Talk:PAMGO work in progress

From GO Wiki
Jump to navigation Jump to search

Some issues I'm encountering whilst working through the file. Please put your thoughts on the matter underneath each issue. Also see the PAMGO terms to be defined page for a list of PAMGO terms which need definitions.


The first draft of the ontology changes can be found in this SourceForge item:

https://sourceforge.net/tracker/index.php?func=detail&aid=1546635&group_id=36855&atid=440764

It would be great if we could discuss any points on the wiki here, since it allows you to organise things much better than SourceForge.


If you want to leave a comment, click on the "edit" tag for that section, and add your comment at the bottom, starting it with your name as a subheader (put three equals signs in front of and after your name, i.e.

===The Invisible Hulk===

). This will automatically create a link at the top of the page, making it easy to navigate to.

If the point you want to talk about isn't already in the list here, just click the + sign at the top of the page; that creates a new comment section.


Thanks!

The 'other organism' string

We decided to put "by other organism" into the terms referring to processes performed by an organism involving a host to clarify which organism was doing what, e.g. "modification by other organism of host morphology or physiology via secreted substance". You would use these terms to annotate the gene products of the smaller organism, i.e. the "other organism" referred to in the term. I think this gets confusing if you then look at the terms for interactions where there isn't a definable host or symbiont, as they also refer to "other organism", e.g. "modification of morphology or physiology of other organism during symbiotic interaction". In these cases, the "other organism" isn't the organism you're annotating! I think it might be better to use the string "by organism", rather than "by other organism", e.g.

  • modification by organism of host morphology or physiology via secreted substance
  • positive regulation by organism of host signal transduction pathway
  • formation by organism of haustorium for nutrient acquisition from host

Does that seem reasonable?

Marcus

COMMENT FROM Mchibuco 13:15, 8 August 2006 (PDT)

I think you raise a good point that it gets confusing when we indiscriminately use the phrase "by other organism," but I think before we just cut the word "other" we need to recall the Biological Process Ontology Guidelines, which state that:

"Symbiotic relationships may be between two organisms of similar sizes or of differing sizes, and most of the processes under symbiosis, encompassing mutualism through parasitism have child terms to specify the sizes of the organisms involved. These terms use the nomenclature host for the larger organism and symbiont for the smaller organism. For interactions where there is no clear host or symbiont, the wording other organism is used, and terms are appended with during symbiotic interaction to make it clear that they represent processes occurring during symbiosis."

In other words, I think we should use the more encompassing phrase "by other organism" or the shortened version "by organism" only in cases where we can not say "by host" or "by symbiont" explicitly.

In the first two examples, I would say:

  • modification by symbiont of host morphology or physiology via secreted substance
  • positive regulation by symbiont of host signal transduction pathway

This is because both definitions already use the word "host" so I think it is appropriate to say "symbiont," because by definition if it's a host, then it is a host to something, e.g. a symbiont. To say anything less is to obfuscate unnecessarily.

In the third example, we have more difficulty, because as I understand it, although a haustorium is a pathogen-derived structure, it is in effect an interface with a host cell membrane. Hence, I would ask this: is a haustorium a pathogenic structure or a joint host-pathogen structure? Although a host cell membrane will not form the invaginations it does in a haustorial context in the absence of a haustorium, I still see it as a joint interface caused by the pathogen, but that requires the host.

Thoughts?

Amelia

Jane and I wrote the documentation, so it would be easy enough to change it to something else if we decide on a new format. The only thing that I think might be problematic about saying 'modification by symbiont of host morphology or physiology' or similar is that it might not be clear from the term string who the term should be used by, whether it is a term to annotate symbiont gps or host gps to. I think if you say 'by organism', it's clear(er) that it is a term for the annotation of symbiont gps.

In the light of Marcus's comments about haustoria, should we have both host and symbiont terms for the formation of a haustorium?

