Continuing Discussion of Response and Detection Terms
Alexander Diehl
adiehl at informatics.jax.org
Wed Mar 30 10:52:06 PST 2005
In late September of last year I made a
SourceForge entry called "Response and Detection
Terms"
(https://sourceforge.net/tracker/?func=detail&aid=1031159&group_id=36855&atid=440764),
that dealt with redundancies, missing terms, and
logical and structural problems in GO terms
involving responses to biotic stimuli. Some of
my proposed changes are non-controversial,
others, however, have generated some concern.
In response to this proposal Suparna Mundodi of
TAIR assembled a group of plant-pathogen experts
to review the proposal and offer a
counter-proposal, which was added to the original
SourceForge entry in early March. I welcome
their involvement in this effort. However,
though I agree with some of the new terms
proposed, I have found a number of problems with
this counter-proposal which I feel render it
inconsistent with the proper logical formation
and established principles of the GO as well as
with the now-accepted term additions and changes
put forth by the PAMGO group (Plant-Associated
Microbe Gene Ontology Interest Group).
I've assembled an email digest below of emails
primarily between myself and members of PAMGO
regarding the proposed term changes put forth by
Suparna and the plant-pathogen experts. I turned
to PAMGO first because they are comprised of
another group of plant biologists who have
studied and made useful additions to the GO, and
I felt that they could offer useful opinions
regarding the development of this section of the
GO.
I am trying to generate additional discussion
regarding this issue on the GO pathogenesis
mailing list (pathogenesis at geneontology.org). I
would urge people who are interested in this
topic to reread my original proposal (attached),
as well as read the counter-proposal (attached)
and the assembled emails below. I would also
request that postings regarding this topic be
made directly to the pathogenesis mailing list,
and cross-posted to the individual email address
above, as some people may not yet be subscribed
to the mailing list.
I will give a summary of responses to this issue
at the GO Consortium meeting at the end of next
week, but discussion can continue on as long as
necessary. Eventually, I will revise my original
proposal to match any consensus that arises. If
necessary, the non-controversial changes will be
split out, and the difficult issues will be
covered at the GO Content Meeting on Defense
Responses and Immunology, which will be held in
October or November of this year.
Thank you for your interest,
Alex
Date: Fri, 4 Mar 2005 15:56:00 -0800 (PST)
From: Suparna Mundodi <smundodi at acoma.stanford.edu>
To: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: response to terms proposal
Status:
Hi Alex,
I am sure you noticed the sourceforge item on response terms has been
updated from TAIR. That means, I was able to organize the meeting between
whole bunch of plant-pathogen experts to get feedback on your proposal. It
was definitely very helpful talking to them and they had very clear idea
on what they wanted.
Anyway, I wanted to let you know that I will be gone any day now on my
maternity leave. Just wanted to warn you on this. You may not hear from me
often now.
Thanx for working on the structure.
Cheers,
Suparna
****************************************************************************
From SourceForge Item 1031159
Comment By: Suparna Mundodi (smundodi)
Date: 2005-03-04 15:41
Message:
Logged In: YES
user_id=614564
Hi Alex,
As promised, I had organized a meeting with
plant-pathogen interaction experts few weeks ago
and we discussed the issues raised by you.
Following the meeting, we also had the email
debate to finalize the tree that suits the plant
community. People involved in resolving this
issue are:
Jeff Dangl – University of North Carolina
Fred Ausabel- Harvard University
Xinnian Dong – Duke University
Shauna Somerville – Stanford
Barbara Baker – UC Berkeley
Richard Michelmore – UC Davis
Pamela Ronald- UC Davis
Based on the feedback I received from all these
experts, I have generated a simple GO tree with
the terminologies that plant people would need.
* We are OK with the obsoletion of the following terms:
GO:0009680 ; response to non-pathogenic bacteria
GO:0009681 ; detection of non-pathogenic bacteria
GO:0042830 ; defense response to pathogenic bacteria
GO:0009618 ; response to pathogenic bacteria
GO:0009621 response to pathogenic fungi
GO:0042831 defense response to pathogenic fungi
GO:0016047 detection of parasitic fungi
GO:0042833 response to pathogenic protozoa
GO:0043019 response to pathogenic insects
* We do not want to obsolete:
GO:0009613 response to pest, pathogen or parasite
GO:0042828 response to pathogen
* Renaming:
We would like to keep response to external biotic
stimulus: Here external is more straight forward
than foreign. Thus do not need to rename it.
* GO:0042829 (formerly "defense response to
pathogen") I propose that GO:0042829 be renamed
"defense response to potential pathogen"
Our response: Prefer the ‘defense response to
pathogen’ term. If needed, include the word
potential in the definition. Do not change the
name to potential. It doesn’t make it any clear.
So, here is the structure and the terms we would
like to have in GO.I have also attached the tree
file to this mail just in case it is hard to see
the parentage and indenting. Let me know if there
is problem with that.
Revised GO tree from Plant-pathogen experts:
Response to external biotic stimulus GO:0043207
--------- Response to bacteria GO:0009617
---------Response to symbiotic bacteria GO:0009609
---------response to bacteria-associated molecule (new)
---------response to bacterial pathogen (new)
--------- Defense response to bacteria GO:0042742
--------- Response to fungi GO:0009620
---------response to fungus-associated molecule (new)
---------response to fungal pathogen (new)
--------- Defense response to fungi GO:0050832
--------- Response to Oomycetes (new)
---------response to oomycetes-associated molecule (new)
---------response to oomycete pathogen (new)
---------defense response to oomycete (new)
---------Response to pests, pathogen and parasites GO:0009613
--------- Response to pathogen GO:0042828
---------response to bacterial pathogen (new)
---------defense response to Bacteria GO:0042742
---------response to fungal pathogen (new)
--------- defense response to fungi GO:0050832
---------response to oomycete pathogen (new)
---------defense response to oomycete (new)
---------response to viral pathogen (new)
--------- response to host-derived signal (new)
---------response to
pathogen-induced salicylic acid (new)
---------response to
pathogen-induced jasmonic acid (new)
---------response to pathogen-induced ethylene (new)
---------response to
pathogen-induced pectin molecules (new)
---------response to
pathogen-induced oxygen radicals (new)
--------- response to symbionts,
commensals, mutualists or saprophytes (new)
---------response to bacterial symbionts,
commensals, mutualists or saprophytes (new)
--------- response to symbiotic bacteria GO:0009609
--------- response to fungal symbionts,
commensals, mutualists or saprophytes (new)
--------- Response to virus GO:0009615
---------response to viral pathogen (new)
--------- Response to parasitic plants (new)
---------Response to Insects GO:0009625
---------response to insect-derived molecules (new)
---------Response to nematode GO:0009624
Definitions and explanations:
1. Response to bacterial pathogen: A change in
state or activity of an organism (in terms of
movement, secretion, enzyme production, gene
expression, etc.) as a result of the perception
of bacterial pathogen.
2. Response to fungal pathogen: A change in state
or activity of an organism (in terms of movement,
secretion, enzyme production, gene expression,
etc.) as a result of the perception of fungal
pathogen.
* We need the terms, ‘response to bacterial
pathogen’ etc in order to be able to annotate
genes when the experiment is set up to have an
interaction between host plant and bacterial
pathogen. The results of these experiments need
to be annotated to ‘response to bacterial
pathogen’. Thus plant community would like to
keep these terms in the tree. However, it is
possible to annotate these genes to 2 different
terms, ie. ‘response to bacteria’ and ‘response
to pathogen’ to give the similar meaning. But at
the time of looking at over-represented processes
in a set of genes and other type of quantitative
analysis, the use of the two broader terms versus
one specific term would have a significant impact
and would help in functionally categorizing these
genes.
3. response to bacteria-associated molecule
Definition: A change in state or activity of an
organism (in terms of movement, secretion, enzyme
production, gene expression, etc.) as a result of
the perception of bacterial associated molecules
such as peptides derived from bacterial flagellin.
4. response to fungal-associated molecule
Definition: A change in state or activity of an
organism (in terms of movement, secretion, enzyme
production, gene expression, etc.) as a result of
the perception of fungal associated molecules
such as chito-octomer oligosaccharide.
5. response to oomycete-associated molecule
Definition: A change in state or activity of an
organism (in terms of movement, secretion, enzyme
production, gene expression, etc.) as a result of
the perception of oomycete associated molecules.
6. response to host-derived signal
Definition: A change in state or activity of an
organism (in terms of movement, secretion, enzyme
production, gene expression, etc.) as a result of
the perception of host-derived signals such as
host cell wall fragments, salicylic acid, or
nitric oxide.
7. response to pathogen-induced salicylic acid (new)
Definition: A change in state or activity of an
organism (in terms of movement, secretion, enzyme
production, gene expression, etc.) as a result of
the perception of pathogen-induced salicylic acid.
These terms will also become children terms of
‘response to xxx’ terms. For example, ‘response
to pathogen-induced salicylic acid’ will also
become a child of ‘response to salicylic acid
stimulus’ GO:0009751
Other definitions for host-derived signals are
similar to the above and they will have
additional parentage under ‘response to xxx
stimulus’
* Currently the response terms to salicylic acid,
ethylene, jasmonic acid are all under 'response
to endogenous stimulus' (except for oxygen
radical). Now that we are presenting these terms
under an additional category of 'response to
host-derived signals', we will have to
instantiate the response to salicylic acid, etc.
terms to make a specific case for responding to
host-derived signals. For example, in order to
distinguish the 'pathogen-induced ethylene
stimulus' from 'ripening or wound-induced
ethylene stimulus', we have created the term
'response to pathogen-induced ethylene' to be a
child of 'response to ethylene' as well as
'response to host-derived signals'.
8. response to symbionts, commensals, mutualists or saprophytes
Definition: A change in state or activity of an
organism (in terms of movement, secretion, enzyme
production, gene expression, etc.) as a result of
the perception of symbionts, commensals,
mutualists or saprophytes.
9. response to bacterial symbionts, commensals, mutualists or saprophytes:
Definition: A change in state or activity of an
organism (in terms of movement, secretion, enzyme
production, gene expression, etc.) as a result of
the perception of bacterial symbionts,
commensals, mutualists or saprophytes.
10. Response to parasitic plant
Definition: A change in state or activity of an
organism (in terms of movement, secretion, enzyme
production, gene expression, etc.) as a result of
the perception of parasitic plant.
Hope to see these terms in GO soon.
Suparna
****************************************************************************
From SourceForge Item 1031159
>Comment By: Alexander Diehl (aledie)
Date: 2005-03-16 12:12
Message:
Logged In: YES
user_id=865072
I welcome the input from Suparna and the plant-pathogen experts
and thank them for their time and effort. As the paradigm for
handling "pathogens" and "defense responses" appears to differ in
the plant community from that that has evolved from studies of
innate immunity and defense responses in mammalian and other
animal systems in recent years, I am seeking additional outside
opinion, and trying to decide how best to resolve the differences.
My main objective remains to create a set of GO terms that work
both for plants and animals and avoid the use of sensu terms
simply because the two communities are locked into separate
paradigms regarding defense responses and the use of the word
"pathogen." Such sensu terms would be self-defeating for the GO.
I will prepare a discussion of the issue for the GO Consortium
meeting, as well as look for areas of consensus where changes
can be implemented directly.
-- Alex
****************************************************************************
My letter of 3/16/05 to Candace Collmer
To: Candace_Collmer
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: We need an opinion from PAMGO
Cc:
Bcc:
Attachments: [added as attachments to main email message]
Hi Candace,
I am writing to ask you or another PAMGO member
to review the response posted by Suparna to my SF
entry regarding the response and detection terms
in the GO
(https://sourceforge.net/tracker/?func=detail&aid=1031159&group_id=36855&atid=440764).
