Mission and Vision
Oliver Ruebenacker
Some suggestions about the future of this effort:
(1) So far, meetings have been based on a small group of active
regulars, and often, not enough of them are available to have a
meeting, which causes uncertainty and frustration. There could and
should be more participants, as there are many people both in the city
and in the world who do use semantic web for biological pathways and
therefore would benefit from us and we from them. We should try to
reach out more and gain new members, both in the Boston area and
beyond.
(2) We started out with the plan to develop an OBO-compliant
alternative to the current draft ("BioPAX-DX") for BioPAX Level 3 and
chose our name - BioPAX-OBO Workgroup - accordingly. The status of the
original plan is at this time, at best, unclear. Meanwhile, I think we
are better described as a group of people exchanging ideas about
Semantic Web methods applied to biological pathways, which has more
tangible immediate benefits than the original plan. To attract new
people, it would be wiser to change our name to reflect this: Some
name that contains the keywords "pathway" and "semantic" would appeal
to many more people than our current name, which appeals only to those
who know and understand the benefits of BioPAX and OBO.
My suggestion for a new name would be "Semantic Pathway Club".
(3) All or almost all of our active regulars seem to be member of
the Health Care and Life Science Interest Group (HCLSIG) of the World
Wide Web Consortium (W3C). The HCLSIG has a handful of sub groups, and
I have long had the idea of suggesting to form a new subgroup for
pathways. I don't know what the process or the requirements are to
form a new subgroup, but my impression is that if a handful of people
show interest, it would be done. I also assume that we could continue
our current meetings as meetings of the subgroup and that it would be
possible to have the meetings of the subgroup open to non-members.
These meetings would then be joint meetings of the HCLSIG subgroup and
us (with most of us being part of both). I would assume that the fact
that something is already happening rather than building something
from scratch would sound good to the HCLSIG.
Becoming an HCLSIG subgroup has a number of advantages. They provide
great infrastructure, including Wiki and meetings that combine web
chat and phone conference. Essentially the type of things we are
using, but provided in a more unified and professional way. We might
also get a staff member of the W3C to help us, perhaps with a more
professional approach to convening our meetings than I can provide,
and perhaps with the capability to book rooms at the MIT (I always
have to ask Jonathan, which complicates matters).
But most importantly, our work would be known to a larger audience
and connect to other interested parties.
I am also planning to build collaborations on SBPAX and hope to one
day submit SBPAX to the HCLSIG for consideration as a W3C
recommendation.
Take care
Oliver
--
Oliver Ruebenacker, Computational Cell Biologist
BioPAX Integration at Virtual Cell (http://vcell.org/biopax)
Center for Cell Analysis and Modeling
http://www.oliver.curiousworld.org
Michel Dumontier
Andrea Splendiani
Michel Dumontier
Oliver Ruebenacker
That's why I was talking about joint meetings. You have to be a
member to be a member, but I am assuming that if the subgroup so
decides, non-members can participate.
Take care
Oliver
andrea splendiani (RRes-Roth)
decision on invited expert status is taken by HCLS itself.
While HCLS gives some visibility, it also makes it difficult to many to
participate.
Not only, remember that at the beginning of this charter they wanted to
make the mailing list private.
It's clear anyway that some affiliation will help us in visibility and
in attracting people. We can also simply register a website, define a
group called something like "pathwayandnetworkontologies working group",
define a charter/mission, regular teleconf and perhaps an annual
meeting. That would be enough to give some structure and an identity.
We can also thinking about joining with other efforts as
semantic-systems biology. Or to explore whether there is some space for
affiliation with other institutions as NCBO.
Bottom line: HCLS has a double sided effect. Gives visibility, but
easily keeps people out.
Ciao,
Andrea
Oliver Ruebenacker
You are right that a subgroup does not have the power to grant
status as invited expert of HCLSIG. But all I am talking about is the
capacity to participate in subgroup meetings only, including phone and
chat conference. Participation in HCLSIG itself is a different story.
Bottom line: We gain much and loose nothing. And HCLSIG also gains
without loosing.
Also, it's not as if the subgroup would now be flooded by a large
number of non-members. Right now, I think, all of our active
participants are HCLSIG members. If new people join us, they might
choose to join HCLSIG as well. The point where a substantial number of
participants are not HCLSIG members is still far down the road, and
until that happens, if ever, we can come up with a new arrangement
that fits all.
Your other suggestion - website and participating in other
organizations - sounds great, too, and one does not exclude the other.
I would say, go for it, find us some groups to join and let us network
as much as possible.
Take care
Oliver
Andrea Splendiani
Ciao,
A.
Apologies for typos, Sent from an iPhone
Oliver Ruebenacker
Sure, but let's first agree on a name and a mission statement.
Take care
Oliver
On Tue, Feb 17, 2009 at 6:18 PM, Andrea Splendiani
JSL913
I want to think about this some. I'll read over this thread and give
it some thought. My comments off the top of my head are:
There may be either a more appropriate group, or way to participate
(with HCLSIG). My sense is (a) if we can identify what part of the
solution we provide, we have something to offer (b) if we have a
compelling use case that is of value because it solves an problem then
we can get others to participate more easily.
So, my suggestion is to ask you to think about or research that - what
are the problems that our effort would solve? and who cares about
those problems?
Joanne
andrea splendiani (RRes-Roth)
Starting to brainstorm about a mission statement...
I think what we are after is the integration of pathways and networks
beyond a syntactic level. In particular the reconciliation of entities,
reactions and pathways across representations.
This is something that the "ontology" in biopax was intended to do,
although the priority of biopax "dx" are making data available in a
common syntax, with a minimal semantic commitment (or better, in case of
tradeoff between the two, it went for "more data").
There has been some experience time ago by Alan and Jeremy (debugging
the bug, after the biopax-manchester meeting), and some relevant work my
Michael. I think also SBPAX fits in line with this goal.
Beside HCLS, it would be interesting to know which level of "contact" we
can have with NCBO. There has been a workshop on ontologies for cellular
networks time ago (organized among others by Alan). Is there some
possibility to be somehow affiliated, like to have a mailing list
managed by them ?
Another important thing, I think, is to arrange a f2f meeting.
JSL913
That's a good start.
You might want to read my paper with Robert Stevens that speaks
directly to integration levels and issues using BioPAX (e-science and
pathways) http://www.biomedcentral.com/1471-2105/8/S3/S3. It's not
the best paper, but it speaks to using biopax to meet the needs of e-
science. The tutorial Jeremy and I did for ISMB 2005 also has good
material.
For me, one of the important aspects (in addition to all the
semantics) is that when one of the databases dumps their data into
biopax, they lose info.
Another important issue is - evaluation - of the 200+ pw databases -
how many have biopax output? How much of their data can be
represented in biopax? what key concepts are kept, what are lost and
what are the implications (for use cases) of what's kept and what is
lost. What questions can and can't be answered?
BioPAX OWL1.1 vs BioPAX OWL 2 vx BioPAX OBO?
Joanne
Oliver Ruebenacker
I want to do outreach. There are thousands of people and dozens of
groups out there who use Semantic Web for cellular pathways. Building
a network of them would create a lot of benefits, including more
people participating in our meetings.
So I suggested to our group to undertake the outreach effort. Most
responses have been very encouraging. However, I have also heard, from
a minority, reservations against outreach, and the insistence that we
can only proceed by consensus, meaning one single person can shut the
effort down.
So let me be clear about this: I want to invite any one who may be
interested and build formal relationships with any group who wants to
participate. And I don't want to depend on every single person. I want
elected leadership and majority decisions.
If this is not compatible with this group, then it should be a
separate effort. We will cooperate. In particular, the outreach will
include advertising of our meetings. And each of you will be invited
to join the effort, of course.
Take care
Oliver
JSL913
I am 100% in favor or outreach.
As far as I am concerned: "Go got it."
On Feb 18, 2009, at 10:20 PM, Oliver Ruebenacker wrote:
Joanne
andrea splendiani (RRes-Roth)
recall having seen reservations.
I think the more inclusive we are the better. If the group gets too big
to be un-manageable this way, we'll see.
We have to find a way to avoid problems with some individuals as there
have been in biopax, but a do-ocracy should work (not sure about the
spelling, though).
Said this, to go on with outreach, we need:
* Mission statement. We have started some discussion about this.
* Something that identify us, a name, a website, a mailing list
* A "seed". Who are we ? Who supports us ? Which guarantees do we offer
to people that wants to get involved that we are serious ?
The last point is tricky.
So, from a strategic point of view, I think it is better to build a core
of people/organizations first. Call it, if you want, controlled
outreach.
If we have a stable and representative core, we can go on enrolling more
people (even wildly).
What do you think ?
Can start to name the people/organizations that are, so far, interested
in this effort ?
Oliver Ruebenacker
The driving force is that we offer people who are interested in
using Semantic Web on cellular pathways to meet others who do the same
to network and exchange information on research, projects, meetings,
institutions, software, standards, recommendations, collaborations,
jobs, grants and policies.