Michelle

I agree with Marcus's suggestion on using "by symbiont" and "by host" whenever possible. I think it is more clear than "by organism" since whoever's gene product one is annotating will be the one in the "by ____" phrase. Also the definitions will say this too. I think using "by organism" is much more vague since both the symbiont and host could be the "organism" referred to.

Midori

I wholeheartedly agree with Michelle about 'by organism'.

David

I also agree with Michelle. Organism is vague.

Amelia

Can anyone remember why it was we decided not to go for specifying 'host' and 'symbiont' in the terms in the first place? I remember there was some specific reason that we didn't do it, but I can't for the life of me remember what it might have been...

Candace

I'm finally getting started on all the discussion on the wiki -- sorry it took so long. I don't think there ever was any specific reason NOT to specify. Remember that when we created the first 30 or so terms, we were doing only those terms useful for annotating genes in a microbe that was interacting with a host (in our case, a plant host). Therefore, in all of those early terms (back when it was blessedly simple), it was clear to say "adhesion to host by other organism" and all was clear. I very much agree with Michelle, Marcus and others that using "other organism" will be very vague and confusing, especially to people who aren't annotating genes to these terms every day. We should use "by host" and "by symbiont" whenever possible, and if that's not possible, then somehow we have to make it very clear who is doing what to whom.....

Hypersensitive response

See this SF item for response from TAIR:

https://sourceforge.net/tracker/?func=detail&atid=440764&aid=1524725&group_id=36855

Marcus

I think Donghui Li’s thoughts (from the link above) on PCD and HR make it clear that [in plants] PCD comprises both developmentally-related forms of cell death and cell death in response to a pathogen. In this light, even if mechanisms underlying HR are not distinct from those of PCD (and the distinction is therefore artificial), I see HR as a useful narrower child term to capture the context of PCD as response to a pathogen. Therefore, I agree with Donghui that the terms should not be merged, and I offer further justification for not merging the terms.

What would happen if host cells were induced to die in response to a non-pathogenic symbiont, e.g. a case of non-HR PCD in response to an organism? It would be interesting to capture this information as distinct from HR type PCD. I found an interesting paper abstract (PMID: 16946237) that documents symbiont-induced apoptosis during light organ formation in the squid Euprymna scolopes by the marine luminous bacterium Vibrio fischeri. I think it would be interesting to capture this case of non-HR PCD induced by another organism as distinct from HR PCD.

Currently there is no term for apoptosis induced by another organism (that I could find), with the exception of virus-infected cell apoptosis (GO:0006926). Although “induction of apoptosis by extracellular signals” (GO:0008624) is a possibility, all its children appear to be endogenous factors such as hormones, and I want to capture that another organism is causing the phenomenon. Perhaps killing of host cells (GO:0001907) would be the best term to use since it is under the IBO node (GO:0051704), but its definition is “any process mediated by an organism that results in the death of cells in the host organism…” In the case of the Euprymna-Vibrio interaction, although host PCD is triggered by a symbiont-induced signal, the killing is done via pathways endogenous to the host. That term is inappropriate for the same reason that it would be to characterize HR.

I believe we should create a new term “symbiont-induced programmed cell death” as a child of programmed cell death. We could move hypersensitive response from its current position as a child of programmed cell death to a child of the new term. We could create a new sibling for HR called “programmed cell death induced by non-pathogen symbiont.” This would allow us clearly to draw the distinction between PCD as a response to a pathogen and PCD as part of a beneficial symbiotic interaction. If HR had been merged upward into PCD, this distinction between symbiont-induced PCD and pathogen-specific PCD would not be possible. Furthermore, if a case arises where it is unclear whether PCD in a host that is induced by a symbiont is “harmful” or “beneficial” (as I imagine would happen sometimes), that case could be annotated to the new parent term “symbiont-induced programmed cell death.”

  • programmed cell death (PCD) GO:0012501, Definition: Cell death resulting from activation of endogenous cellular processes.
    • symbiont-induced programmed cell death GO:XXXXXXX, Definition: Cell death in a host resulting from activation of endogenous cellular processes after direct or indirect interaction with a symbiont. Comment: An example of direct interaction is penetration by hyphae, and examples of indirect interaction include encounter with secreted enzyme and interaction with MAMP such as cell wall fragment.
      • hypersensitive response (HR) GO:0009626, Definition: The rapid death of plant cells in response to invasion by a pathogen.
      • programmed cell death induced by non-pathogen symbiont GO:XXXXXXX, Definition: Cell death in a host resulting from activation of endogenous cellular processes after direct or indirect interaction with a non-pathogenic symbiont. Comment: This includes interactions typically viewed as beneficial from the perspective of the host, such as luminous bacteria in the light organs of marine squid.

==

Trudy

I agree with Donghui Li and Marcus that HR should not be merged with PCD. In addition to the reasons that both Donghui Li and Marcus present, it has not been conclusively shown that these two phenomena are the same. Infact most papers refer to HR as a type of PCD.

I am still thinking through Marcus’s response on creating the term "programmed cell death induced by non-pathogen symbiont GO:XXXXXXX" as I am not sure the use of the term non-pathogen is appropriate in light of the way PAMGO defines symbiosis.... All the same, I thought of throwing in this bit about apoptosis so we can consider it whilst working on the PCD/HR node. Apoptosis is referred to as a type of PCD (can be found in GO defn of apoptosis GO:0006915) and HR is also sometimes said to be a type of apoptosis (see below)


Eur. J. Biochem. 267, 5078-5084 (2000) Apoptosis (programmed cell death) is a ubiquitous active process that occurs during development and in response to environmental stimuli. Apoptosis has been dAescribed in animal cells in great detail at the morphological, biochemical and genetic levels [1-3]. In plants, apoptosis has been associated with various phases of development and senescence [4,5], germination [6], response to salt stress [7] or cold [8]. A particularly interesting type of apoptosis has been observed during the plant response to pathogen attack, and was termed ‘hypersensitive response’ (HR) [5,9]. Signal transduction pathways are activated during HR, leading to biosynthesis or release of potential anti-microbial effector molecules, which are thought to contribute to both host and pathogen cell death [5,9]. The molecular mechanism of plant apoptosis and HR are being disclosed, and involve small GTP-binding proteins [10], arabinogalactan proteins [11], subcellular reorganization of mitochondria [12], Rubisco proteases [13], mannose-inducible endonucleases [14] and Bax [15]. Among other signals, rapid generation of reactive oxygen species has been implicated in HR of plants against pathogens [16-18]. In particular, hydrogen peroxide (H2O2) has been shown to be a crucial component of the HR control circuit, to such an extent that a short pulse of exogenous H2O2 is sufficient to activate the hypersensitive cell death programme [19,20]. Reactive oxygen species, H2O2 and lipid peroxides have been long considered crucial elements of apoptosis in animals [21,22]. More recently, the peroxides produced by lipoxygenase activity have received attention as mediators of apoptosis in animal cells [23-27]. Remarkably, lipoxygenase-dependent pathways are implicated also in plant response to abiotic stress [28] and development of HR [29]. Lentil (Lens culinaris), a member of the Fabaceae, is an annual legume crop of nutritional quality higher than that of cereals, meat and fish, and is severely affected by pathogens [30]. Lentil seedlings contain different lipoxygenase (LOX) isozymes and we have characterized [31], cloned [32] and expressed in Escherichia coli [33] the isozyme most abundant in shoots. Lentil roots contain a different LOX, which shares several properties (e.g. molecular weight, pH optimum, substrate specificity) with the shoot enzyme and is recognized by the same monoclonal antibodies [34,35]. Therefore, lentil root cells were chosen to investigate the possible role of the lipoxygenase pathway in plant response to oxidative stress by exogenous H2O2.

Do we need to incorporate apoptosis in the PCD hierarchy as well?

Modification of host protein function

At the moment, modification of host protein function is a child of 'modification ... via secreted substance'. This seems OK to me, but are there any ways in which host protein function can be modified directly or via some other method?

Modification/modulation of host defenses

Host defenses consist of the induced defenses - the defense response mounted by the host - and the preexisting defenses, such as the cell wall and other defensive structures. We've got modification of host structure and modulation of host defense response, but nothing to encapsulate both - logically it seems like a term we should have since elsewhere we have structures like this:

  • evasion or tolerance of host defenses
    • evasion or tolerance of host defense response

Currently, we've got the term 'response to host defenses ; GO:0044413' (was 'avoidance of host defenses' but we renamed it), which is defined thus:

Any process, either constitutive or induced, by which an organism evades, suppresses or tolerates the effects of its host organism's defense(s). Host defenses may be induced by the presence of the organism or may be preformed (e.g. physical barriers). The host is defined as the larger of the organisms involved in a symbiotic interaction.

I am no longer sure that this term is correctly named, as there are situations where the symbiont induces host defenses, but the def of 'response to host defenses' only covers the "negative" aspects of the response. Perhaps 'response to host defenses' should be used as a term to encapsulate modification/modulation of host defenses?

As yet I don't have any suggestion for a new name for 'response to host defenses', unfortunately!

Symbiont defenses

I am just contemplating adding terms for the following:

  • modulation by organism of symbiont defense response
    • modulation by organism of symbiont inflammatory response
    • modulation by organism of symbiont innate immunity

The first one is OK, but I'm not sure about the second and third. Are there/might there be symbioses out there where the smaller partner is a multicellular organism with inflammatory responses and innate immunity which the host species regulates in some way? Most of the symbioses between larger organisms seem to be more behavioural...

Innate immunity

At the moment our terms refer to 'innate immunity', whereas in the main GO, they talk about 'innate immune response'. Is it OK to change the terms to that string instead, e.g. modulation of host innate immune response

Acquisition of nutrients from host

...is defined as

"The production of structures and/or molecules in an organism that are required for the acquisition and/or utilization of nutrients obtained from its host. The host is defined as the larger of the organisms involved in a symbiotic interaction."

How are we going to differentiate "formation by organism of specialized structure for nutrient acquisition from host" from this term? Is the def for "acquisition of nutrients from host" too detailed?

Michelle

Well, it seems an example of two things that fall under the parent term would be haustoria formation and hemolysin. The haustoria is a structure for gaining nutrients, while hemolysin is a molecule produced by the pathogen which causes nutrients to be dumped into the environment which the pathogen then slurps up - no specialized structures needed. So I think the "formation by organism of specialized...." term is correctly a more specific child of the parent. There could also be "cytolysis of host cells for the purpose of nutrient acquisition" as another child. I don't think we necessarily need such a term, and concurrent annotations should be able to handle it, but I am just trying to illustrate what I mean.

Entry of organism into host cell by promotion of host phagocytosis

def: "The invasion by an organism of a cell of its host organism by utilizing the host phagocytosis mechanism."

Just to check - does this term mean that the organism upregulates host phagocytosis?


Michelle

Yes, at least that was my understanding.

Recognition of other organism during symbiotic interaction

This is defined as "The specific processes that allow an organism to detect the presence of a second organism via physical or chemical signals, where the two organisms are in a symbiotic interaction.".

Could this be a standard detection term? The std def for detection terms is:

The series of events in which a [...] stimulus is received by a cell and converted into a molecular signal.

or is recognition more of an organismal-level process?

I was just wondering as it would be nice to have a structure that mimics the main ontology, i.e.

response to other organism

  • detection of other organism

Induction terms

At the moment, under positive regulation, we have various 'induction' and 'activation' terms. What specifically should these terms mean? Are they induction as in kicking off an inactive process, or are the induction meaning general positive regulation, i.e. activating an inactive process and perpetuating or upregulating an existing process? The terms in question are:

  • GO:0052103 activation by organism of induced systemic resistance in host
  • GO:0052104 activation by organism of systemic acquired resistance in host
  • GO:0052030 induction by organism of host apoptotic programmed cell death
  • GO:0052065 induction by organism of host calcium ion flux
  • GO:0044416 induction by organism of host defense response
  • GO:0052105 induction by organism of host defensive cell wall thickening
  • GO:0052063 induction by organism of host nitric oxide production
  • GO:0012504 induction by organism of host non-apoptotic programmed cell death
  • GO:0052062 induction by organism of host phytoalexin production
  • GO:0052044 induction by organism of host programmed cell death
  • GO:0052064 induction by organism of host reactive oxidative species production
  • GO:0052101 induction by organism of host resistance gene-dependent defense response
  • GO:0052033 pathogen-associated molecular pattern dependent induction by organism of host innate immunity

Michelle

I seem to recall that for the induction terms we meant "kicking off an inactive process" and for upregulating we used "enhancement" however, I am not sure that we put in "enhancement" terms everywhere we needed them. We may need to create analagous enhancement terms for many of the induction terms.

Protein secretion systems

We've got the terms for modification of morphology and physiology using substances secreted by various types of secretion systems. Are these systems only found in symbionts, or might there be symbioses where the host organism had type II / III / IV (i.e. a bacterium was the host)? Also, we've got the term names as follows:

  • GO:0052049 interaction with host via protein secreted by type III secretion system
  • GO:0052050 interaction with host via substance secreted by type IV secretion system
  • GO:0052051 interaction with host via protein secreted by type II secretion system

Should these all be "... *protein* secreted by type ..."? Do other substances ever get secreted by these systems?

T4SS secretes protein and DNA (Donghui Li)

Three distinct secretion pathyways have been studied in plant pathogens.

The Type II secretion systems (T2SS) exports enzymes that are involved in degrading the plant cell wall, including pectinases, endo-glucanases and cellulases.

T3SS is the most widely studied secretion system in plant pathogens. Pathogens use T3SS to inject effectors (proteins) into the host cell. There is a great diversity of effectors; they have diverse enzymatic activities. Bacterial effectors have a prominent role in promoting the virulence in plants.

The T4SS has a critical role in the pathogenesis of Argobacterium and its capacity to form galls on plants. The T4SS mediates the trafficking of bacterial proteins and DNA into the plant cell. The bacterial DNA is integrated into the host genome to produce hormones that modify the phsyiology of plants.

Many pathogens rely on multiple secretion systems. For example, several strains of Xanthomonas have T2SS, T3SS and T4SS.

Based on the above discussion, it is clear that the use of 'substance' instead of 'protein' for T4SS is appropriate: T4SS secretes both protein and DNA, while T2SS and T3SS only secrete proteins.

Ref: Abramovitch et al (2006) Nature Reviews Molecular Cell Biology, 7:601-611.

Protein secretion systems (again)

Def of "modification by organism of host morphology or physiology via secreted substance":

The process by which an organism effects a change in the structure or function of its host organism, mediated by a substance secreted by one of the organisms.

Is this supposed to be suitable for referring to either host- or symbiont-secreted substances, or is it supposed to be specific to the organism secreting the substance (i.e. the symbiont in this case)?

MAPK pathways

At the moment, we have regulation of the defense-related MAPK signalling pathways as an is-a of regulation of ethylene-mediated defense response. I vaguely remember discussing that but I just want to check it's correct.

Avoidance, Evasion and Tolerance, oh my!

I'm trying to work out a systematic way to deal with the terms under 'response to ... defenses', because at the moment, the terms have various names (evasion, evasion and tolerance, suppression, avoidance, response to, etc.) and the defs all seem to differ. This is the current structure (I've used a generic 'org' instead of spelling out all the host / symbiont / other organism terms):

  • response to [org] defenses
    • avoidance of [org] defenses
      • evasion or tolerance of [org] defenses
        • evasion by virus of host defenses
          • child terms include suppression of host processes
        • evasion or tolerance of [org] defense response
          • evasion or tolerance of [org] immune response
            • active evasion terms
            • passive evasion terms
            • evasion or tolerance of [org] oxidative burst / [org] phytoalexins / [org] NO
      • suppression of [org] defenses
    • modulation of [org] defense response

I would like to try to work out what we're actually meaning by "evasion", "tolerance" and so on.


  • "passive evasion" is defined as a process by which an organism deals with a defense response without affecting the other organism
  • "active evasion" is a process by which an organism deals with a defense response and affects the other organism into the bargain. It is not clear
  • "tolerance" isn't used on its own, so I am not sure what it represents
  • "suppression" seems to be very similar, if not identical, to "negative regulation" - i.e. stopping a process, impeding a process, or preventing an inactive process from starting.
  • "avoidance" seems to be the conjunction of "tolerance", "evasion" and "suppression"

Some of the term defs mention "constitutive" and "induced" processes (the avoidance terms), and others talk about "active" and "passive" (evasion or tolerance terms). I am not sure whether "active" and "passive" here means the same as "active evasion" and "passive evasion".


This is how I have been thinking of defenses and the defense response:

  • "defenses" applies to everything that an organism has developed over its lifespan and through evolution to deal with threats.
  • you can separate defenses into two types, structural defenses and defensive processes. I *think* this correlates with constitutive and induced defenses (aka the defense response). I can't think of any constitutive defenses that are a GO process, rather than just a protein sitting there holding the cell wall together.
  • structural defenses can't harm you, so the only thing you're likely to do to them is to attack them
  • the defense response of the other organism is likely to harm you in some way. There are three ways you could deal with this:
  1. build up fortifications to protect yourself against the effects of the other organism's process (e.g. cell wall)
  2. neutralize the effects of the other organism's process (e.g. detoxifying harmful chemicals)
  3. stop the other organism from performing the process

No. 1 is your structural defenses. They are constitutive and passive, and they can't be represented by a GO process term (I believe). No. 2 and 3 are your 'defense response'.

So your response to the defenses of the other organism could be split up thus:

response to [org] defenses
[i] modification of [org] defenses
---[i] physical modification of [org] defensive structures
---[i] modulation of [org] defense response (#3)
[i] response to [org] defense response
---[i] modulation of [org] defense response (#3)
---[i] examples of neutralizing the effects of the other org's defense response (#2)

No. 2 would include the new terms like 'evasion or tolerance of [org] phytoalexins'.

From this perspective, 'suppression of [org] defenses' doesn't make much sense, as you can't suppress a physical structure. I think it would be better to alter this term to be 'suppression of [org] defense response'.

Is 'tolerance' supposed to represent the passive / structural defenses?

[this is unfinished at the moment due to temporary loss of mind]

Later thoughts

For concepts where we have both modulation terms and evasion/tolerance terms, I've grouped them together under a 'response to' term, e.g.

response to [org] phytoalexin production
[i] evasion or tolerance of [org]-produced phytoalexins
[i] modulation of [org] phytoaxelin production
---[i] positive regulation of [org] phytoalexin production
------[i] induction of [org] phytoalexin production

I really need to know, for the 'avoidance' and the 'evasion or tolerance' terms, exactly what is being meant. I am wondering if the 'avoidance' terms could themselves be avoided. Hmmmm.

all is_a relationships

I'm interested in the rationale for building this portion of the ontology with terms that represent collections of processes rather than global processes that are then broken down into part_of subprocesses. Is there a plan to break these terms down into parts at some point? It seems for the nervous system development portion of the ontology we took the converse approach. We considered global processes and broke them down into parts. We are now going back and adding the is_a parents that will eventually have is_a relationships to the more 'primitive' terms in the process ontology.