All,
I've been following the discussion here with interest the last two weeks
- with the Neurocommons project, Science Commons is taking on both
issues of intellectual property (on ontologies and databases) and the
semantic web in neuroscience. We're text mining the open content and
indexing with public ontologies with a focus on autism and epilepsy;
draft RDF release is planned in the november time frame.
For what it's worth...we were planning on holding a Neurocommons meeting
in the late fall / early winter and I'd be happy to expand that meeting
to more of a global "meeting of the minds" between SWeb and Neuro, if
that's of use. Let me know...
jtw
kc28 wrote:
Hi Bill,
Thanks for your passionate response. When I said "outreach", I did imply
to establish a mutually beneficial relationship between the semantic web
and neuroscience communities. I agree with you that such a relationship
would help bring scientific/techological advances to both communities. I
also agree that it's an excellent idea to have a face-to-face meeting
with the neuroscientists you listed. However, it could be a challenge to
get all of them agree to come and show up at the same meeting. I don't
mean we shouldn't try. I think we should try even if we can only get
some but not all of them. Also, I think we should also invite folks from
NCBO and MGED to join if possible. In addition to the face-to-face
meeting, I think we can still try to invite these neuroscientists (as
well as some of the NCBO/MGED folks) to participate in some of future
telconf's to establish an ongoing interaction. For your suggestions on
the BioRDF wiki pages (I think they are very good suggestions), I
suggest that we set it as one of the agenda items to discuss in our
BioRDF telconf call. Other folks may also have other suggestions.
Cheers,
-Kei
William Bug wrote:
I do run on, sometimes, don't I, Kei?
I emphatically agree with the general tenor of your suggestion.
I would word it a bit differently.
I wouldn't call this outreach so much as going to the "customer" and
asking them to help us - the technology experts - to define their
user requirements. I would word it this way to the technologists, at
least. The Neuroscientists should be pitched using "civilian"
colloquialisms, but the point is I believe the onus is on those
developing and applying the technology to stay in sync with the needs
of the neuroscientists.
I realize many of us on this list are in fact trained biomedical and/
or computer science researchers. I myself was originally trained as
a molecular biophysicist studying neuromodulation of presynaptic,
Voltage-dependent, Ca++-channels using single-channel and whole-cell
electrophysiological techniques. That places us at the extremely
valuable nexus where we possess specific insight into the information
needs of broader community of neuroscientists we hope will benefit
from the technological resources we develop, while also possessing
the technological insight required to determine what is practical.
My sense is it's important to develop credibility on both sides of
this equation - the technology developers need to clearly demonstrate
they're sensitive the needs of "bleeding edge" researchers. They are
developing tools to revolutionize a scientist's ability to perform
their research tasks effectively and efficiently - transform them
from 19th century cottage scientists where all knowledge mining must
be done laboriously and with very limited scope by their lonely brain
into 21st century informaticists where large scale, data/knowledge
mining against the evolving "World Brain" (H.G. Wells term - http://
sherlock.berkeley.edu/wells/world_brain.html) is a routine practice.
The scientists also need to demonstrate they recognize the value
provided by the technologists. This will again derive from clear
demonstrations of the value the technological solutions can provide
to the researcher. This latter issue is often a hard one to get
across, but its lack of such recognition/trust that can lead the
technologists to go at it on their own out of frustration (Kei, Don,
and others who attended the Human Brain Project meeting in April can
attest to the fact that I am just as subject to this frustration as
any other bioinformatics developer - :-) ).
Along these lines, I'd suggest:
1) Presentations by neuroscientists who have done seminal work in
neuroinformatics:
I think Kei's suggestion is an excellent. However, I'd suggest a
F2F meeting, where these folks are invited as speakers. It will be
hard to get the full effect of what they have to say on a phone or
video conference. They are likely to take a talk at a meeting more
seriously and a greater level of commitment is likely to derive from it.
I would suggest there be a session of neuroinformatics
presentations by neuroscientists, and also a session of semantic web
technology presentations by participants of this group. The focus
should be on neuroinformatics projects using semantic web technology
with one intro talk on semantic web technology applied to biomedical
informatics to provide a context for those neuroscientists who've not
yet got the take home message.
My suggestion for neuroscientists would be - in no particular
order of importance:
1) Gordon Shepherd (SenseLab) - integration of various
modalities of neuro-data with a focus on the olfactory system
2) Doug Bowden (NeuroNames) - unified, mammalian
neuroanatomical lexicon
3) Maryann Martone (CCDB, SMART Atlas, & BIRN) /Mark Ellisman
(BIRN)/ Jeff Grethe (BIRN infrastructure) - broad-field, neuroimaging-
centric neuroinformatics infrastructure
4) Rolf Kütter (CoCoMac) - literature informatics
("bibliomics") system with a focus on neuro-connectivity
5) Rob Williams (GeneNetwork/WebQTL/Mouse Brain Library) -
genetic variability and brain phenotypes from molecules through
anatomy and behavior
6) Peter Hunter (CellML and parametric spatial modeling of the
brain)
6) Dan Gardner (BrainML) - XML schema for neuroscience data
There are other folks, but I believe this core of people cut across a
variety of neuroscientific sub-domains and levels of technical
complexity. I'd also recommend someone from the field of 3D digital
brain atlasing (atlas data set/computer vision algorithm/atlas tool
development), but as I'm in this field myself, I don't feel it's
appropriate for me to suggest which of the several researchers would
be the most appropriate. I would only say it's important to
recognize the distinction between spatially-based, neuroscience data
sets (GENSAT, Allen Brain Atlas, Desmond Smith's "voxelized"
microarray data sets) and the use of brain atlases to provide a
canonical coordinate space and algorithmic tool set via which one can
perform large-scale integration & atlas mapping of spatially-based,
neuroscience data sets. This task - integration of spatially-mapped
neuroscience data sets - is obviously one for which semantic web
technologies will be a critical catalytic factor.