As I wrote in my SourceForge proposal:
The proposal is a major reorganization of "response to . . .",
"detection of . . .", and "defense response to . . ." terms.
The major goal is to systematize the terms describing how a
given organism responds to the stimulus provided by other
organisms, so that the types of other organisms providing
the stimulus are treated more or less the same, with basic
"response to . . .", "detection of . . .", and "defense
response to . . ." terms. The perspective for the terms is
that of the principal organism doing the responding; thus
defense responses to particular organisms, bacteria, fungi,
protozoa, and viruses (not quite an organism, of course),
are grouped under "defense response to potential pathogen,"
both to please those who would like to have the concept of
pathogenicity retained in the GO, and those who believe that
pathogenicity ought to be viewed from the perspective of the
threatened organism.
Suparna writes that she has had a meeting with a
number of plant pathogen experts, and has
formulated a counter-proposal based on that
discussion. She appears to agree with the basic
idea of systematizing the terms, and accepts that
a number of terms should be obsoleted. However,
she proposes a new set of terms such as "response
to bacterial pathogen," "response to fungal
pathogen," etc., and then places the "defense
response to" terms under these terms.
--Response to pests, pathogen and parasites GO:0009613
----Response to pathogen GO:0042828
------response to bacterial pathogen (new)
--------defense response to Bacteria GO:0042742
------response to fungal pathogen (new)
--------defense response to fungi GO:0050832
I have major problems with this structure. As
was discussed at the GO content meaning, a
response to some type of pathogen, based on its
presumed danger to the responding organism, is a
defense response. Thus I consider "response to
bacterial pathogen" and "defense response to
bacteria" to be redundant, and the second term to
be the correct formulation. Also, as I have
mentioned many times, animals and plants are
constantly making little defense responses to
bacteria in their environment prior to the
appearance of any pathology caused by the
bacteria, and for a healthy animal or plant, many
types of bacteria will not cause disease, and not
be considered pathogens by the human observer
even though such defense responses are being made
against them. Thus, I am not happy at all with
this proposed structure, and I stand by my
original proposal, which looks like this
(considerably simplified):
------------% defense response ; GO:0006952
--------------% defense response to potential
pathogen ; GO:0042829 (as renamed)
----------------% defense response to bacteria ; GO:0042742
----------------% defense response to fungi ; GO:0050832
----------------% defense response to protozoa ; GO:0042832 (as renamed)
----------------% defense response to virus ; GO:NewTerm
Suparna states that her structure is what the
plant community wants. I have problems with it
for the reasons stated and more, but I would like
your opinion or that of someone else in PAMGO,
since I need formulate a sensible response. My
main objective is to have a set of common GO
terms that work both for plants and animals and
avoid the use of sensu terms simply because the
two communities are locked into separate
paradigms regarding defense responses and
pathogens.
I would recommend reading through both my
original proposal and Suparna's response. I have
attached these as documents to this email for
your convenience, although you may look on
SourceForge as well. I know that this may take
some time.
Thank you very much.
Alex
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
To: adiehl <adiehl at informatics.jax.org>
From: David Bird <david_bird at ncsu.edu>
Subject: Response to pests, pathogen and parasites GO:0009613
Date: Thu, 17 Mar 2005 10:19:05 -0500
X-PMX-Version: 4.7.1.128075, Antispam-Engine:
2.0.3.0, Antispam-Data: 2005.3.17.6
X-Spam-Status: No, Hits=7%
X-Spam-Level: IIIIIII
Status:
Hi Alex,
I am loosely associate with the PAMGO group, and
have an interest in nematode-plant interactions.
I have been following some of the discussion of
terms on the entries: "Response to pests,
pathogen and parasites GO:0009613" and have a
couple of comments. I believe that Candice an
John Collmer will be replying later with no doubt
more coherent thoughts.
Part of the complexity comes from the fact that
plants interact with a wide range of other
organisms in a wide range of ways (mutualism,
parasitism, etc), and yet the underlying
mechanisms might be shared. I suspect that the
outcome and complexity of the plant-microbe
signaling is actually an important determinant in
whether a particular interaction will lead to
pathology. As an example, a mutualistic symbiont
such as rhizobia still needs to convince the
plant that it is not a pathogen. Perhaps it does
elicit a defense response, but then suppress
that. It certainly interacts with the plant to
elicit a response that is not defense related
(nodule formation, for example). A rhizobium that
reneges on its "agreement" with the plant to fix
nitrogen (and hence becoming a pathogen) is
subsequently detected by the plant, which then
mounts a response. We now know that parasitic
symbionts such as root-knot nematodes also signal
the plant through the same receptors and signal
transduction pathways used by rhizobia. A defense
response is presumably suppressed and the
pathogen interacts with the plant to elicit a
response that is not defense related (giant cell
formation). But there also can be a defense
response too (leading to host resistance). Thus
"response to bacterial pathogen" and "defense
response to bacteria" are most definitely not
redundant and the second term is not the
"correct formulation."
Best regards
dave
David Bird
Director, Center for the Biology of Nematode Parasitism
Professor of Plant Pathology
Box 7253, NCSU
Raleigh NC 27695
+1 (919) 515-6813
****************************************************************************
X-Sender: ccollmer at henry.wells.edu
Date: Thu, 17 Mar 2005 10:30:00 -0500
To: Alexander Diehl <adiehl at informatics.jax.org>
From: Candace Collmer <ccollmer at wells.edu>
Subject: Re: We need an opinion from PAMGO
Status:
Hi Alex --
will work on this today -- as soon as I finish
some stuff ahead of it in line..... I've also
sent it out to other PAMGO members, and I think
Dave Bird has already emailed his response to
you. -Candace
At 10:26 AM 3/16/2005 -0500, you wrote:
[deleted for brevity, direct quote of my original email above]
---------------------------
Candace W. Collmer
Professor of Biology
Wells College
Aurora, NY 13026
phone: 315-364-3271
fax: 315-364-3464
email: ccollmer at wells.edu
****************************************************************************
X-WebMail-UserID: ccollmer
Date: Thu, 17 Mar 2005 16:34:17 -0500
Sender: ccollmer <ccollmer at wells.edu>
From: ccollmer <ccollmer at wells.edu>
To: adiehl at informatics.jax.org
Cc: arc2 at cornell.edu, ccollmer at wells.edu
X-EXP32-SerialNo: 00003289
Subject: some PAMGO thoughts, more coming
Status:
Hi Alex --
Alan and I are still going to work on this later today (or else on the
weekend, if we don't get finished), but I had some more discussion with David
Bird, and wanted to pass that on to you as I'm not sure whether he did that.
Will send more thoughts later.
You can see my question to him, and his response to me, below. -Candace
>===== Original Message From David Bird <david_bird at ncsu.edu> =====
I don't know if Alex's terms cover what I intend (I really just don't
know). But I do feel that the Suparna terms are clear, and seem to deal
with what I think reflects the biology best. The reality is that plants
(and animals too, I think) will have non-defense responses to
pathogens. But I will defer to you as you certainly have thought about
this much more than have I.
dave
David Bird
Director, Center for the Biology of Nematode Parasitism
Professor of Plant Pathology
Box 7253, NCSU
Raleigh NC 27695
+1 (919) 515-6813
On Mar 17, 2005, at 3:02 PM, ccollmer wrote:
> It seems to me that one of the main points of contention is
> whether, as
> Alex says, "response to bacterial pathogen" [notice the "bacterial
> pathogen"
> here] is really always the same as "defense response to bacteria."
> Suparna et
> al, on the other hand, think there is a need to have 2 separate terms,
> as seen
> in the tree below (from Suparna et al) [I added the numbers just in
> case the
> indentations did not come through in the email -- each number relates
> to a
> different amount of indenting]:
>
> 1Response to external biotic stimulus GO:0043207
> ---------2Response to bacteria GO:0009617
> ---------3Response to symbiotic bacteria GO:0009609
> ---------3response to bacteria-associated molecule (new)
> ---------3response to bacterial pathogen (new)
> ---------4Defense response to bacteria GO:0042742
>
> Now look at what Alex proposes, where those 2 terms are combined, but
> there
> are also separate terms related to symbiotic interactions that are
> separate
> from those. I've written below that part from Alex's tree. Would his
> tree,
> as detailed here, take care of the kind of concerns you have?
>
> response to bacteria ; GO:0009617
> --% defense response to bacteria ; GO:0042742
> ----% antibacterial humoral response ; GO:0019731 +
> ----% defense response to Gram-negative bacteria ; GO:0050829 +
> ----% defense response to Gram-positive bacteria ; GO:0050830 +
> ----% defense response to bacteria, incompatible interaction ;
> GO:0009816 (as
> renamed)
> ----% induced systemic resistance ; GO:0009682
> ----% male-specific defense response to bacteria c +
> --< detection of bacteria ; GO:0016045
> ----% detection of pathogenic bacteria ; GO:0009598
> ------< detection of bacterial lipoprotein ; GO:0042494 +
> ----% detection of symbiotic bacteria ; GO:0009604
> --% response to symbiotic bacteria ; GO:0009609
> ----< detection of symbiotic bacteria ; GO:0009604
>
>
----------------------------
Candace W. Collmer, Ph.D.
Professor of Biology
Wells College
Aurora, NY 13026
phone: 315-364-3271
email: ccollmer at wells.edu
****************************************************************************
From SourceForge Item 1031159
Comment By: Michelle Gwinn (mlgwinn)
Date: 2005-03-18 11:05
Message:
Logged In: YES
user_id=482944
Hi Alex and Suparna,
Wow, and we thought the "interaction with other organisms" subtree was
hard to work out!
This one is really complex and difficult. I
think that there is a lot here still
to resolve. I feel it is definately going to require a face-to-face meeting to
work out. I think its a good idea to raise the issue at the GOC meeting,
but I have a feeling (unless a lot of time is available for discussion of this
at the GOC meeting) that it will require a content meeting to work out - one
of which has not yet been scheduled for this year, as far as I know.
I look forward to seeing both of you in Pasadena.
Michelle
****************************************************************************
Subject: RE: PAMGO: opinion needed from GO
Date: Fri, 18 Mar 2005 11:09:46 -0500
Thread-Topic: PAMGO: opinion needed from GO
Thread-Index: AcUq+inZ/Hsw5nKkQaOEXB/WYmccIgA2DG6L
From: "Gwinn, Michelle L." <MLGwinn at tigr.org>
To: "ccollmer" <ccollmer at wells.edu>,
"xt_contact Collmer, Alan" <arc2 at cornell.edu>,
<adiehl at informatics.jax.org>
Cc: "xt_contact Collmer, Alan" <arc2 at cornell.edu>, <bmtyler at vt.edu>,
<david_bird at ncsu.edu>, <jeremy at genome.wisc.edu>,
"Hannick, Linda" <lhannick at tigr.org>, <perna at ahabs.wisc.edu>,
<ralph_dean at ncsu.edu>, "Buell, Robin" <rbuell at tigr.org>,
"xt_contact Setubal, Joao" <setubal at vbi.vt.edu>, <sutripa at vbi.vt.edu>,
<thomas_mitchell at ncsu.edu>, "Gwinn, Michelle L." <MLGwinn at tigr.org>
Status:
Hi Candace, Alex, and everyone,
This issue is really complex and I feel it will
definately take a face-to-face meeting to work
out. I started to formulate specific responses
and suggestions to the two proposals but found
that it's getting rather too complex for the
SF/email format - at least I think so.
So, I just posted a note to the SF item saying
that I think it will likely need a content
meeting for this topic. That would allow reps
from all interested parties the opportunity to
get in one room and work through this item by
item.
Michelle
****************************************************************************
My first letter of 3/18/05 to David Bird
To: David Bird <david_bird at ncsu.edu>
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: Response to pests, pathogen and parasites GO:0009613
Cc: Candace_Collmer, MLGwinn at tigr.org
Bcc:
Attachments:
David,
Thanks for your response. I'm glad to see that
there are some other people paying attention to
this discussion.
I agree with much of what you say regarding the
range and complexity of plant-microbe or
-nematode interactions. My main problem with the
use of the word "pathogen," as part of GO terms
because of the loose way the word is used and the
resulting true-path violations that can result.
The word "pathogen" means of course "something
that causes disease in another organism." As you
point out, particular organisms are pathogens in
some situations and not in others. Furthermore,
the host response to a bacteria (or whatever) may
well vary depending on the site of interaction
and contextual information such as the detection
of specific bacterial products.
What's more, calling something a pathogen is a
very human perspective. In mammals, we do not
normally consider normal gut dwelling bacteria in
healthy individuals as pathogens, yet should that
bacteria get into a cut on the skin (pick a
mechanism), it will be in most cases instantly
killing through any of a variety of mechanisms
such as complement activation, phagocytosis by
macrophages, or superoxide production. A defense
response has been made, even though the bacteria
itself is not one humans normally consider a
pathogen, and even though the bacteria itself
lacks adaptations to its new environment. And,
in an immunocompromised individual, the normally
non-pathogenic bacteria might become pathogenic.
In the DAG structure below there is therefore a true-path violation.
--Response to pests, pathogen and parasites GO:0009613
----% Response to pathogen GO:0042828
------% response to bacterial pathogen (new)
--------% defense response to Bacteria GO:0042742
A gene product shown to be involved in the
killing of the gut derived bacteria in the
skin-wound environment would logically be
annotated to GO:0042742 defense response to
bacteria. However because this DAG structure
states that GO:0042742 is a type of "response to
bacterial pathogen" a true-path violation occurs
because we have demonstrated that a defense
response to bacteria can occur which is not a
response to a bacterial pathogen.
One might argue, well, when based in the in the
context of the wound, that the responding
organism decides that the infecting organism is a
pathogen, even though we, the human observers,
would not call as such since no disease is being
induced in the responding organism. If our
criteria for what we call a pathogen then becomes
"any organism that the body makes a defense
response to," we've stepped away from the
original, if vague, definition of "pathogen." In
fact, we could probably put a million different
kinds of normally non-pathogenic bacteria into a
wound, and decide they are all really pathogenic,
at which point the distinction between pathogenic
and non-pathogenic bacteria is meaningless, and
there is no point is having any GO terms such as
"response to bacterial pathogen" or even
"response to pathogen."
Moreover, even in our gut, the innate and
adaptive immune systems are always at work to
maintain homeostasis of the bacterial
populations. Although some gut host responses
are symbiotic, many are actually low-level
defense responses which occur at every moment
following colonization of the gut following
birth, without evident pathology, but without
which we could not survive.
All in all, the distinction between between
pathogens and non-pathogens seems rather
arbitrary, and not one I think is useful in the
GO. Plants and animals respond to molecular
signals from their environment, from their
infecting organisms -- they don't read some
human-approved list what is or is not a
"pathogen," and they don't necessarily wait for
the onset of human-detectable symptoms to mount a
defense response.
However, I do realize that for some users of the
GO, the use of the word "pathogen" provides a
sense of the familiar and thus in my
reformulation of the terms I proposed using the
GO term GO:0042829 with the changed name "defense
response to potential pathogen," to reflect the
fact the response is based on the sensing of
molecular signals indicative of the fact an
organism which could potentially cause disease is
present. This way gene products involved in a
defense response to, say, LPS, from a
non-pathogenic bacteria used in an experimental
system can be annotated without a true-path
violation.
------------% defense response ; GO:0006952
--------------% defense response to potential
pathogen ; GO:0042829 (as renamed)
----------------% defense response to bacteria ; GO:0042742
----------------% defense response to fungi ; GO:0050832
----------------% defense response to protozoa ; GO:0042832 (as renamed)
----------------% defense response to virus ; GO:NewTerm
(other potential pathogen types may be added as needed)
I don't want it hard-coded into the GO that all
bacteria are pathogens or all fungi are pathogens
simply because a defense response can be made
against them, which is what the first DAG above
implies.
There is still the proposed term, "response to
bacterial pathogen," and the implied term,
"response to nematode pathogen." If such
responses are not defense responses, then why and
how is the plant or animal distinguishing between
pathogens and non-pathogens? I don't know
anything about giant cell formation in response
to nematodes or the role of giant cells in plant
physiology. What are the consequences to the
plant if it can't form the giant cells? Are
there nematodes which are not pathogenic but
which still possess the ability to form giant
cells? Is the existing GO term GO:0009624
response to nematode insufficient for annotation
of a gene product which is involved in giant cell
formation in response to a molecular signal from
a nematode, pathogen or not?
I apologize for the length of this message, but I
feel the distinctions I am trying to explain are
sufficiently subtle to warrant a longer
explanation. I also apologize for using animal
systems as explanations, but GO terms need to
work for both plants and animals.
-- Alex
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
Cc: ccollmer Collmer <ccollmer at wells.edu>
From: David Bird <david_bird at ncsu.edu>
Subject: Re: Response to pests, pathogen and parasites GO:0009613
Date: Fri, 18 Mar 2005 14:59:24 -0500
To: Alexander Diehl <adiehl at informatics.jax.org>
X-PMX-Version: 4.7.1.128075, Antispam-Engine:
2.0.3.0, Antispam-Data: 2005.3.18.11
X-Spam-Status: No, Hits=7%
X-Spam-Level: IIIIIII
Status:
Dear Alex,
I am new to GO syntax, and so I greatly
appreciate your patience. GO is an excellent
resource that has been very useful to my
research, and I am keen to make sure that the
terminology works.
>All in all, the distinction between between
>pathogens and non-pathogens seems rather
>arbitrary, and not one I think is useful in the
>GO. Plants and animals respond to molecular
>signals from their environment, from their
>infecting organisms -- they don't read some
>human-approved list what is or is not a
>"pathogen," and they don't necessarily wait for
>the onset of human-detectable symptoms to mount
>a defense response.
I agree 100% with this. The term pathogen has no logical place in GO.
>There is still the proposed term, "response to
>bacterial pathogen," and the implied term,
>"response to nematode pathogen." If such
>responses are not defense responses, then why
>and how is the plant or animal distinguishing
>between pathogens and non-pathogens? I don't
>know anything about giant cell formation in
>response to nematodes or the role of giant cells
>in plant physiology. What are the consequences
>to the plant if it can't form the giant cells?
>Are there nematodes which are not pathogenic but
>which still possess the ability to form giant
>cells? Is the existing GO term GO:0009624
>response to nematode insufficient for annotation
>of a gene product which is involved in giant
>cell formation in response to a molecular signal
>from a nematode, pathogen or not?
So let's drop "pathogen." But not all responses
are defense responses. In the case of my
nematodes (and I suspect other organisms
interacting with plants and animals) there can be
a defense response mounted by the host, but it is
far from being the primary interaction. There is
clearly a molecular conversation between the
plant and nematode prior to them making direct
physical contact. It is likely that the
interaction continues during penetration of the
host, and up to the induction of giant cells
24-36 hours later. If the plant is going to mount
a defense response (and this is conditioned in
some instances by single Medelian loci), this
happens at this point. I think "response to a
nematode" is sufficient.
dave
David Bird
Director, Center for the Biology of Nematode Parasitism
Professor of Plant Pathology
Box 7253, NCSU
Raleigh NC 27695
****************************************************************************
My second letter of 3/18/05 to David Bird
To: David Bird <david_bird at ncsu.edu>
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: Response to pests, pathogen and parasites GO:0009613
Cc: Candace_Collmer
Bcc:
Attachments:
Dear David,
Thanks again for your email. I absolutely agree
that not all responses are defense responses.
Indeed, I would welcome such terms as "response
to nematode, giant cell formation," if an expert
such as yourself would find a use for such a term
in annotating particular gene products involved
in this type of specialized plant response.
With your permission, I would like to collect and
post email exchanges on this issue at some point
on the relevant SourceForge entry and as the
start of a more public discussion on the GO
pathogenesis mailing list, which I would cross
post to you.
Thanks,
Alex
>Dear Alex,
>I am new to GO syntax, and so I greatly
>appreciate your patience. GO is an excellent
>resource that has been very useful to my
>research, and I am keen to make sure that the
>terminology works.
>
>>All in all, the distinction between between
>>pathogens and non-pathogens seems rather
>>arbitrary, and not one I think is useful in the
>>GO. Plants and animals respond to molecular
>>signals from their environment, from their
>>infecting organisms -- they don't read some
>>human-approved list what is or is not a
>>"pathogen," and they don't necessarily wait for
>>the onset of human-detectable symptoms to mount
>>a defense response.
>
>I agree 100% with this. The term pathogen has no logical place in GO.
>
>>There is still the proposed term, "response to
>>bacterial pathogen," and the implied term,
>>"response to nematode pathogen." If such
>>responses are not defense responses, then why
>>and how is the plant or animal distinguishing
>>between pathogens and non-pathogens? I don't
>>know anything about giant cell formation in
>>response to nematodes or the role of giant
>>cells in plant physiology. What are the
>>consequences to the plant if it can't form the
>>giant cells? Are there nematodes which are not
>>pathogenic but which still possess the ability
>>to form giant cells? Is the existing GO term
>>GO:0009624 response to nematode insufficient
>>for annotation of a gene product which is
>>involved in giant cell formation in response to
>>a molecular signal from a nematode, pathogen or
>>not?
>
>So let's drop "pathogen." But not all responses
>are defense responses. In the case of my
>nematodes (and I suspect other organisms
>interacting with plants and animals) there can
>be a defense response mounted by the host, but
>it is far from being the primary interaction.
>There is clearly a molecular conversation
>between the plant and nematode prior to them
>making direct physical contact. It is likely
>that the interaction continues during
>penetration of the host, and up to the induction
>of giant cells 24-36 hours later. If the plant
>is going to mount a defense response (and this
>is conditioned in some instances by single
>Medelian loci), this happens at this point. I
>think "response to a nematode" is sufficient.
>dave
>
>David Bird
>Director, Center for the Biology of Nematode Parasitism
>Professor of Plant Pathology
>Box 7253, NCSU
>Raleigh NC 27695
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
From: David Bird <david_bird at ncsu.edu>
Subject: Re: Response to pests, pathogen and parasites GO:0009613
Date: Fri, 18 Mar 2005 15:48:28 -0500
To: Alexander Diehl <adiehl at informatics.jax.org>
X-PMX-Version: 4.7.1.128075, Antispam-Engine:
2.0.3.0, Antispam-Data: 2005.3.18.12
X-Spam-Status: No, Hits=7%
X-Spam-Level: IIIIIII
Status:
> Indeed, I would welcome such terms as "response
>to nematode, giant cell formation," if an expert
>such as yourself would find a use for such a
>term in annotating particular gene products
>involved in this type of specialized plant
>response.
The PAMGO has a proposal pending to really
support the plant-microbe GO development, and a
small component will be some of the
nematode-plant (and animal too) terms (such as
"response to nematode, giant cell formation").
Fingers crossed for the funding.
>With your permission, I would like to collect
>and post email exchanges on this issue at some
>point on the relevant SourceForge entry and as
>the start of a more public discussion on the GO
>pathogenesis mailing list, which I would cross
>post to you.
Sure, that is fine.
dave
>>Dear Alex,
>>I am new to GO syntax, and so I greatly
>>appreciate your patience. GO is an excellent
>>resource that has been very useful to my
>>research, and I am keen to make sure that the
>>terminology works.
>>
>>>All in all, the distinction between between
>>>pathogens and non-pathogens seems rather
>>>arbitrary, and not one I think is useful in
>>>the GO. Plants and animals respond to
>>>molecular signals from their environment, from
>>>their infecting organisms -- they don't read
>>>some human-approved list what is or is not a
>>>"pathogen," and they don't necessarily wait
>>>for the onset of human-detectable symptoms to
>>>mount a defense response.
>>
>>I agree 100% with this. The term pathogen has no logical place in GO.
>>
>>>There is still the proposed term, "response to
>>>bacterial pathogen," and the implied term,
>>>"response to nematode pathogen." If such
>>>responses are not defense responses, then why
>>>and how is the plant or animal distinguishing
>>>between pathogens and non-pathogens? I don't
>>>know anything about giant cell formation in
>>>response to nematodes or the role of giant
>>>cells in plant physiology. What are the
>>>consequences to the plant if it can't form the
>>>giant cells? Are there nematodes which are not
>>>pathogenic but which still possess the ability
>>>to form giant cells? Is the existing GO term
>>>GO:0009624 response to nematode insufficient
>>>for annotation of a gene product which is
>>>involved in giant cell formation in response
>>>to a molecular signal from a nematode,
>>>pathogen or not?
>>
>>So let's drop "pathogen." But not all responses
>>are defense responses. In the case of my
>>nematodes (and I suspect other organisms
>>interacting with plants and animals) there can
>>be a defense response mounted by the host, but
>>it is far from being the primary interaction.
>>There is clearly a molecular conversation
>>between the plant and nematode prior to them
>>making direct physical contact. It is likely
>>that the interaction continues during
>>penetration of the host, and up to the
>>induction of giant cells 24-36 hours later. If
>>the plant is going to mount a defense response
>>(and this is conditioned in some instances by
>>single Medelian loci), this happens at this
>>point. I think "response to a nematode" is
>>sufficient.
>>dave
>>
>>David Bird
>>Director, Center for the Biology of Nematode Parasitism
>>Professor of Plant Pathology
>>Box 7253, NCSU
>>Raleigh NC 27695
>
>--
>Alexander Diehl, Ph.D.
>Scientific Curator
>Mouse Genome Informatics
>The Jackson Laboratory
>600 Main Street
>Bar Harbor, ME 04609
****************************************************************************
X-Sender: brtyler at pop.vt.edu
Date: Sun, 20 Mar 2005 19:28:00 -0500
To: ccollmer at wells.edu, mlgwinn at tigr.org, adiehl at informatics.jax.org
From: Brett Tyler <bmtyler at vt.edu>
Subject: Re: PAMGO: opinion needed from GO
Cc: arc2 at cornell.edu, ccollmer at wells.edu, david_bird at ncsu.edu,
jeremy at genome.wisc.edu, lhannick at tigr.org, perna at ahabs.wisc.edu,
ralph_dean at ncsu.edu, rbuell at tigr.org, setubal at vbi.vt.edu,
sutripa at vbi.vt.edu, thomas_mitchell at ncsu.edu
Status:
Folks,
I agree this is important. Revision of this part
of the ontology is well overdue.
I think there are positive aspects to both Alex's
and Suparna's proposals. I like the idea of
trying to make the terms as general as possible
to both plant and animal hosts.
The panel of experts that Suparna has assembled
are indeed leading lights in the plant side of
the plant-pathogen interaction field. However,
it seems that the terms have been proposed
without close attention to the terms that were
just accepted from PAMGO, nor to some of the key
philosophies of the GO.
Specific comments and suggestions:
1. the term "Symbiotic" is used incorrectly as a
synonym for mutualist rather than indicating any
biotic association.
2. Another needed correction is the replacement
of "saprophyte" with "neutralist". Saprophytism
does not describe a relationship whereas
neutralism does. Saprophytism refers to a
nutritional strategy.
3. many terms, such as "response to bacterial
pathogen" assume that the host can distinguish
the "intention" of the microbe. For example,
just because a microbe is a pathogenic bacterium
and the host is responding to the microbe does
not mean that the host is responding specifically
because the bacterium is a pathogen. "Response
to parasitic plant" also suffers from the same
problem. Thus I prefer Suparna's alternative
suggestion that "response to bacteria" (or
response to <some organism>) and "response to
pathogen" be separated. This would then be
congruent to what we have done on the microbe
side and fits with what Alex has proposed.
4. In similar vein, I think terms like "response
to pathogen-induced salicylic acid" are
over-specified. In other words, a plant gene may
respond to salicylic acid, but it cannot
distinguish between pathogen-induced salicyclic
acid and any other source or salicyclic acid,
unless it is also responding to some other
stimulus which is not salicylic acid, in which
cause it should be annotated with "response to
<something else>" e.g. defense response to
pathogen.
5. There are also path violations. For example
"response to a host-derived signal" cannot be a
"response to external biotic stimulus".
"Response to host-derived signal" should really
be worded "response to "defense signal" and
should be a a child of "response to biotic
stimulus" and sibling of "response to external
biotic signal"
6. Alex however is not correct in saying that
"response to bacterial pathogen" is synonymous
with "defense response to bacteria" because there
are many plant responses to bacterial pathogens
which are not defense responses. There are many
plant responses which are triggered by pathogen
in order to make the plant more susceptible.
Surely this is true also in the animal world.
Since I am arguing that terms like "response to
bacterial pathogen" should not be used, I would
propose that the proper annotation would be with
"(defense) response to bacteria" and/or
"(defense) response to pathogen"
7. Alex raises the somewhat philosophical issue
of whether to use "potential pathogen" instead of
"pathogen". I think this is too much
hair-splitting. A microbe doesn't suddenly stop
being a pathogen if the host successfully defeats
the microbe. The microbe can cause a wide range
of damage, from killing just a single cell to
killing the entire host as a result of its
activities. This squishiness is why PAMGO
actually proposed eliminating the term
"pathogenicity" altogether. However, given that
"pathogenicity" and "pathogen" are here to stay
(at the recent Fungal Genetics meeting there were
lots of raised eyebrows when I mentioned that
PAMGO had proposed deleting "pathogenicity"),
let's just keep things clean and use "pathogen"
and "pathogenicity".
I have attached my modification of Suparna's
proposal that incorporates my comments and
Alex's. A few of the existing terms may be
incorrect because I was writing this on the plane
and couldn't check them.
BTW. Folks at the Fungal Genetic Meeting were
very enthusiastic about the new PAMGO terms. I
gave four different presentations on them.
Cheers
Brett
[direct quote of earlier email deleted]
--
*****************
Brett Tyler
Professor
Virginia Bioinformatics Institute
Bioinformatics I, Washington Street
Virginia Polytechnic and State University
Blacksburg, VA 24061-0447
Telephone: (540) 231-7318
Fax: (540) 231-1678
Email: bmtyler at vt.edu
*****************
****************************************************************************
X-WebMail-UserID: ccollmer
Date: Mon, 21 Mar 2005 10:38:16 -0500
Sender: ccollmer <ccollmer at wells.edu>
From: ccollmer <ccollmer at wells.edu>
To: adiehl <adiehl at informatics.jax.org>, Brett Tyler <bmtyler at vt.edu>,
mlgwinn <mlgwinn at tigr.org>
Cc: "david_bird" <david_bird at ncsu.edu>, "ralph_dean" <ralph_dean at ncsu.edu>,
"thomas_mitchell" <thomas_mitchell at ncsu.edu>, arc2 <arc2 at cornell.edu>,
jeremy <jeremy at genome.wisc.edu>, lhannick <lhannick at tigr.org>,
perna <perna at ahabs.wisc.edu>, rbuell <rbuell at tigr.org>,
setubal <setubal at vbi.vt.edu>, sutripa <sutripa at vbi.vt.edu>
X-EXP32-SerialNo: 00003289
Subject: RE: PAMGO: opinion needed from GO
Status:
Dear Alex and PAMGO:
Alan and I also spent quite a bit of time pondering the dilemma here. I
have not talked with Alan about Brett's comments (below) since they just came
in this morning for me, but most of them seem true to me. Overall, I'd say we
feel that both Alex's proposal and Suparna's have important points, so a bit
more work (at least) will be required to get to a tree of terms that works for
all situtations pointed out by everyone responding so far. I certainly agree
that Suparna's group did not seem to try to be consistent with the
newly-adopted PAMGO terms, which is a bit distressing, considering how much
time was devoted to trying to get those in a form that would work for
everyone. And we certainly agree with Alex's approach to making broad terms
that will fit the needs of annotating both plant and animal genes, as well as
genes of both pathogens and other types of symbionts (because so many
underlying processes will be the same)---the main phyilosophical stance taken
by PAMGO all along.
Alan and I mainly talked about several points that we realize are very
complex, and wonder whether Michelle is correct here in thinking that we need
a content meeting to work out the complexities face to face. We certainly
agree with all who point out that a plant or an animal has at least some
responses to a bacterium (or other microbe) that are the same whether that
organism is a potential pathogen or not, and thus a broad term that
encompasses all bacteria recognized by the host seems most appropriate (and
thus we agree with Alex here). As pointed out by others, this also means that
in the tree proposed by Suparna et al., having "defense response to bacteria"
as a child to "response to bacterial pathogen" would result in some TPVs,
because not all defense responses would be in response to a bacterial
"pathogen" (i.e. some would be to non-pathogens -- e.g. the defense systems
that recognize flagella, or LPS molecules, no matter what type of bacterium
produces them). We're not sure whether merely changing the parent term here to
"response to **potential** bacterial pathogen" would solve this problem---but
that is certainly better than merely "response to bacterial pathogen," which
cannot work as written now.
The other major problem that we wrestled with is again what others have
also mentioned. This is the fact that plants (and animals too) have responses
to bacteria (for example) that are definitely not "defense responses." (And
thus we'd agree more with Suparna's group than with Alex here.) That would
include all the responses that some at least are calling "susceptibility
respones." So, not only does a microbe have the ability (in some cases) to
overcome a plant-mounted defense (perhaps by destroying defense molecules made
by the plant), but some pathogen-produced molecules definitely elicit a
response by the plant that makes that plant more susceptible to infection
(whether that will lead to nodule formation with Rhizobium, or giant cell
formation with a nematode, or to a disease state in an interaction with a
pathogen). Since scientists like those gathered together by Suparna are
intensely interested in these plant-microbe molecular interactions, and which
plant responses (molecules) are defensive versus "helpful" to the microbe
knocking at the door (whichever it is), it seems that we need to have terms to
accommodate these different types of responses by the host. Whether that means
we need merely a general term "response to bacteria" or a more specific one
that says, for example, "susceptibility response to bacteria," is perhaps an
important point to debate in the wider community. The problem, of course, is
that there may well be plant molecules that are made in both cases (both as
defense molecules and as susceptibility molecules, perhaps with different
organisms [and the molecules they do or do not make] or under different
situations), and more importantly, there will be many molecules that we can
tell are being made in response to a potential pathogen, but we may not know
what their actual role is in the complex interaction between microbe and host
(for instance, molecules that are shown in microarray experiments to be made
by the plant uniquely in response to inoculation). Again, since there is much
interest in finding all such moelcules in the plant-microbe interaction world,
we do need terms that will capture the attention, and meet the needs, of these
scientists.
So, where do we go from here? Is there a need for a content meeting? Or
do others think this can be hammered out via Sourceforge and email? I would
like to know soon, as I'm currently considering going to the GO Annotation
camp this coming June, and if I need to do this instead, I'll want to make
plans. --Candace
[direct quote of earlier email deleted]
----------------------------
Candace W. Collmer, Ph.D.
Professor of Biology
Wells College
Aurora, NY 13026
phone: 315-364-3271
email: ccollmer at wells.edu
****************************************************************************
My letter of 3/21/05 to Brett Tyler
To: Brett Tyler <bmtyler at vt.edu>, Candace_Collmer
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: PAMGO: opinion needed from GO
Cc:
Bcc:
Attachments:
Dear Brett and Candace,
Thank you both for your responses. My immediate
plan is to write up some comments to your
responses, which I will send to you in the next
day or two. Then I plan to post a compendium of
the email exchanges I've had with the two of you
and David Bird, unedited, to the pathogenesis
mailing list (pathogenesis at geneontology.org),
which I will cross post to the plant-pathogen
expert group assembled by Suparna, as well as the
PAMGO group (perhaps someone could send me a
definite listing of PAMGO members and emails), to
facilitate more public discussion. I encourage
you to subscribe to the pathogenesis mailing list
if you haven't already. The mailing list has not
been used much yet, but seems like the
appropriate forum for continuing discussion.
I will also speak on this issue at the GO
Consortium meeting to let people know the state
of the discussion.
The GO content meeting for discussion of issues
in immunology and defense responses is loosely
scheduled for October or November of this year.
I would love to get this topic cleared up before
then, but if not, this content meeting seems like
the appropriate place for its resolution.
-- Alex
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
My letter of 3/22/05 to Brett Tyler
To: Brett Tyler <bmtyler at vt.edu>, ccollmer at wells.edu, mlgwinn at tigr.org
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: PAMGO: opinion needed from GO
Cc: arc2 at cornell.edu, ccollmer at wells.edu,
david_bird at ncsu.edu, jeremy at genome.wisc.edu,
lhannick at tigr.org, perna at ahabs.wisc.edu,
ralph_dean at ncsu.edu, rbuell at tigr.org,
setubal at vbi.vt.edu, sutripa at vbi.vt.edu,
thomas_mitchell at ncsu.edu, owhite at tigr.org
Bcc:
X-Attachments:
Brett,
I'm going to intersperse my comments throughout.
At 7:28 PM -0500 3/20/05, Brett Tyler wrote:
>Folks,
>
>I agree this is important. Revision of this part
>of the ontology is well overdue.
>
>I think there are positive aspects to both
>Alex's and Suparna's proposals. I like the idea
>of trying to make the terms as general as
>possible to both plant and animal hosts.
>
>The panel of experts that Suparna has assembled
>are indeed leading lights in the plant side of
>the plant-pathogen interaction field. However,
>it seems that the terms have been proposed
>without close attention to the terms that were
>just accepted from PAMGO, nor to some of the key
>philosophies of the GO.
>
>Specific comments and suggestions:
>
>1. the term "Symbiotic" is used incorrectly as a
>synonym for mutualist rather than indicating any
>biotic association.
>
>2. Another needed correction is the replacement
>of "saprophyte" with "neutralist". Saprophytism
>does not describe a relationship whereas
>neutralism does. Saprophytism refers to a
>nutritional strategy.
>
>3. many terms, such as "response to bacterial
>pathogen" assume that the host can distinguish
>the "intention" of the microbe. For example,
>just because a microbe is a pathogenic bacterium
>and the host is responding to the microbe does
>not mean that the host is responding
>specifically because the bacterium is a
>pathogen. "Response to parasitic plant" also
>suffers from the same problem. Thus I prefer
>Suparna's alternative suggestion that "response
>to bacteria" (or response to <some organism>)
>and "response to pathogen" be separated. This
>would then be congruent to what we have done on
>the microbe side and fits with what Alex has
>proposed.
I agree with you that the host cannot distinguish
the intention of the microbe, but is simply
responding to particular microbial products in an
evolutionarily defined way to promote its own
survival.
Furthermore, I would argue that "response to
pathogen" should simply be obsoleted. See below
>
>4. In similar vein, I think terms like "response
>to pathogen-induced salicylic acid" are
>over-specified. In other words, a plant gene
>may respond to salicylic acid, but it cannot
>distinguish between pathogen-induced salicyclic
>acid and any other source or salicyclic acid,
>unless it is also responding to some other
>stimulus which is not salicylic acid, in which
>cause it should be annotated with "response to
><something else>" e.g. defense response to
>pathogen.
I absolutely agree with you here.
>
>5. There are also path violations. For example
>"response to a host-derived signal" cannot be a
>"response to external biotic stimulus".
>"Response to host-derived signal" should really
>be worded "response to "defense signal" and
>should be a a child of "response to biotic
>stimulus" and sibling of "response to external
>biotic signal"
Again, I agree with you on these points.
>
>6. Alex however is not correct in saying that
>"response to bacterial pathogen" is synonymous
>with "defense response to bacteria" because
>there are many plant responses to bacterial
>pathogens which are not defense responses. There
>are many plant responses which are triggered by
>pathogen in order to make the plant more
>susceptible. Surely this is true also in the
>animal world. Since I am arguing that terms
>like "response to bacterial pathogen" should
>not be used, I would propose that the proper
>annotation would be with "(defense) response to
>bacteria" and/or "(defense) response to pathogen"
Here, I have to point out a key aspect of the GO,
which should not be overlooked in the formulation
of GO terms. The GO is not intended for the
annotation of gene product involvement in
abnormal or disease processes for that organism.
This is clear from the documentation of the GO
and from discussions at GO Consortium meetings.
A plant response which is triggered by a pathogen
in order to make the plant more susceptible is a
disease process, and the plant's gene products
involved in that process have other normal
processes for which they ought to be annotated
using the GO. Response process terms in the GO
should only cover normal plant and animal
biology, not disease biology. Defense responses
are normal. Other types of responses such as
establishing a symbiotic relationship or trapping
a fly for nutritive uptake (a potential plant
process) are normal.
This is one of the main reasons why "response to pathogen" is not a valid term.
If a plant-infecting bacterium utilizes a
signaling pathway in a plant normally used for
establishing symbiosis as a trick to evade or
suppress the plant's defenses, the plant gene
products involved in that pathway should
nevertheless be annotated to GO:0009606 response
to symbiotic bacteria because that is their
normal function, assuming there is evidence from
another experiment for that role. No plant would
have evolved a pathway for promoting its own
demise, unless the process was part of a defense
response, such a the induction of an apoptosis
pathway to hinder the spread of an infecting
organism, or a demise that somehow promoted the
survival of other members of the same species (I
can't think of a plant example of this, but there
probably is one).
I have included three links and quotes from the GO Consortium documentation.
What is a GO term? http://www.geneontology.org/GO.usage.shtml
As explained in An Introduction to GO , the
purpose of GO is to define particular attributes
of gene products. Practically speaking, a term is
simply the text string used to describe an entry
in GO, e.g. cell ,fibroblast growth factor
receptor binding or signal transduction . A node
refers to a term and all its children. GO does
not contain the following:
Gene products: e.g. cytochrome c is not in the
ontologies, but attributes of cytochrome c, such
as electron transporter , are.
Processes, functions or components that are
unique to mutants or diseases: e.g. oncogenesis
is not a valid GO term because causing cancer is
not the normal function of any gene.
Attributes of sequence such as intron/exon
parameters: these are not attributes of gene
products and will be described in a separate
sequence ontology (see the OBO web page for more
information).
Protein domains or structural features.
Protein-protein interactions.
http://www.geneontology.org/GO.annotation.shtml
A gene product should be annotated with terms
reflecting its normal activity and location. A
function, process, or localization (component)
observed only in a mutant or disease state is
therefore not usually included. In some
circumstances, however, what is "normal" depends
on the point of view taken by the annotator. For
example, many viruses use host proteins to carry
out viral processes. The host protein is then
doing something abnormal from the perspective of
the host, but completely normal from the
perspective of the virus. GO annotators handle
these cases by including two taxon IDs in the
"Taxon" column of the gene association file (see
Annotation File Format, below). When two taxon
IDs appear, the first is that of the organism
that encodes the gene product, and the second ID
is that of the organism that uses the gene
product, and whose perspective is considered
"normal" for that annotation. Beyond GO
http://www.geneontology.org/GO.doc.shtml
A family of open-source ontologies
GO allows us to annotate genes and their products
with a limited set of attributes. For example, GO
does not allow us to describe genes in terms of
which cells or tissues they're expressed in,
which developmental stages they're expressed at,
or their involvement in disease. It is not
necessary for GO to do these things because other
ontologies are being developed for these
purposes. The GO consortium supports the
development of other ontologies and makes its
tools for editing and curating ontologies freely
available. A list of freely available ontologies
that are relevant to genomics and proteomics and
are structured similarly to GO can be found at
the OBO (open biology ontologies) website. A
larger list, which includes the ontologies listed
at OBO and also other controlled vocabularies
that do not fulfil the OBO criteria is available
at the Ontology Working Group page of the
Microarray Gene Expression Data Society (MGED).
>7. Alex raises the somewhat philosophical issue
>of whether to use "potential pathogen" instead
>of "pathogen". I think this is too much
>hair-splitting. A microbe doesn't suddenly stop
>being a pathogen if the host successfully
>defeats the microbe. The microbe can cause a
>wide range of damage, from killing just a single
>cell to killing the entire host as a result of
>its activities. This squishiness is why PAMGO
>actually proposed eliminating the term
>"pathogenicity" altogether. However, given that
>"pathogenicity" and "pathogen" are here to stay
>(at the recent Fungal Genetics meeting there
>were lots of raised eyebrows when I mentioned
>that PAMGO had proposed deleting
>"pathogenicity"), let's just keep things clean
>and use "pathogen" and "pathogenicity".
As I explained in my response to David Bird,
having "defense response to bacteria" as a child
of "defense response to pathogen" assumes that
all bacteria are pathogens, which is not the
case. Moreover, even non-pathogenic bacteria can
provoke defense responses. The responding plant
or animal makes a defense response based on
molecular clues given off by the bacteria without
knowing whether the bacteria is truly a threat.
It is clear, however, the bacteria is a potential
threat, and the getting rid of the potential
threat is a better bet that letting it hang
around and maybe do something nasty.
My initial feeling was to dispose of the "defense
response to pathogen" term completely because of
the potential true-path violation. However, I
feel that modifying the term to "defense response
to potential pathogen" actually make it a
stronger term because the term is written from
the perspective of the responding organism which
cannot know for sure what is a pathogen and what
is not, but is only making a response based on
the possibility that the continuing presence of
the infecting organism is not likely to be a good
thing. Furthermore, modifying the term still
allows GO users to search on the word "pathogen"
and come up with a list of genes annotated to the
term or its children.
I would like to promote the general idea that
responses in the GO (including defense responses)
should be limited to responses that increase the
fitness of an organism or a group of organisms of
the same species living in close proximity. This
includes symbiotic responses, because forming a
symbiotic relationship promotes the fitness of at
least one of the partners.
>
>I have attached my modification of Suparna's
>proposal that incorporates my comments and
>Alex's. A few of the existing terms may be
>incorrect because I was writing this on the
>plane and couldn't check them.
I agree with basically all of your changes, and
will bear them in mind as I further revise my
proposal regarding these terms.
Thank you very much for your time and interest,
Alex
>BTW. Folks at the Fungal Genetic Meeting were
>very enthusiastic about the new PAMGO terms. I
>gave four different presentations on them.
