell or a multicellular organism. The first sentence is fine. However the second sentence is not necessarily true and should be omitted entirely from the definition. Cells produce signals without regard to whether those signals will be ever be received and interpreted. Signals in many biological contexts are just molecules released from one cell in response to external stimuli or internal changes. The potential targets of such signals are not guaranteed to be nearby or available or receptive to such signals. Saying the signaling entity "will be transmitted, received, and interpreted" simply cannot be assured or assumed in all cases. Of course, we can argue that a signal is not a signal unless it is actually received, but I think that the core idea of a signal is the "intention" to transmit a message to any potential receivers, not the necessary reception of the message. 2) The definition of "biological signaling" provided on the PowerPoint presentation, Any biological process in which a signal is generated, transmitted, received and interpreted by a cell or multi-cellular organism. A signal is an entity used to transmit or convey information. suffers from implying that signaling is necessarily a linked process from start to finish. However, when scientists study signals, they rarely study the entire process in a single experiment, and any evidence that a signal has been produced or transmitted is not necessarily evidence that it will be received and interpreted. Thus, relative to the points I have raised above, I think biological signaling might be defined better as follows: Any biological process involved in the generation, transmission, reception, or interpretation of a signal. A signal is an entity used to transmit or convey information." Similarly, "cellular signaling" ought best be defined: Any biological process carried out by an individual cell involved in the generation, transmission, reception, or interpretation of a signal. and "multicellular organismal signaling" should be defined as Any biological process involved in the generation, transmission, reception, or interpretation of a signal that is carried out within a multicellular organism . and "multi-organismal" should be defined as Any biological process involved in the generation, transmission, reception, or interpretation of a signal between two organisms. 3) I am not completely pleased with formulation and definitions of "directed signaling" and "systemic signaling." It appears that you have considered neuronal (directed) and hormonal signaling (systemic) as the only types of signaling that exist within organisms. However quite a lot of signaling is done locally, but not necessarily in a directed manner. Cytokines almost always act locally, not systemically, but not in the "direct" manner in your first definition. There are hundreds of cytokines acting continuously in development and in homeostasis of the adult organism, so this is not an insignificant type of signaling nor one limited to the immune system. In fact, I wonder if you should avoid the word "directed" completely. Do you mean direct as in going straight between two points, or direct, as in acting without any intermediate steps? In a third sense, most soluble signal mediators act in a direct manner, in that only cells with the specific receptors for those mediators can receive the signal, no matter where those cells are located relative to the cell producing the signal. I suspect what you mean by directed signaling is really "guided signaling," where the signal is guided to its destination or receptor directly along an axon or across a synapse. This may be a better term to use. I think one ought to be careful about the "systemic hormone signaling" term, as I think that the basic definition of a hormone is a diffusible molecule that can act at locations remote to the point of production. Many hormones do not act systemically at all, but rather in distinct locations. There are, of course, some metabolic hormones, such as insulin or epinephrine, that do have broader systemic effects, but we should not confuse them with other hormones that have much more targeted effects. Thus, I am not opposed to "systemic signaling", but clearly you need two additional terms, "localized signaling" to cover the vast number of cytokines that signal locally and "long-range signaling" to account for the fact that many hormones do not in fact act systemically. In fact all these terms ought to be placed under a parental term "signaling via diffusible molecule" as "systemic signaling via diffusible molecule," "long-range signaling via diffusible molecule," and "localized signaling via diffusible molecule." We may define locally as acting within the same tissue or an adjacent tissue, and long-range as acting within any tissue not directly adjacent to the source of the signaling molecule. So we have multicellular organismal signaling --% guided signaling ----% guided nervous system signaling --% signaling via diffusible molecule ----% localized signaling via diffusible molecule ----% long-range signaling via diffusible molecule ------% long-range hormone signaling ----% systemic signaling via diffusible molecule ------% systemic hormone signaling --% nervous system signaling ----% guided nervous system signaling ----% nervous system signaling via diffusible molecule Arguably, it may be sufficient to have only the "signaling via diffusible molecule" itself, and skip all the child terms, since I'm not sure if they represent real or worthwhile distinctions. But this can be discussed. I would be happy to provide additional definitions.
19th March 2009
Meeting with Sandra Orchard, signaling expert at EMBL-EBI, and Jennifer Deegan.
Sandra's comments on terms and Alex's ideas:
When does cellular stop and multicellular start? It's not clear what we are meant to annotate to these. Even in multicellular signaling the signal must originally come from single cells and it must eventually be received by single cells. cell-cell signaling = intercellular singaling Synonym addition? Alex's point 1: change def to 'may be transmitted' Alex said he wanted long range signaling but we can't define that. Alex's point 2: def changes fine. Just semantics and not a big problem. diffusable is nice alternative to systemic. Hormone class: Should we really have hormone? It is like drug. We should have response to signaling molecule but not class under hormone. Cytokines also act at small concentrations and have big effects. There is little difference. Hormones are mostly lipids in higher mammals, whereas cytokines are proteins produced by blood cells. Otherwise there are few differences. Directed: Not sure that 'directed' is great, and whilst the idea makes sense once explained, the def does not give a clear idea of what we are getting at.'Guided' is not good either Can't think of aything to anotate to 'transmision' term. 'Secretion' should be part of 'creation' not 'transmission'. this from Alex is good: Arguably, it may be sufficient to have only the "signaling via diffusible molecule" itself, and skip all the child terms, since I'm not sure if they represent real or worthwhile distinctions. But this can be discussed. I would be happy to provide additional definitions.