We can reach out to people related to BioPAX, SBML, VCell, OBOF,
NCBO, HCLS, etc.
Doocracy (sp?) is a great attitude, but not a leadership structure.
We like people to be as productive as they can and discuss as much as
they want. The problem is only when a group's resources are at stake,
such as the group's data, website or mailing lists, or if nothing
else, the power to use the group's name and to define what the group
stands for.
I understand that after we got burned by a deadlock with certain
individuals, it is tempting to think, if only those individuals are
gone, we can move ahead. But I can guarantee you that any group, once
it moves beyond the listen and learn stage to actually achieving
specific aims, will find that some of its members have irreconcilable
differences. The question then is not, can we reconcile all
disagreements, but can we move ahead in spite of them. Instead of
waiting until every one agrees, you define procedures to make
decisions if people do not agree.
Besides, outreach means getting to know new people, and you don't
know how those new people will turn out. Also, what if some one
suggests in front of the whole group to invite some individual that
you prefer not to have in the group? Would you want to argue in front
of the whole group why not to have this individual?
So I think the trick is not trying to avoid problematic individuals,
but to make sure they can make no harm.
In the very beginning, we start with selected people, until we have
a format in place. Then, we should invite everybody.
Take care
Oliver
On Thu, Feb 19, 2009 at 5:06 AM, andrea splendiani (RRes-Roth)
Andrea Splendiani
So:
Name
Mission
Initial seed (people).
(later on... governance)
Is all the people on this list interested ?
ciao,
Andrea
Il giorno 19/feb/09, alle ore 16:55, Oliver Ruebenacker ha scritto:
Alan Ruttenberg
http://neurocommons.org/page/BioPAX-OBO
-Alan
andrea splendiani (RRes-Roth)
This thread originated from the consideration that we should have a more
general scope, as we didn't yet produce a sound mapping of biopax, but
we ended up having more discussion on the representation of networks and
pathways.
Also the name, biopax-obo, that doesn't refer anymore much to the
original goals (albeit "new ones" are not defined).
Should we add notes to this page ?
Oliver Ruebenacker
What name should we give the new outreach effort?
First, there has to be a reference to Semantic Web. The word
"semantic" is most precise, but "web" is more familiar, and these
days, any one interested in the Web would be interested in the
Semantic Web.
Next, there should be a reference to biological pathways.
And it should be as short as possible.
Here is an idea: BioPathWeb
And guess what, Google finds nothing under "BioPathWeb"! No one
seems to have taken this seemingly obvious name. That's an
opportunity, isn't it?
Take care
Oliver
On Thu, Feb 19, 2009 at 12:40 PM, andrea splendiani (RRes-Roth)
JSL913
semantic biopathwebs
sbpw
(just a (weird) thought)
Jonathan Rees
History: Science Commons needed an OBOF-compatible biomolecular
interaction ontology so that we could do effective integration of the
same kinds of information that BioPAX-DX is targeting. The need was
felt by others, and we started talking. The problem now is that we
seem starved for labor. I'm sorry I don't have time to work on it; I
was really praying that you all would do the work for me. If Science
Commons comes into money we'll hire someone to do this. If you want to
do outreach in order to recruit more hands to the job that's fine. I
wish you luck, but the project is known on the BioPAX list and in the
OBO community, so I thought we had identified all the people who cared
already. I don't see what this group's challenges have to do with
semantic web (whatever that is) or control issues (do-ocracy vs.
consensus). It's mostly a matter of getting a job done. With a good
goal in mind technical argument is usually productive. It's only if
you don't know what you need, or if differences in goals are not
surfaced (e.g. if "use OWL" means different things to different
people), that things get acrimonious. But I don't see acrimony or even
decision-making as being a problem in this particular group. There is
plenty to do, it is just a matter of doing it.
Getting better organized is good. Making sure that everyone who might
contribute has a chance to is good. Knowing what you want to do, and
doing it, are even better.
It's important to keep focus. If you need to start up a new activity
of the kind you describe, create a new project with a new name and new
discussion group. I would just be sad about competition for
person-hours - selfishly I'd rather you all work on what I need, which
is a principled, open semantic model of biomolecular interactions
suitable for data integration, than something else, like servicing the
linked data community.
If you think continuing with an unchanged goal under a new name would
help for recruitment, that's a different story, and we should talk
about it. But I am skeptical that the name would be a deciding factor
in anyone's mind. I could be wrong but I would think BioPAX and OBO
both have name brand recognition among the people who are experienced
enough to contribute, so the name ought to help, if anything.
By the way the "do-ocracy" of the BioPAX group was what led to its
becoming closed after supposedly being open. Basically it means that
the first person to get a grant gets to take control of the whole
project. That's not how any of the OBO projects works.
Jonathan
andrea splendiani (RRes-Roth)
So... I didn't want to get into governance discussion... I agree with you and it's anyway not a priority at this stage.
What I wanted to say (in one of the earliest emails of this thread) is:
>Getting better organized is good. Making sure that everyone who might
>contribute has a chance to is good.
>Knowing what you want to do, and
>doing it, are even better.
>help for recruitment, that's a different story, and we should talk
>about it.
*) At least for me it would be better to have a monthly meeting that people commit to attend, and that will take place even if people is absent, than a more frequent meeting, but with random attendance. Can be that this is a problem of mine as I'm not in the Boston area... but this model in practice pushes this project down my priority list.
Can we try a monthly meeting with notes, TODO and so on ? This doesn't prevent to have more "informal" ones...
*) Our scope it's not so clear, at least for me. It's clear in the goals, not so much in the methods: strategies based on, say, topological comparisons of pathways (or any other method not related to ontologies and OBO), are they within our scope ?
*) As for the name/mission/outreach and so on, I agree that biopax-obo clearly identifies the target community, but perhaps is keeping outside people that don't commit to biopax, but who are anyway dealing with RDF/OWL representations of pathways, interactions and ontologies... and for which biopax-obo would be the next step (same goes for linked data and semantic web... but this is a weak opinion). Do you think that having "biopax" in the name is helping or "hurting" us ?
Oliver Ruebenacker
What could be more philosophical than a historical argument?
I am not aware of the needs of Science Commons being a factor in our
discussions so far, but that doesn't mean we can not deliver what
Science Commons needs. Do you have a more detailed formulation of what
Science Commons needs and why?
SBPAX is linked extensively with BFO to be OBOF-compatible and I
have software to convert from BioPAX to SBPAX. Is that what you need?
Decision-making has not been an issue in this group, because no
tough decision has been on the table so far. We have disagreements
that have not been resolved yet. We have formulated goals loosly
enough to disagree on their interpretation, or their priority. Every
one thinks they know what they want and their own goals good. Once the
final product moves closer, tough decisions will have to be made.
Without a process other than doocracy or consensus, we will be unable
to make these decisions.
Take care
Oliver
Jonathan Rees
On Feb 20, 2009, at 9:30 AM, Oliver Ruebenacker wrote:
> Hello Jonathan, All,
>
> What could be more philosophical than a historical argument?
>
> I am not aware of the needs of Science Commons being a factor in our
> discussions so far, but that doesn't mean we can not deliver what
> Science Commons needs. Do you have a more detailed formulation of what
> Science Commons needs and why?
Not detailed, but I can outline the general shape. The Science Commons
Data Project is trying to explain and demonstrate to the world the
value of porting data and "knowledge" to one (or, if necessary, more
than one) common 'platform' - think Linux distributions and their
populations of packages, noting that you can load packages together
into the same system build by virtue of their having and using more or
less compatible interfaces. The demonstration platform consists of
RDF, OWL, SPARQL, and other web standards, and a method for steering
'knowledge representation' so that the same information tends to be
expressed in the same or similar ways even when it comes from
different sources - this is hard, and while OBO Foundry methodology
may not be perfect, it's the best thing we've got, and it has a great
community around it, so we're pushing its use for this goal.
So for pathway information, such as that from Reactome, KEGG, Ecocyc,
etc., which we would definitely like to include in our distribution
and triple store ASAP, the best path forward we've come up with is to
do ports that are along the OWL and OBO Foundry lines. If there's a
better way that has the promise of promoting scalable data
integration, I'd like to hear about it. The bottom-up approach
advocated by Tim Berners-Lee and others (including, I would say,
BioPAX DX), where you just do whatever's easiest and wait to fix
incompatibilities until you get stuck, seems to me less efficient than
the more coordinated approach taken by OBO (GO, etc.), but only time
will tell how the mix of coordination practices will evolve.
So if you want to do something for Science Commons and its mission,
figure out how to represent Reactome et al. in a way that will
contribute to a scalable data/knowledge ecosystem, i.e. has the best
possible chance of permitting graceful connections with unknown other
information (phyolgenetic, chemical, classical genetic, anatomical,
simulation, etc.) and attempts to forestall the threat of future
refactoring crises.
My guess is that your interest is not too far off from this; I'm just
stating my project's viewpoint so as not to presume about anyone
else's needs.
We're not paying you, so anything you do that helps is a gift.
Jonathan