>
>Cheers
>
>Brett
>
>
>
>--
>*****************
>Brett Tyler
>Professor
>Virginia Bioinformatics Institute
>Bioinformatics I, Washington Street
>Virginia Polytechnic and State University
>Blacksburg, VA 24061-0447
>
>Telephone: (540) 231-7318
>Fax: (540) 231-1678
>Email: bmtyler at vt.edu
>*****************
--
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
My letter of 3/22/05 to Candace Collmer
To: ccollmer <ccollmer at wells.edu>, Brett Tyler
<bmtyler at vt.edu>, mlgwinn <mlgwinn at tigr.org>
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: RE: PAMGO: opinion needed from GO
Cc: "david_bird" <david_bird at ncsu.edu>,
"ralph_dean" <ralph_dean at ncsu.edu>,
"thomas_mitchell" <thomas_mitchell at ncsu.edu>,
arc2 <arc2 at cornell.edu>, jeremy
<jeremy at genome.wisc.edu>, lhannick
<lhannick at tigr.org>, perna
<perna at ahabs.wisc.edu>, rbuell <rbuell at tigr.org>,
setubal <setubal at vbi.vt.edu>, sutripa
<sutripa at vbi.vt.edu>, owhite at tigr.org
Bcc:
Attachments:
Candace,
I will try to repeat what I just wrote back to
Brett. My comments are interspersed.
At 10:38 AM -0500 3/21/05, ccollmer wrote:
>Dear Alex and PAMGO:
> Alan and I also spent quite a bit of time pondering the dilemma here. I
>have not talked with Alan about Brett's comments (below) since they just came
>in this morning for me, but most of them seem true to me. Overall, I'd say we
>feel that both Alex's proposal and Suparna's have important points, so a bit
>more work (at least) will be required to get to a tree of terms that works for
>all situtations pointed out by everyone responding so far. I certainly agree
>that Suparna's group did not seem to try to be consistent with the
>newly-adopted PAMGO terms, which is a bit distressing, considering how much
>time was devoted to trying to get those in a form that would work for
>everyone. And we certainly agree with Alex's approach to making broad terms
>that will fit the needs of annotating both plant and animal genes, as well as
>genes of both pathogens and other types of symbionts (because so many
>underlying processes will be the same)---the main phyilosophical stance taken
>by PAMGO all along.
I really like and support the PAMGO viewpoint.
> Alan and I mainly talked about several points that we realize are very
>complex, and wonder whether Michelle is correct here in thinking that we need
>a content meeting to work out the complexities face to face. We certainly
>agree with all who point out that a plant or an animal has at least some
>responses to a bacterium (or other microbe) that are the same whether that
>organism is a potential pathogen or not, and thus a broad term that
>encompasses all bacteria recognized by the host seems most appropriate (and
>thus we agree with Alex here). As pointed out by others, this also means that
>in the tree proposed by Suparna et al., having "defense response to bacteria"
>as a child to "response to bacterial pathogen" would result in some TPVs,
>because not all defense responses would be in response to a bacterial
>"pathogen" (i.e. some would be to non-pathogens -- e.g. the defense systems
>that recognize flagella, or LPS molecules, no matter what type of bacterium
>produces them). We're not sure whether merely changing the parent term here to
>"response to **potential** bacterial pathogen" would solve this problem---but
>that is certainly better than merely "response to bacterial pathogen," which
>cannot work as written now.
We are basically in agreement.
I just want to stress that my proposed term is
"defense response to potential pathogen." I
think it is important to remember that defense
responses are a subset of all responses. The
structure of the GO I would like to see has both
"response to" and "defense response to" terms,
with the latter being an is-a of the former. The
term "defense response to potential pathogen"
provides a second parent for terms like "defense
response to bacteria," and is itself an is-a
child of "defense response."
> The other major problem that we wrestled with is again what others have
>also mentioned. This is the fact that plants (and animals too) have responses
>to bacteria (for example) that are definitely not "defense responses." (And
>thus we'd agree more with Suparna's group than with Alex here.)
In general terms I agree with you here.
Paramecia eat bacteria, and in most cases not as
a defense response. Plants form symbiotic
relationships with root bacteria. Responses to
commensal bacteria in the colon of mice promotes
epithelial homeostasis (Cell 118:229). These are
all types of responses to bacteria which are not
necessary defense responses.
>That would
>include all the responses that some at least are calling "susceptibility
>respones." So, not only does a microbe have the ability (in some cases) to
>overcome a plant-mounted defense (perhaps by destroying defense molecules made
>by the plant), but some pathogen-produced molecules definitely elicit a
>response by the plant that makes that plant more susceptible to infection
>(whether that will lead to nodule formation with Rhizobium, or giant cell
>formation with a nematode, or to a disease state in an interaction with a
>pathogen). Since scientists like those gathered together by Suparna are
>intensely interested in these plant-microbe molecular interactions, and which
>plant responses (molecules) are defensive versus "helpful" to the microbe
>knocking at the door (whichever it is), it seems that we need to have terms to
>accommodate these different types of responses by the host.
As I have pointed out in my response to Brett,
susceptibility responses are in fact disease
responses or at least abnormal biology. The GO
is concerned with the normal function and biology
of gene products. Susceptibility responses
result from the microbe utilizing a plant pathway
designed for a different purpose to promote its
own interests, essentially tricking the plant.
The plant gene products involved in such pathways
ought be annotated to their proper functions or
processes for the plant, as they did not evolve
for the purpose of facilitating non-symbiotic
microbial infection. Thus there is no need for
the term "response to pathogen," as the only
self-interested responses to pathogens are
defense responses or responses which co-opt the
pathogen into being a symbiont (which can be seen
a defensive as well).
>Whether that means
>we need merely a general term "response to bacteria" or a more specific one
>that says, for example, "susceptibility response to bacteria," is perhaps an
>important point to debate in the wider community.
The GO editors and others including me would
argue that a "susceptibility response to
bacteria" term is really describing a phenotype
and not gene function or process.
> The problem, of course, is
>that there may well be plant molecules that are made in both cases (both as
>defense molecules and as susceptibility molecules, perhaps with different
>organisms [and the molecules they do or do not make] or under different
>situations), and more importantly, there will be many molecules that we can
>tell are being made in response to a potential pathogen, but we may not know
>what their actual role is in the complex interaction between microbe and host
>(for instance, molecules that are shown in microarray experiments to be made
>by the plant uniquely in response to inoculation). Again, since there is much
>interest in finding all such moelcules in the plant-microbe interaction world,
>we do need terms that will capture the attention, and meet the needs, of these
>scientists.
I would argue that the GO is the wrong ontology
for annotating resistance to or susceptibility to
pathogens or disease. These are phenotypes and
this issue were decided in the GO several years
ago. I don't know if plant phenotype ontologies
have been created or not, but since there seems
to be an ontology for everything these days,
there probably is one already. The GO is great
for many purposes but not every purpose.
> So, where do we go from here? Is there a need for a content meeting? Or
>do others think this can be hammered out via Sourceforge and email? I would
>like to know soon, as I'm currently considering going to the GO Annotation
>camp this coming June, and if I need to do this instead, I'll want to make
>plans. --Candace
Let's see how far we can get through the
pathogenesis mailing list, before committing this
issue to the GO content meeting later in the
year. I will of course get feedback at the GO
Consortium meeting regarding this issue. I also
don't want to dissuade you from attending the
annotation camp, as that too could be very
positive.
Thank you very much for taking time to comment,
Alex
>----------------------------
>Candace W. Collmer, Ph.D.
>Professor of Biology
>Wells College
>Aurora, NY 13026
> phone: 315-364-3271
> email: ccollmer at wells.edu
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
X-WebMail-UserID: ccollmer
Date: Thu, 24 Mar 2005 09:34:39 -0500
Sender: ccollmer <ccollmer at wells.edu>
From: ccollmer <ccollmer at wells.edu>
To: Alexander Diehl <adiehl at informatics.jax.org>, Brett Tyler <bmtyler at vt.edu>,
mlgwinn <mlgwinn at tigr.org>
Cc: "david_bird" <david_bird at ncsu.edu>, "ralph_dean" <ralph_dean at ncsu.edu>,
"thomas_mitchell" <thomas_mitchell at ncsu.edu>, arc2 <arc2 at cornell.edu>,
jeremy <jeremy at genome.wisc.edu>, lhannick <lhannick at tigr.org>,
owhite <owhite at tigr.org>, perna <perna at ahabs.wisc.edu>,
rbuell <rbuell at tigr.org>, setubal <setubal at vbi.vt.edu>,
sutripa <sutripa at vbi.vt.edu>
X-EXP32-SerialNo: 00003289
Subject: important -- question for PAMGO
Status:
Dear Alex and PAMGO members:
Thanks, Alex, for so patiently addressing our comments and those of Brett.
Basically, I agree with most all of what you have said, and I do realize that
we do not want to annotate plant genes beased on their playing a part in
inducing some particular phenotype. Are you now suggesting (in relation to
bacteria) that that there be a parent term "response to bacteria" and then two
children terms, "response to symbiotic bacteria" [i.e. mutualistic through
parasitic] and "defense response to potential pathogen" (recognizing that this
latter term would actually have to be a child of "response to symbiotic
bacteria" -- given the definition of symbiosis we're using)? Would these
terms then cover all plant genes that are turned on in response to the
appearance of a bacterium in the vicinity? (I think that's importnant,
obviously, as people working on the plant side will want to annotate these
genes using GO, or they may just ignore GO.) Do others agree that these terms
would cover all needs on the plant side? I guess I'm still a bit concerned
about whether all "susceptibility responses" on the part of a plant would
necessarily be considered as abnormal responses, outside the scope of GO, or
whether there are some that will be hard to disentangle as to their exact role
-- normal or not. But perhaps those will then be annotated only to "response
to bacteria"? What do others think? --Candace
[quote of my directly preceding email deleted]
----------------------------
Candace W. Collmer, Ph.D.
Professor of Biology
Wells College
Aurora, NY 13026
phone: 315-364-3271
email: ccollmer at wells.edu
****************************************************************************
Cc: mlgwinn <mlgwinn at tigr.org>, Alexander Diehl <adiehl at informatics.jax.org>,
arc2 <arc2 at cornell.edu>, "thomas_mitchell" <thomas_mitchell at ncsu.edu>,
setubal <setubal at vbi.vt.edu>, perna <perna at ahabs.wisc.edu>,
"ralph_dean" <ralph_dean at ncsu.edu>, owhite <owhite at tigr.org>,
Brett Tyler <bmtyler at vt.edu>, lhannick <lhannick at tigr.org>,
rbuell <rbuell at tigr.org>, jeremy <jeremy at genome.wisc.edu>,
sutripa <sutripa at vbi.vt.edu>
From: David Bird <david_bird at ncsu.edu>
Subject: Re: important -- question for PAMGO
Date: Thu, 24 Mar 2005 09:47:23 -0500
To: ccollmer <ccollmer at wells.edu>
X-PMX-Version: 4.7.1.128075, Antispam-Engine:
2.0.3.0, Antispam-Data: 2005.3.24.6
X-Spam-Status: No, Hits=7%
X-Spam-Level: IIIIIII
Status:
Hi Candice,
If I substitute "nematode" for "bacteria", then I
think the proposed solution is ideal. It appears
to cover the full range of nematode-plant
interactions (from simple herbivores to those
which engage in complex parasitic symbioses) and
captures what we know about resistance responses.
And I think it also covers the situation where a
nematode (or bacterium, etc) might signal a plant
from a distance, but then move on. Or, penetrate
a plant, but then leave (such as often occurs
when a nematode enters a "non-host" plant.
But I am not quite sure what is meant by your
comment (I admit I have not read all the e-mail
traffic on this subject):
>I guess I'm still a bit concerned
>about whether all "susceptibility responses" on the part of a plant would
>necessarily be considered as abnormal responses, outside the scope of GO
dave
David Bird
Director, Center for the Biology of Nematode Parasitism
Professor of Plant Pathology
Box 7253, NCSU
Raleigh NC 27695
+1 (919) 515-6813
[direct quote of Candace Collmer's immediately preceding email deleted]
****************************************************************************
Date: Thu, 24 Mar 2005 10:03:12 -0500
To: ccollmer <ccollmer at wells.edu>
From: Brett Tyler <bmtyler at vt.edu>
Subject: Re: important -- question for PAMGO
Cc: Alexander Diehl <adiehl at informatics.jax.org>, mlgwinn <mlgwinn at tigr.org>,
david_bird <david_bird at ncsu.edu>, ralph_dean <ralph_dean at ncsu.edu>,
thomas_mitchell <thomas_mitchell at ncsu.edu>, arc2 <arc2 at cornell.edu>,
jeremy <jeremy at genome.wisc.edu>, lhannick <lhannick at tigr.org>,
owhite <owhite at tigr.org>, perna <perna at ahabs.wisc.edu>,
rbuell <rbuell at tigr.org>, setubal <setubal at vbi.vt.edu>,
sutripa <sutripa at vbi.vt.edu>
Status:
All,
I spent some time trying to think about the issue
of plant "susceptibility" responses that are a
"normal" part of the interaction with a
mutualistic organism, but have been co-opted by a
pathogen.
I think there is not a problem.
Since we won't know if a "susceptibility"
response is really a coopted mutualistic response
until we have characterized the mutualistic
response, genes involved in the response would
initially be annotated just with "response to
<specific organism> and later with "response to
mutualist"
I think that the set of terms "response to
<specific organism>", "response to symbiotic
<specific organism>, mutualistic through
parasitic" and "defense response to potential
pathogen" will be sufficient assuming the
eventual addition of terms like "response to
mutualist" and "response to mutualistic <specific
organism>"
Cheers
Brett
[quote of Candace Collmer's immediately preceding email deleted]
--
*****************
Brett Tyler
Professor
Virginia Bioinformatics Institute
Virginia Polytechnic Institute and State University
Washington Street
Blacksburg, VA 24061-0447
Telephone: (540) 231-7318
Fax: (540) 231-2606
Email: bmtyler at vt.edu
Web: https://www.vbi.vt.edu/article/articleview/141
*****************
****************************************************************************
Cc: mlgwinn <mlgwinn at tigr.org>, Alexander Diehl <adiehl at informatics.jax.org>,
arc2 <arc2 at cornell.edu>, setubal <setubal at vbi.vt.edu>,
perna <perna at ahabs.wisc.edu>, owhite <owhite at tigr.org>,
lhannick <lhannick at tigr.org>, rbuell <rbuell at tigr.org>,
ralph_dean <ralph_dean at ncsu.edu>,
thomas_mitchell <thomas_mitchell at ncsu.edu>,
jeremy <jeremy at genome.wisc.edu>, ccollmer <ccollmer at wells.edu>,
sutripa <sutripa at vbi.vt.edu>
From: David Bird <david_bird at ncsu.edu>
Subject: Re: important -- question for PAMGO
Date: Thu, 24 Mar 2005 10:14:25 -0500
To: Brett Tyler <bmtyler at vt.edu>
X-PMX-Version: 4.7.1.128075, Antispam-Engine:
2.0.3.0, Antispam-Data: 2005.3.24.6
X-Spam-Status: No, Hits=7%
X-Spam-Level: IIIIIII
Status:
All,
I like the broad definition of symbiont, and in
fact I am not so sure that we need to try to
codify subterms such as "response to mutualist"
or "response to parasite", because how do we know
that they are not the same plant pathways.
Indeed, rhizobia, mycorrhizae and root-knot
nematodes all interact through an at least partly
shared plant pathway.
dave
David Bird
Director, Center for the Biology of Nematode Parasitism
Professor of Plant Pathology
Box 7253, NCSU
Raleigh NC 27695
+1 (919) 515-6813
[direct quote of Brett Tyler's immediately preceding email deleted]
****************************************************************************
From: "Nicole T. Perna" <perna at svm.vetmed.wisc.edu>
Subject: Re: important -- question for PAMGO
To: David Bird <david_bird at ncsu.edu>, ccollmer <ccollmer at wells.edu>
Cc: mlgwinn <mlgwinn at tigr.org>, Alexander Diehl
<adiehl at informatics.jax.org>,
arc2 <arc2 at cornell.edu>, "thomas_mitchell" <thomas_mitchell at ncsu.edu>,
setubal
<setubal at vbi.vt.edu>, "ralph_dean" <ralph_dean at ncsu.edu>,
owhite
<owhite at tigr.org>, Brett Tyler <bmtyler at vt.edu>,
lhannick
<lhannick at tigr.org>, rbuell <rbuell at tigr.org>,
jeremy
<jeremy at genome.wisc.edu>, sutripa <sutripa at vbi.vt.edu>
Date: Thu, 24 Mar 2005 09:43:13 -0600
Status:
Candace et al.,
I’ve been quiet throughout this discussion, but am
basically on board with the current direction. As with
the rest of you, I’m dedicated to the broad definition of
symbiosis. Unlike David, I think it may be important to
develop the subterms for types of symbiosis. Should the
same gene product prove to be involved with both mutualism
and parasitism, both terms could be included. You might
argue that in the absence of evidence that the gene is
involved with both, you should always choose the more
general term, but that seems like a usage issue, not an
ontology problem. This issue may even occur much earlier
in the tree if the same gene products are involved with
detection or response to both nematodes and bacteria. I
suppose the beauty of the system is that commonalities
among terms can be detected even when gene products use
different child terms originating from a common parent.
Nicole
[direct quote of David Bird's email of Thu, 24
Mar 2005 09:47:23 -0500 (above) deleted]
****************************************************************************
My letter of 3/28/05 to Chris Mungall
To: Chris_Mungall
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: is_a question
Cc:
Bcc:
Attachments:
Chris,
I'm working with the PAMGO group to clarify some
term issues dealing with responses and defense
responses.
The partial DAG I've proposed is below (all edges are is_a):
Additional parental has been proposed certain
members of the PAMGO group for the "defense
response to potential pathogen" term making it
also an is_a of "response to symbiont." However
not all defense responses to potential pathogens
are types of responses to symbiont. So in the
graph below, the red edges are is_a's, but only
part-time is_a's:
Is it acceptable to have an is_a child when the
child term is only a subtype of the parent term
in certain situations?
Thanks,
Alex
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
My letter of 3/28/05 to Candace Collmer and PAMGO
To: ccollmer <ccollmer at wells.edu>, Brett Tyler
<bmtyler at vt.edu>, mlgwinn <mlgwinn at tigr.org>
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: important -- question for PAMGO
Cc: "david_bird" <david_bird at ncsu.edu>,
"ralph_dean" <ralph_dean at ncsu.edu>,
"thomas_mitchell" <thomas_mitchell at ncsu.edu>,
arc2 <arc2 at cornell.edu>, jeremy
<jeremy at genome.wisc.edu>, lhannick
<lhannick at tigr.org>, owhite <owhite at tigr.org>,
perna <perna at ahabs.wisc.edu>, rbuell
<rbuell at tigr.org>, setubal <setubal at vbi.vt.edu>,
sutripa <sutripa at vbi.vt.edu>
Bcc:
X-Attachments:
Candace,
Hi again, here are some delayed comments to your email of last Thursday.
"Defense response to potential pathogen" is not a
child of "response to bacteria," or "response to
symbiont" in my scheme. It rather a grouping
term to allow defense response terms for
particular types of organisms such as bacteria or
fungi which are sometimes considered. A partial
DAG for my proposed hierarchy is below (all edges
are is_a):
I can, however, see why "defense response to
potential pathogen" might also be considered an
is-a to response to symbiont, given the broad
range of interaction codified in the GO term
GO:0044403 symbiosis, mutualism through
parasitism. The question becomes whether the
condition of symbiosis is triggered under
whatever circumstances two organisms are in the
immediate vicinity of each other, which is overly
broad in my point of view, or not. If a bacteria
has just been brushed in a skin wound, that isn't
exactly symbiosis at the start, although there
may be a defense response immediately. Perhaps
the situation is different for plants, since they
don't move very much, but the GO terms have work
for all of biology.
I've experimented adding additional parentage to
DAG to allow "defense response to potential
pathogen" to be an is_a child to "response to
symbiont," although as I wrote above, this is not
necessarily the case at all times. The resulting
graph looks like this, where the red edges are
"part-time" is-a's:
I've asked Chris Mungall of FlyBase to comment on
whether such part-time term parentage is
acceptable in the GO. Chris has a particular
interest in the logical structure of the GO and
will probably have a relevant opinion. I am
awaiting his reply.
As far as susceptibility responses go, I would
have to say that annotation to the GO is simply
incorrect to cover experimental results that
describe disease. I will use a human example
(sorry): One of the way HIV kills T-cells is
through direct cross linking of CD4 molecules on
the cell surface by soluble gp120 shed from HIV
virions (Banda, N. K. et al., J Exp Med
176:1099). Yet I would not annotate CD4 to
"response to virus," based on this experiment as
the result is a disease process, not the normal
function of the CD4 molecule. In other
situations, based on its role in T-cell
activation in cell-mediated immune responses to
viruses, CD4 might warrant an annotation to
"defense response to virus," but not from an
experiment that is really telling us something
about virus biology, not host biology. The
distinction is that the virus is using CD4 to
provoke a response in the host which may help it
to evade the host's immune system, rather than
that the host displaying that response as a
defense response against the virus per se. The
CD4 molecule did not evolve for the purpose of
promoting HIV infection, although the gp120
molecule did.
Naturally, I want to promote use of the GO across
biology, but the GO is not appropriate for all
types of annotation. People annotating
host-pathogen interactions in plants and animals
may indeed find that the GO is inadequate when
what one really wishes to describe is a phenotype
or disease process.
-- Alex
At 9:34 AM -0500 3/24/05, ccollmer wrote:
>Dear Alex and PAMGO members:
> Thanks, Alex, for so patiently addressing our comments and those of Brett.
>Basically, I agree with most all of what you have said, and I do realize that
>we do not want to annotate plant genes beased on their playing a part in
>inducing some particular phenotype. Are you now suggesting (in relation to
>bacteria) that that there be a parent term "response to bacteria" and then two
>children terms, "response to symbiotic bacteria" [i.e. mutualistic through
>parasitic] and "defense response to potential pathogen" (recognizing that this
>latter term would actually have to be a child of "response to symbiotic
>bacteria" -- given the definition of symbiosis we're using)? Would these
>terms then cover all plant genes that are turned on in response to the
>appearance of a bacterium in the vicinity? (I think that's importnant,
>obviously, as people working on the plant side will want to annotate these
>genes using GO, or they may just ignore GO.) Do others agree that these terms
>would cover all needs on the plant side? I guess I'm still a bit concerned
>about whether all "susceptibility responses" on the part of a plant would
>necessarily be considered as abnormal responses, outside the scope of GO, or
>whether there are some that will be hard to disentangle as to their exact role
>-- normal or not. But perhaps those will then be annotated only to "response
>to bacteria"? What do others think? --Candace
>
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
Date: Mon, 28 Mar 2005 14:18:58 -0800 (PST)
From: Chris Mungall <cjm at fruitfly.org>
X-X-Sender: cjm at skerryvore.dhcp.lbl.gov
To: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: is_a question
X-Virus-Scanned: by amavisd-new using ClamAV at fruitfly.org
Status:
Hi Alex
I'm afraid part time 'is_a's are not allowed
Surely what you want is this:
response to bacteria
defense response to bacteria
defense response to symbiotic bacteria
response to symbiotic bacteria
defense response to symbiotic bacteria
I can see why you might want to avoid the cross-product but there's not
really an alternative here as far as I can see
Perhaps the above is problematic because a defense response to a symbiotic
organism would be by definition a pathological process and therefore not
in GO?