20th March 2009
Emily Dimmer asks whether we will include the signal starting a given signaling pathway in the definition of that pathway. So for example in a hormone signaling pathway, is the binding of the hormone to the receptor part of the pathway, or is the pathway considered to start after the binding event? Also how will we define the end of the pathway?
David says we should name by the receptor, not the signal as signals can trigger different pathways.
Jen to follow up signaling alone with Cambridge experts. David now working on lungs.
multicell v cell signal answer: You can annotate to both term, but if you don't know which cells in an organ are signaling then you can just use the organ signaling term.
Plan: Jen to ask Sandra if they can do double act on signal overhaul and present proposal to working group. (I have written to ask if we could meet one hour a week.)
23rd March 2009
Sandra has been thinking further about possible improvements to the top signaling terms and suggests the following structure. At present there are no definitions, but the meaning is fairly clear from the names. The graph is shown in two different visual formats in case people find one version easier than the other. The two pictures show identical graphs though.
Further request from Emily:
In a case of a term like insulin signaling pathway, where insulin is the pathway that is started by insulin, we need to know whether the insulin is actually considered part of the pathway. So should the annotator annotate insulin to 'insulin signaling pathway' or to 'positive regulation of insulin signaling pathway'?
Further request from Lenov:
Beside "signal transduction via ionic flux" and "signal transduction via diffusible mediator", I think we need signal transduction via conformational transition. This is the case of lattice of receptors or large assemblies of scaffold and enzymes, but more simply the case of GPCR transduction. Yes, G-protein alpha diffuses next to the membrane from the GPCR to the target, but the signal transduction parts are the stabilisation of a GTPase form of Galpha that itself stabilise e.g. the active form of adenylate cyclase (and Galphai actually stabilises the inactive form of AdC, and do not compete with Galphas as we sometimes read). Some GPCRs also directly affect the final targets without diffusion of G-alpha protein at all.
Comments from Alex
Standard GO annotation practice has been to avoid annotating the extracellular ligand as being part of the signaling pathway. So insulin is not annotated to the 'insulin signaling pathway' but rather to 'positive regulation of insulin signaling pathway'.
Also, last time I checked, lipids are chemicals, and proteins and peptides are chemicals too. We use the word chemical in some vague and imprecise way here and elsewhere in the GO, and it would be much better to use a term to describe which subset of the chemical world we mean. I can't solve that problem here, but I insist that we do so. It's not enough to say all chemicals which are not lipids and not proteins or peptides -- we need a positive and specific term and definition.
Lenov: I agree. Many efforts such as BioPAX use "small molecules", but this is not satisfactory either, because macrolides are bigger than small polypeptides, and radicals and ions are not always molecules. In SBGN we ended up using "simple chemical" by contrast to "macromolecules". Macromolecules are build upon scaffolds made up of repetitive elements. They are: polypeptides, nucleic acids, polysaccharides and lipids. So calcium ion, ATP, free radicals are simple chemicals.
I am further concerned that the meaning of the term 'signal tranduction' is being expanded from its current definition in the GO to mean something more like 'signal transmission'. Wikipedia (amazingly enough) has a quite a good and succinct definition of 'signal transduction': "any process by which a cell converts one kind of signal or stimulus into another." The current GO definition is "The cascade of processes by which a signal interacts with a receptor, causing a change in the level or activity of a second messenger or other downstream target, and ultimately effecting a change in the functioning of the cell," which says something similar, but not as cleanly.
'Signal transmission' is simply the movement of a signal from one place to another, with or without a change in form. 'Signal transduction' is a type of 'signal transmission.' Thus most or all of the places that say 'signal transduction' in the graph above ought to be renamed 'signal transmission.'
Comments from Sandra
I agree with the comments about "chemical" but couldn't quickly come up with a term that encompassed biochemicals (other than proteins and lipids) but included quite complex small molecules, monosaccharides etc, so was using it as a placeholder until a better term came along.
And yes I totally forget about conformational changes but it should be included.
Discussion on list
Peter D'Eustachio suggested that we might use encoded and non-encoded to distinguish different molecules/chemicals.
6th April 2009
Participants: Sandra Orchard, Jennifer Deegan
Encoded is no good as lipids account for most of the hormones, and they should be split out quite high up. ChEBI regard lipids as 'chemicals'. Lipids are not encoded.
'Conformational' is fine.
'Signal transmission' is fine if preferred. Add 'transduction' as synonym
We made a couple of changes to the proposed structure to accommodate these thoughts. For now 'chemical' has been changed to '(chemical?)' until we get a better option.
We merged the proposed structure into a branch on the live GO and checked it into scratch as go/scratch/signaling3.obo. Sandra started sorting the pre-existing terms under the new ones. I am going to mail the file to her so she can think further about what to do. We will then show her ideas to the list in the form of this file and ask them to send their thoughts.
4th June 2009
Meeting with David Hill, Sandra Orchard, Varsha Khodiyar and Jennifer Deegan to edit SMAD signaling terms.
Sandra Orchard currently has editing control of the signaling branch file but is having trouble finding time to work on it as her group is currently short staffed.
29th June 2009
Since Sandra is a bit too busy to work on signaling just now I am organising a meeting of all the contributers to see if there is another route that we can try.
Made doodle poll to arrange meeting. Fixed on: 13th July at 8a.m. PDT, 9a.m. MDT, 10a.m. CDT, 11a.m. EDT, 4p.m. BST, 5p.m. CEST?
Wrote to participants to arrange date.
Participants to be Nicolas Le Novere, Alex Diehl, Ruth Lovering, Peter D'Eustachio, Stan Lauderkind, Harold Drabkin, Sandra Orchard, Val Wood, Susan Tweedie. Jane lomax also indicated an interest but is not available at the time that fits for everyone else.
For continuation of these notes see: