new terms requested by PAMGO

[gocentral]

Dear GO Consortium members: This is Candace Collmer from the PAMGO (plant- associated microbe gene ontology) interest group. We are a group of scientists who work on the genome projects of various organismic plant pathogens
bacteria, oomycetes, fungi, and nematodes. We are trying to develop GO terms (for integration into the existing 3 ontologies) that will be appropriate for annotating the gene products of not only various plant pathogens, but also animal pathogens as well as organisms that enter into symbiotic or other helpful or neutral relationships with plants or animals, like Rhizobium (nitrogen fixation) and mycorrhizae. We feel that many of the initial processes involved in the interactions between both pathogens AND symbionts and their hosts are similar, and that, as was pointed out in the recent discussion about pathogenesis and disease on the GO Consortium email, the outcome (i.e., pathogenesis, symbiosis, or repulsion of the invader by the host) of such an initial interaction can vary depending on many things (e.g. strength of host defense, environment, etc.). Thus, we feel it would be very useful for scientists to be able to search across organisms of many types for gene products involved in similar types of initial (and later) interactions, no matter how those interactions may end up. This has been the thinking behind what we are trying to do. We obviously realize that we will need your help (and blessing) to succeed at this goal. We are sending our suggestions for our first attempt at higher order terms as well as the changes proposed in order to integrate those into the exisiting GO. We are certain that you will have additional perspectives on this, and welcome your suggestions as we try to find the best way to do this. The initial major change that we are proposing relates to the earlier discussion on the GO Consortium list about pathogenesis and similar terms. We have suggested eliminating both pathogenesis (GO:0009405) and host-pathogen interaction (GO:0030383) for the reasons given above to allow the annotation of gene products of both pathogens and symbionts to processes that are likely to be shared by both. In addition, we feel that having those two currently existing, different terms under Physiological Process and Cellular Process, respectively, may not be necessary and may in fact make it more difficult for newer users of GO to find the appropriate terms for annotation. As you will see below, we have proposed the new GO term, interaction with host organism to encompass both of these terms. Michelle Gwinn has been working with us on this proposal and she concurs with this solution to the pathogenesis problem. I think this is enough introduction. Our proposed scheme is given in the word file attached to this request (first the skeleton of the proposed tree, then the expanded tree with suggested definitions and embedded questions and comments). We hope to attend an upcoming consortium meeting to meet everyone fact-to-face (this was mentioned by Michelle in an email to your group last fall and met a favorable response). If this is especially useful to work on the proposed terms, please advise us as to which meeting would be the best. Thanks for your consideration of our proposal.
--Candace Collmer

Reported by: ccollmer

Original Ticket: "geneontology/ontology-requests/1769":https://sourceforge.net/p/geneontology/ontology-requests/1769

[gocentral]

new terms, definitions, and tree from PAMGO

Original comment by: ccollmer

[gocentral]

Logged In: YES user_id=614564

I like the new structure for the defense tree. However, I don't agree that the defense response to pathogenic bacteria and other similar terms should be obsoleted. Pathogenic and non-pathogenic responses are very well recognized branches of host response to pathogens in plant system. Quite a few plant genes are annotated to these terms using literature.

Suparna

Original comment by: smundodi

[gocentral]

Logged In: YES user_id=614564

Are the nonhost terms similar to pathogenic and host non- pathogenic? How are we accomodating the existing response terms to pathogenic and non-pathogenic organisms in this tree?

Original comment by: smundodi

[gocentral]

Logged In: YES user_id=1035029

I'm experimenting to see if this is the way to add a commetn related to disucssion about our submission. Please advise me if I should do this in a different way. However, please note that our proposed terms, at least so far, relate only to annotating the genes of microbes -- pathogens, symbionts, etc. We have not sent in anything about terms related to defense responses by the host.
Thus, since the upper level term of our scheme is "interaction with other organism", the term "interaction with host organism" refers to the situation where a pathogen or symbiont is interacting with an organism (often a plant or animal) that can serve as a host to it. And "interaction with nonhost organism" encompasses interactions between organisms that are not involved in a host-parasite or host- symbiont interaction, e.g. like the organisms in a biofilm, or free-living organisms in the rhizosphere around a plant root. Hope that that clarifies. --Candace Collmer

Original comment by: ccollmer

[gocentral]

Logged In: YES user_id=616548

Hi Candace,

I'm working on a genome database for Candida albicans (CGD) and have looked over your proposal in light of its applicability to a single-celled fungal pathogen of humans. In general, I think it looks really good. There are some specific terms that we might want to add for Candida and other fungal pathogens, but I think we'd wait until the proposal in general is accepted.

This may be too much detail for right now, but I have a comment on the term 'biofilm formation'. Currently it's a child of 'interaction with non- host organism'; to my mind, this would describe the interactions between the different organisms that make up the biofilm. But what about when a biofilm is formed inside a host organism, as frequently happens with C. albicans? In this case, it's clearly a type of interaction with the host as well. Would something like this work?

-%interaction with other organism --%interaction with non-host organism ---%biofilm formation (GO:0042710) ----%biofilm formation in or on host organism

--%interaction with host organism ---%adhesion to host ----%biofilm formation in or on host organism

Maria

Original comment by: mariacostanzo

[gocentral]

pamgo obo file

Original comment by: jenclark

[gocentral]

Logged In: YES user_id=1035029

Dear GO members I want to alert you to the posting on Sourceforge of the revised PAMGO proposal, as requested following discussion at the GO content meeting held at the Carnegie Institute August 22-23, 2004. At that meeting we presented the original PAMGO proposal for terms for annotating gene products in microbes that are involved in the process of pathogenesis as well as in other types of symbiosis with a host organism. We specifically proposed general terms that we hoped would serve communities studying both plant and animal pathogens, as well as microbes (and nematodes) involved in mutualism, commensalism and other types of symbiosis. At that time our proposal did not include the words symbiosis or pathogenesis. While GO members at that meeting were generally supportive of our proposal, they also felt strongly that scientists wanting to use GO would naturally search for the terms pathogenesis and symbiosis, and that those should be incorporated. At that time we were asked to add trees for both of these terms, under which would be found all of our proposed general terms, but made specific for each of those processes. We would also keep the tree of general terms, but have as children for each general term, that same term applied to both pathogenesis and symbiosis. In the process of offering comments on the minutes of that meeting, and then trying to make definitions for new terms to be added to our original tree, Michelle Gwinn and I (Candace Collmer) realized that the most widely used definition of symbiosis included not only mutualism and commensalism (where one or both partners of the interaction benefits), but also parasitism (and thus pathogenesis). Because of that, we could not make a new proposal that would separate the process of pathogenesis from the process of symbiosis. I should point out that the original PAMGO proposal sought specifically not to separate the different types of symbiosis (parasitism, mutualism, commensalism) because some microbes, and thus their gene products, can be involved in different ones of these processes at different times, depending on the condition of the host, environmental conditions, etc. As a result, we had to modify our revised proposal to acknowledge that pathogenesis is one type of symbiosis, and thus must be a child of that term rather than a sibling of it.
We then made a revised tree that reflected that relationship, where our original process terms have been made specific to symbiosis, but also to pathogenesis (the latter as a child of the former). That proposal is the attached word document: GO-PAMGO-sym-path.doc. However, from looking at that proposal it is clear that there are some redundancies that occur as a result, and some of the trees seem to be more complex than desirable. This complexity is especially important considering that the long term goal is to add to these trees additional terms that will allow the annotating of gene products from the host side of the various symbiotic interactions. Therefore, we are also attaching a word document of a more simplified tree of terms, where symbiosis and pathogenesis are included as part of the terms (and thus would allow searching for them), but where the process terms proposed stay as general ones under symbiotic interaction between host and another organism during mutualism, commensalism, or parasitism (pathogenesis). That can be viewed in the second attached word document: GO-PAMGO-sym-only.doc. Please note that I have left spaces between the different categories of general terms in both of the documents, so that they can be more easily compared to each other. Those will eventually come out, of course. Please note that in both options we have proposed the term symbiotic interaction between host and another organism during mutualism, commensalism, or parasitism (pathogenesis) as a child of symbiosis, including relationships of mutualism, commensalism, or parasitism (pathogenesis), and the parent of our terms.
This recognizes that there are some symbiotic interactions that do not involve a true host (e.g. lichens), and thus there is room for expansion to address those if needed in the future. We hope that this proposal (and these two options) can be discussed at the upcoming GO meeting in Chicago, and that one of the two trees can be approved for inclusion in GO. (We recognize that we have not yet submitted definitions for all of the terms, but will complete that as soon as we know which tree will be accepted. There are a number of terms that would not need to be included if the simpler tree were approved.) Several groups are waiting to be able to use these terms for annotating genes in bacteria and other microbes.

Original comment by: ccollmer

[gocentral]

PAMGO-symbiosis only, PAMGO-symbiosis and pathogenesis

Original comment by: ccollmer

[gocentral]

Original comment by: ccollmer

[gocentral]

Logged In: YES user_id=1035029

and here is the second word document

Original comment by: ccollmer

[gocentral]

Logged In: YES user_id=865072

Having thought about this issue a lot, I have to admit I am in favor of dispensing with the distinction between symbiosis and pathogenesis, and thus favor the simpler tree that does not try to force a distinction.

The vast majority of these terms are for lower level processes that may or may not been seen at different times in an organism's lifecycle depending on external stimuli as well as the genetic program of the organism. Just as many simple processes are collection of distinct functions, more complicated processes, such as undertaking a virulence program or establishing a symbiotic relationship with another organism, really rely on execution of many lower level processes that are the same whether or not the overall goal of the organism is acting in a pathogenic or symbiotic way and are potentially utilized by both types of interactions.
Thus, type III or type IV protein secretion is mechanistically the same whether or not the organism involved is doing the secretion during pathogenesis or symbiosis. And since in many cases the distinction between the two higher level processes is unclear (as Candace has articulated quite clearly), why should the GO attempt to apply the distinction to all the lower processes that may be involved in one or both at different times.

Another problem from the curation point of view, is the tendency of gene annotators (who propose the most terms) not to think beyond the paper they are looking at and the particular model system involved. A hypothetical situation: If a paper says "we conclude that gene X is involved in type III protein secretion by the pathogenic bacteria p. aquaductus during root colonization of plant Y" that is not sufficient to conclude that gene product X is only involved in type III protein secretion during pathogenesis.
Yet the tendency appears throughout the GO to make overly specific terms, and annotations to such terms can lead to false conclusions. Thus if gene product X above is annotated to "type III protein secretion during pathogenesis" because the bacteria involved is called "pathogenic" according to some vague criteria (leading to the false presumption that all of the bacteria's processes that affect another organism are derived some pathogenic intent on its part) then a reasonable person looking at GO annotations might conclude that gene product X is only involved in type III protein secretion during pathogenesis, when in fact it may be a required participant in all types of type III protein secretion, whether or not the outcome is something an unbiased human observer would call pathogenesis or symbiosis.

-- Alex

Original comment by: addiehl

[gocentral]

Logged In: YES user_id=482944

Hello, Following the Consortium meeting in Chicago, there were remaining concerns by some members on the (then) final structure of the new pathogensis and interaction with host terms. Another round of emails later we have a new structure that seems agreeable to all parties so far consulted.

The new structure does not have separate subtrees for symbiosis and pathogenesis, but rather encompasses all of the original general host interaction terms as children of symbiosis with pathogenesis as a sibling to them and grandchild of symbiosis.

This removes the problematic separation of the general terms like "recognition of host" into pathogenic and symbiotic forms - something that was causing much concern since it is sometimes impossible to tell apriori whether something is a pathogen or not.

The new structure retains the concept of pathogenesis without specific children. The expectation is that an annotator who knows their gene is involved in recognition of host during a pathogenic process can annotoate to both "recognition of host" and "pathogenesis".

Below is the latest version.

Please look it over and make any final comments.

Thanks, Michelle

physiological process -interaction between organisms --biofilm formation (GO:0042710) --competition with another, non-host, organism --cell killing (and its children terms) --symbiosis ---interaction between host and another organism ----pathogenesis ----recognition of host ----adhesion to host -----cytoadherence to microvasculature (GO:0020035) ----growth on or near host surface ----growth within host ----entry into host -----viral entry (GO:0046718) ------viral cell invasion -----entry into host through natural portals -----entry into host through host barriers -----cell invasion ------viral cell invasion ----avoidance of host defenses -----suppression of host defenses -----evasion of host defenses ------viral host defense evasion (GO:0019049) ------evasion of host defense response -------evasion of host immune response ----induction of host defense response -----viral induction of host immune response (GO:0046730) ----translocation of molecules into host -----translocation of DNA into host -----translocation of peptides or proteins into host ------type III protein secretion system (GO:0030254) ------type IV protein secretion system (GO:0030255) ----movement within host -----viral spread within host (GO:0046739) -----migration within host ----acquisition of nutrients from host ----modification of host morphology or physiology -----viral host cell process manipulation (GO:0019054) -----viral transformation (GO:0019087) ------viral immortalization (GO:0019088) -----disruption of host cells ------killing of host cells (and its children terms) -----induction in host of a tumor, nodule, or growth ------induction in host of a tumor, nodule, or growth containing transformed cells ----dissemination or transmission of an organism from a host -----viral transmission (GO:0019089) ------viral transmission via vector -----dissemination or transmission of an organism from a host by a vector ------viral transmission via vector

Original comment by: mlgwinn

[gocentral]

Logged In: YES user_id=614564

Michelle and others,

I checked with few plant symbiosis experts here on campus and asked their opinion on symbiosis.

Here are their comments:

Scientist one:(Melanie Jean Barnett from Sharon long's lab) "I suppose the answer lies in how one defines symbiosis--because we study rhizobia, we use symbiosis to refer the mutualistic interaction between bacteria & plant. Within our field, if we started referring to a crown gall on a legume, as a "symbiosis" people would be confused. However, I realize that symbiosis is used more broadly as a term to include interactions where one or both partners benefit. I suppose you can use the term however you like, as long as you clearly define how you are using it."

Comment by another scientist:(Robert Fischer from Sharon Long's lab)

"As to your question on the use of the term "symbiosis," as long as you are very clear about how you define the term, then you can use it how you want: you can either be very limiting (interactions where both partners benefit) or very expansive (to include interactions where one partners inflicts harm on the other).

Personally, I do not consider pathogenic interactions to be examples of symbiosis. I guess that's because I choose to define it in a more limited way."

So based on above comments and a discussion here with Sue Rhee, here is what I have to say:

Do you think it would be good to make symbiosis a synonym of "interaction between host and another organism". Thus the term symbiosis is not very obvious in the tree to avoid confusion among biologists but it stays as synonym because it is known to include all types of interactions in Biology-101. Also, if it includes mutualism,parasitism, pathogenesis etc, it is nothing but all the interactions between host and other organisms right? Thus, we can merge the symbiosis and interaction between host and organism terms.

What do you think?

Suparna

Original comment by: smundodi

[gocentral]

Logged In: YES user_id=482944

Hi Suparna,

First let me say that my scientific intuition also balked at making pathogenesis a child of symbiosis, it just didn't seem right, but the more I read up on symbiosis the more it became clear to me that symbiosis does indeed encompass all forms of interactions between species where the species in question live intimately with each other. I think the reason we don't generally use symbiosis this way day to day is because we have other more specific terms to describe the bad relationships (parasitism, pathogenesis).

I have looked now at many sources for definitions for symbiosis (encyclopedia, biology texts, dictionaries) and virtually all include parasitism (of which pathogenesis is a part) in the definition.

Second, I don't think we can make symbiosis a synonym of "interaction between host an another organism" since there are kinds of interactions which do not involve a host but which are also symbiotic, many types of biofilms for example (but not all biofilms are symbiotic since some have only one species.)

Likewise we can not merge symbiosis with "interaction with another organism" since not all interactions are symbiotic - some are predatory, some are competitive.

I really think we need to keep a separate symbiosis term.

It just occurs to me as I was re-reading some of the symbiosis definitions, that maybe the term we should have in the tree is "parasitism" rather that "pathogenesis". This removes that thorny "from the host perspective" issue that pathogenesis always did and still does have as well as the issue over whether something acutally makes something sick or not. Then we could have "pathogenesis" as a synonym to "parasitism".

Michelle

Original comment by: mlgwinn

[gocentral]

Logged In: YES user_id=629839

Suparna and I are OK with the use of the term 'symbiosis' and accept the latest structure proposed.

But we don't agree that 'parasitism' could be a synonym of 'pathogenesis'. You can 'parasitize' another organism without causing disease. Let's leave it as is. In the future, we could talk about whether to add 'parasitism' to the structure and what would be the most appropriate place if needs arise.

Original comment by: syrhee

[gocentral]

PAMGO-final-defs

Original comment by: ccollmer

[gocentral]

Logged In: YES user_id=1035029

Dear GO Consortium members: I want to alert you to the posting on Sourceforge (965023) of the revised-yet-again PAMGO (Plant-Associated Microbe Gene Ontology) proposal, which has taken into account relevant discussion at the GO content meeting held in California in August, 2004, at the GO meeting held in Chicago in October, 2004, and via email since then. The most recent addition accommodates the process of biofilm formation in or on a host, which required a further development of terms related to biofilm formation. We believe that this final proposal accommodates all comments and concerns and represents consensus. The original PAMGO proposal offered terms for annotating gene products in microbes that are involved in various types of symbiosis, including pathogenesis as one type of symbiosis involving interaction with a host organism. We specifically proposed general terms that we hoped would serve communities studying both plant and animal pathogens, as well as microbes (and nematodes) involved in mutualism, commensalism and other types of symbiosis. This final proposal achieves that goal, as symbiosis is now specifically defined with a broad definition that includes relationships along a continuum from mutualism through parasitism. The term pathogenesis is also present in the tree as a child of symbiosis, and as a sibling to all of the original general terms (e.g. recognition of host, entry into host, etc.) The idea is that a gene product involved in recognition of a host by a pathogen would be annotated to both the general term recognition of host and the term pathogenesis. A gene product involved in recognition of a host by a microbe about to initiate a mutualistic relationship would be annotated to the general term only (as that term is a child of symbiosis). Specific process terms such as mutualism could be added later, if desirable, to be siblings of pathogenesis and children of symbiosis. This posting includes two word documents one with the revised tree of GO terms (GO-PAMGO-fnlfnl-tree.doc), the other with the definitions for those terms (GO-PAMGO- fnlfnl-def.doc) as well as proposed obsoletions of existing GO terms, modifications of definitions of existing GO terms, and remaining questions. Thanks for all the work everyone did on this proposal. Candace Collmer

Original comment by: ccollmer

[gocentral]

PAMGO-final tree of terms

Original comment by: ccollmer

[gocentral]

Logged In: YES user_id=865072

Look for my interspersed comments below.

From the PAMGO definition document "GO-PAMGO-fnlfnl-def.doc":

Other aspects of the PAMGO proposal:

1) The currently existing GO term pathogenesis would get moved from its current location to the location indicated in the new PAMGO-proposed tree. GO needs to decide what to do with its three currently existing children terms invasive growth, necrosis, and hemolysis:

a) invasive growth GO:0001404 = Growth of a pathogenic organism that results in penetration into cells or tissues of the host organism. This often (but not necessarily) includes a filamentous growth form, and also can include secretion of proteases and lipases to break down host tissue. There are currently no genes annotated to this term. We propose making it obsolete. We have essentially introduced 2 new terms entry into host and growth within host that break up this term into 2 different pieces see above. We wrote a possible definition of growth within host that uses most of the current definition of this old GO term see if that works OK.

Alex: I think this is a good solution.

b) necrosis as we recall, this is not currently in Alexs cell killing terms we believe Alex has talked about making this term obsolete. However, there are currently 13 genes annotated to this term what happens to those?

Alex: As I have mentioned in previous posts and emails, the term "necrosis" is used in the literature by groups in different areas of biology with a number of varying explicit and implicit definitions, including cell death by non-PCD mechanisms and cell death by particular PCD mechanisms which may or may not include apoptosis-related mechanisms. The word "necrosis" is thus imprecise and debased, and accurate annotation to the GO term GO:0008220 necrosis may be difficult given the tendency of annotators to match lexical similarities rather than semantic similarities. I recommend that GO:0008220 necrosis be obsoleted, with the recommendation that existing and new annotations be mapped to GO:0008219 cell death, GO:0001906 cell killing, GO:0019835 cytolysis, or GO:0012501 programmed cell death (or an appropriate child term of any of these), depending on the data and assays employed.

c) hemolysis we believe this currently existing GO term now is located within Alexs cell killing tree. Is it OK there, and does it then automatically end up in our tree?

Alex: After the recent entry of the cell killing terms, GO:0019836 hemolysis is an is-a child of GO:0001897 cytolysis of host cells, and will retain that parentage after GO:0009405 pathogenesis is obsoleted, so there should be no problem as GO:0019836 is already a descendent of GO:0001906 cell killing. It is thus already in the GO in the way specified by PAMGO.

Thanks again to Candace and PAMGO for their hard work on this proposal.

-- Alex

Original comment by: addiehl

[gocentral]

Logged In: YES user_id=451873

A version of this file with these changes included is available here:

http://www.ebi.ac.uk/~jane/gene\_ontology.obo

This will be committed on Monday 31st Jan, pending any further changes.

Original comment by: jl242

[gocentral]

• status: open --> closed

Original comment by: jl242

[gocentral]

Logged In: YES user_id=451873

Implemented the changes outlined in GO-PAMGO-fnlfnl-tree.doc and GO-PAMGO-fnlfnl-def.doc - new ids listed below. There were a few other minor changes too:

Name change: cell invasion -> entry into host cell

Name change: viral host defense evasion -> evasion of host defense by virus

Name change: competition with another, non-host, organism -> competition with other, non-host, organism

Name change: interaction between host and another organism -> interaction between host and other organism

Name change: viral entry -> viral entry into host cell

Also made 'viral entry' a child of 'host cell invasion' like this:

entry into host cell ---[i] viral entry into host cell

Changed the definition of 'interaction between organisms' slightly from:

The processes by which an organism has an observable effect on an organism of a different species.

to

The processes by which an organism effects an organism of a different species.

Added extra term term: 'interaction with other, non-host organism' as child of 'symbiosis', like this:

interaction between organisms ---[i] symbiosis ------[i] interaction between host and another organism ------[i] interaction with other, non-host organism

New ids:

id: GO:0042000 name: translocation of peptides or proteins into host

id: GO:0043298 name: symbiotic interaction with other, non-host organism

id: GO:0044000 name: movement within host

id: GO:0044001 name: migration within host

id: GO:0044002 name: acquisition of nutrients from host

id: GO:0044003 name: modification of host morphology or physiology

id: GO:0044004 name: disruption of host cells

id: GO:0044005 name: induction in host of a tumor, nodule, or growth

id: GO:0044006 name: induction in host of a tumor, nodule, or growth containing transformed cells

id: GO:0044007 name: dissemination or transmission of an organism from a host

id: GO:0044008 name: dissemination or transmission of an organism from a host by a vector

id: GO:0044009 name: viral transmission by a vector

id: GO:0044010 name: single-species biofilm formation

id: GO:0044011 name: single-species biofilm formation on inanimate substrate

id: GO:0044399 name: multi-species biofilm formation

id: GO:0044400 name: multi-species biofilm formation on inanimate substrate

id: GO:0044401 name: multi-species biofilm formation in or on host organism

id: GO:0044402 name: competition with other, non-host, organism

id: GO:0044403 name: symbiosis, mutualism through parasitism

id: GO:0044404 name: symbiotic interaction between host and other organism

id: GO:0044405 name: recognition of host

id: GO:0044406 name: adhesion to host

id: GO:0044407 name: single-species biofilm formation in or on host organism

id: GO:0044408 name: growth on or near host surface

id: GO:0044409 name: entry into host

id: GO:0044410 name: entry into host through natural portals

id: GO:0044411 name: entry into host through host barriers

id: GO:0044412 name: growth within host

id: GO:0044413 name: avoidance of host defenses

id: GO:0044414 name: suppression of host defenses

id: GO:0044415 name: evasion of host defenses

id: GO:0044416 name: induction of host defense response

id: GO:0044417 name: translocation of molecules into host

id: GO:0044418 name: translocation of DNA into host

id: GO:0044419 name: interaction between organisms

Let me know if you spot any problems,

thanks,

jane

Original comment by: jl242

[gocentral]

• milestone: --> TermGenie
• assigned_to: nobody --> jl242
• status: closed --> closed-accepted

Original comment by: jl242