[direct quote of my letter to Chris Mungall deleted]
****************************************************************************
Date: Mon, 28 Mar 2005 17:59:10 -0500
To: Alexander Diehl <adiehl at informatics.jax.org>
From: Brett Tyler <bmtyler at vt.edu>
Subject: Re: important -- question for PAMGO
Cc: ccollmer <ccollmer at wells.edu>, mlgwinn <mlgwinn at tigr.org>,
david_bird <david_bird at ncsu.edu>, ralph_dean <ralph_dean at ncsu.edu>,
thomas_mitchell <thomas_mitchell at ncsu.edu>, arc2 <arc2 at cornell.edu>,
jeremy <jeremy at genome.wisc.edu>, lhannick <lhannick at tigr.org>,
owhite <owhite at tigr.org>, perna <perna at ahabs.wisc.edu>,
rbuell <rbuell at tigr.org>, setubal <setubal at vbi.vt.edu>,
sutripa <sutripa at vbi.vt.edu>
Status:
Alex and Candace,
Wouldn't adding the word "potential" to create
"response to potential symbiont" cover the
situation where the outcome of the interaction
was in doubt, such as the skin wound, and so
enable it to be a full-time parent of "defense
response to potential pathogen".
"Response to potential symbiont" could also be
used to annotate "susceptibility" genes that are
"normally" used to "help" establish mutualisms
but have been coopted by pathogens, since again
we avoid the need for the host to be able to
"predict" the outcome of the interaction.
Cheers
Brett
[direct quote of my letter of 3/28/05 to Candace Collmer deleted]
--
*****************
Brett Tyler
Professor
Virginia Bioinformatics Institute
Virginia Polytechnic Institute and State University
Washington Street
Blacksburg, VA 24061-0447
Telephone: (540) 231-7318
Fax: (540) 231-2606
Email: bmtyler at vt.edu
Web: https://www.vbi.vt.edu/article/articleview/141
*****************
****************************************************************************
My letter of 3/29/05 to Brett Tyler, Chris Mungall, and PAMGO
To: Brett Tyler <bmtyler at vt.edu>, Chris Mungall <cjm at fruitfly.org>
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: important -- question for PAMGO
Cc: ccollmer <ccollmer at wells.edu>, mlgwinn
<mlgwinn at tigr.org>, david_bird
<david_bird at ncsu.edu>, ralph_dean
<ralph_dean at ncsu.edu>, thomas_mitchell
<thomas_mitchell at ncsu.edu>, arc2
<arc2 at cornell.edu>, jeremy
<jeremy at genome.wisc.edu>, lhannick
<lhannick at tigr.org>, owhite <owhite at tigr.org>,
perna <perna at ahabs.wisc.edu>, rbuell
<rbuell at tigr.org>, setubal <setubal at vbi.vt.edu>,
sutripa <sutripa at vbi.vt.edu>
Bcc:
Attachments:
Brett, Chris, Candace, and other interested parties,
I've received two alternate proposals, from Chris
Mungall and Brett Tyler, for how to solve the
parentage of "defense response to potential
pathogen." I've also created a combined proposal
incorporating elements of both. The following is
an embedded pdf file.
Chris suggests logical cross products to form the
terms "defense response to symbiotic bacteria"
and "defense response to symbiotic fungi." I
think that these terms are acceptable with the
expanded definition of symbiosis now used by the
GO.
Brett suggests changing the name of "response to
symbiont" to "response to potential symbiont,"
and then adding it as a full-time parent to
"defense response to potential pathogen." In
some regards I find this okay, since the initial
responses to another organism are likely to be
tentative and reinforced by cross talk. But on
the other hand, symbiosis presupposes that the
two organisms are already in intimate
association, so I'm not sure about the use of the
word potential here. Also I dislike the idea of
using this as a way to create a term to be used
for annotations of genes known only for their
role in susceptibility responses, without other
evidence as to their role in mutualistic
responses. I think it is bad science to presume
a role for a gene in mutualistic response based
simply on data concerning its role in a
susceptibility response, and again, the GO is not
intended for annotation of diseases or phenotypes.
In the combined proposal I use "response to
symbiont" rather than "response to potential
symbiont" because of my arguments in the
preceding paragraph. But one can argue that any
symbiont is a potential pathogen, and thus
"defense response to potential pathogen" becomes
a de facto cross-product of "response to symbiont
and defense response." We can even add a synonym
of "defense response to symbiont" to make it
relationship clearer. This however, does not
solve the part-time problem because terms like
"defense response to bacteria" are still
grandchildren of "response to symbiont."
All in all I favor Chris's proposal rather than
Brett's or the combined proposal.
I look forward to your comments.
-- Alex
At 2:18 PM -0800 3/28/05, Chris Mungall wrote:
[direct quote of Chris Mungall's letter deleted]
At 5:59 PM -0500 3/28/05, Brett Tyler wrote:
[direct quote of Brett Tyler's letter deleted]
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
Date: Tue, 29 Mar 2005 16:26:06 -0500
To: Alexander Diehl <adiehl at informatics.jax.org>
From: Brett Tyler <bmtyler at vt.edu>
Subject: Re: important -- question for PAMGO
Cc: Chris Mungall <cjm at fruitfly.org>, ccollmer <ccollmer at wells.edu>,
mlgwinn <mlgwinn at tigr.org>, david_bird <david_bird at ncsu.edu>,
ralph_dean <ralph_dean at ncsu.edu>,
thomas_mitchell <thomas_mitchell at ncsu.edu>, arc2 <arc2 at cornell.edu>,
jeremy <jeremy at genome.wisc.edu>, lhannick <lhannick at tigr.org>,
owhite <owhite at tigr.org>, perna <perna at ahabs.wisc.edu>,
rbuell <rbuell at tigr.org>, setubal <setubal at vbi.vt.edu>,
sutripa <sutripa at vbi.vt.edu>
Status:
Chris' proposal is fine with me (with one trivial correction - see below).
As I think about it, the term "Response to
symbiont" already is agnostic of the outcome of
the interaction because "symbiont" spans many
possible outcomes of the association.
Also a clarification: I did not intend to propose
that "response to potential symbiont" would be
used to annotate genes known ONLY for their role
in susceptibility responses, without other
evidence as to their role in mutualistic
responses - I had in mind that annotation would
be applied AFTER their role in a mutualistic
response was discovered.
Finally a trivial correction: Chris' diagram
shows yeast as a distinct organism from fungi.
All yeasts re of course fungi, so any terms
specific to yeast should first be a child of the
relevant fungi term.
Cheers
Brett
[direct quote of my previous letter deleted
--
*****************
Brett Tyler
Professor
Virginia Bioinformatics Institute
Virginia Polytechnic Institute and State University
Washington Street
Blacksburg, VA 24061-0447
Telephone: (540) 231-7318
Fax: (540) 231-2606
Email: bmtyler at vt.edu
Web: https://www.vbi.vt.edu/article/articleview/141
*****************
****************************************************************************
To: Brett Tyler <bmtyler at vt.edu>
From: Alexander Diehl <adiehl at informatics.jax.org>
Subject: Re: important -- question for PAMGO
Cc: Chris Mungall <cjm at fruitfly.org>, ccollmer
<ccollmer at wells.edu>, mlgwinn
<mlgwinn at tigr.org>, david_bird
<david_bird at ncsu.edu>, ralph_dean
<ralph_dean at ncsu.edu>, thomas_mitchell
<thomas_mitchell at ncsu.edu>, arc2
<arc2 at cornell.edu>, jeremy
<jeremy at genome.wisc.edu>, lhannick
<lhannick at tigr.org>, owhite
<owhite at tigr.org>, perna <perna at ahabs.wisc.edu>,
rbuell <rbuell at tigr.org>, setubal
<setubal at vbi.vt.edu>, sutripa
<sutripa at vbi.vt.edu>
Bcc:
Attachments:
Brett,
Thanks for your correction. Here is an updated
DAG based on Chris's suggestion and your
correction:
At this point I'm going to send out an email
digest of these discussions to the pathogenesis
(pathogenesis at geneontology.org) mailing list to
engage a larger group of people in these
discussions.
-- Alex
>Chris' proposal is fine with me (with one trivial correction - see below).
>
>As I think about it, the term "Response to
>symbiont" already is agnostic of the outcome of
>the interaction because "symbiont" spans many
>possible outcomes of the association.
>
>Also a clarification: I did not intend to
>propose that "response to potential symbiont"
>would be used to annotate genes known ONLY for
>their role in susceptibility responses, without
>other evidence as to their role in mutualistic
>responses - I had in mind that annotation would
>be applied AFTER their role in a mutualistic
>response was discovered.
>
>Finally a trivial correction: Chris' diagram
>shows yeast as a distinct organism from fungi.
>All yeasts re of course fungi, so any terms
>specific to yeast should first be a child of the
>relevant fungi term.
>
>Cheers
>
>Brett
>
>--
>*****************
>Brett Tyler
>Professor
>Virginia Bioinformatics Institute
>Virginia Polytechnic Institute and State University
>Washington Street
>Blacksburg, VA 24061-0447
>
>Telephone: (540) 231-7318
>Fax: (540) 231-2606
>Email: bmtyler at vt.edu
>Web: https://www.vbi.vt.edu/article/articleview/141
>*****************
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
****************************************************************************
End of email digest.
Please address further comments to pathogenesis at geneontology.org.
--
Alexander Diehl, Ph.D.
Scientific Curator
Mouse Genome Informatics
The Jackson Laboratory
600 Main Street
Bar Harbor, ME 04609
email: adiehl at informatics.jax.org
work: +1 (207) 288-6427
fax: +1 (207) 288-6131
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment.html
-------------- next part --------------
A non-text attachment was scrubbed...
Name: digraph_Responses_S.pdf
Type: application/mac-binhex40
Size: 25199 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment.hqx
-------------- next part --------------
A non-text attachment was scrubbed...
Name: digraph_Responses_S_extra.pdf
Type: application/mac-binhex40
Size: 25403 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment-0001.hqx
-------------- next part --------------
A non-text attachment was scrubbed...
Name: digraph_Responses_S.pdf
Type: application/mac-binhex40
Size: 25199 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment-0002.hqx
-------------- next part --------------
A non-text attachment was scrubbed...
Name: digraph_Responses_S_extra.pdf
Type: application/mac-binhex40
Size: 25403 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment-0003.hqx
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Combined_Potentials.pdf
Type: application/pdf
Size: 14332 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment.pdf
-------------- next part --------------
A non-text attachment was scrubbed...
Name: digraph_Responses_S-#181EBB.pdf
Type: application/mac-binhex40
Size: 25545 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment-0004.hqx
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Response_and_Detection.doc
Type: application/msword
Size: 59904 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment.doc
-------------- next part --------------
A non-text attachment was scrubbed...
Name: host-pathogenmeeting#16CD95.doc
Type: application/mac-binhex40
Size: 44206 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment-0005.hqx
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Suparna_host_terms_BT.doc
Type: application/msword
Size: 65024 bytes
Desc: not available
Url : http://fafner.stanford.edu/pipermail/pathogenesis/attachments/20050330/ba820d86/attachment-0001.doc
More information about the Pathogenesis
mailing list