2) The BioRDF Wiki page:
I'd suggest this focus on semantic web applications in the
neuroscience. There is already a link to a list of projects (e.g.,
SWAN, Semantic Synapse, NeuroCommons). Rather than place substantive
info on these 3 projects 3 clicks away, I'd suggest you list them
right there on main BioRDF Wiki along with a 1 - 2 sentence summary
of each project. This will guarantee the widest possible recognition/
visibility for these efforts.
I'd also suggest that in listing of "other" neuroscience
resources on the web, rather than creating an ad hoc collection of a
few projects (which can effect general credibility - e.g., "Where are
all those neuroscience resources I think are important - why just
BrainML & GENSAT?" - I'd point to the several consortia and/or
registries/"yellow pages" already compiled - e.g., the Society for
Neuroscience's Neuroscience Database Gateway (http://big.sfn.org/NDG/
site/), David Kennedy's Internet Analysis Tools Registry (mainly
neuroscience tools, though this scope is expanding - http://
www.cma.mgh.harvard.edu/iatr/display.php?spec=all), fMRI Tools
(http://www.fmritools.org/), The Neuroinformatics Portal Pilot
(http://www.neuroinf.de/), etc.
3) Licensing:
To say one final thing about licensing, I completely agree with Don
that it is a hideous, unworkable mess. Go back to the single
statement in Article 8 of the U.S. Constitution, and you clearly get
the sense of what was originally intended by establishing copyright
and patent law as a legal entities (http://www.archives.gov/national-
archives-experience/charters/constitution_transcript.html):
"The Congress shall have Power...To promote the Progress of Science
and useful Arts, by securing for limited Times to Authors and
Inventors the exclusive Right to their respective Writings and
Discoveries;"
It was recognized even 200 years ago the creative commons is of great
value to society. For this value to be realized, these resources
must be a part of the commons and available to all - including latter
day inventors, artists, and scientists seeking to build on what came
before. This need, however, must be balanced again the desire of the
artist, scientists, inventor to make a productive living from the
fruits of their labor (otherwise, the creation stops).
I'd guess most folks on this list would certainly agree with the need
to establish this right. Where the founders went wrong was in the
statement "The Congress shall have Power To...", as this left the
door wide open for Congress to redefine what copyright was all
about. As most of you probably know, the balance began to shift from
the "...Authors and Inventors (and scientists)..." to publishers
(those solely in business to make $$$ off the efforts of the creative
persons) starting in the late 19th Century with the proliferation of
pirated sheet music. This trend worsened through the last century,
but really took a significant, qualitative leap away from the
original intentions as outlined in Article 8 above with the DMCA.
Given how significant a driver IP is for the engines of the economy
(and greed), I'm still uncertain how we can over turn this trend and
get back to the original principles. The work sponsored by the
CreativeCommons - and specifically The ScienceCommons - will
certainly help to get us there**. This is the case despite the
extremely clear detriment the current trend has toward society as a
whole*** and to the communication amongst scientists in particular.
Though still problematic, I actually endorse the use of licensing by
the NeuroNames folks (as you might have been able to gather already),
as I see their application going right back to that original
statement in the U.S. Constitution. It's one thing to bulk download
sequence records and "cleanse" their semantic content in order to
promote powerful knowledge mining efforts. When it comes to highly
curated, knowledge resources, the onus is on the user to be careful
both to clearly understand the original intentions and limitations of
the resource, as well as to work to protect the integrity of the
resource. It does none of us any good to create a "better" or more
"open" NeuroNames, if that just becomes another version of
NeuroNames. If we are not ALL using the same NeuroNames (or at least
using compatible and consistent versions), then we defeat the purpose
of using NeuroNames for large-scale data integration and semantic
mining.
What is needed is for there to be an established authority to
arbitrate when issues of curation and usage of a knowledge resources
come into conflict. Here again, I'd suggest going to NCBO for help.
Not that they have an infinite supply of resources and can solve all
the problems, but at least they understand this complex issue from
both sides - that of the curation authority and of the biomedical
informatics scientist trying to make productive use of the resource -
and have some resources and authority to grease the wheels of science
in this domain.
Again - just my $0.02. I hope this helps to clarify what I've been
trying to communicate in this thread.
Cheers,
Bill
** I expect it's a bit superfluous to mention here, but I'd suggest
checking out the SC info resources, if you've not already at http://
sciencecommons.org/resources.
***see the excellent article by Richard Nelson posted by John
Wilbanks on the Science Commons weblog a few months back [http://
sciencecommons.org/weblog/archive/2006/02/15/richard-nelson-on-the-
scientific-commons] for an excellent treatment of how this directly
impedes the pursuit and accumulation of scientific knowledge.
On Jun 6, 2006, at 7:42 PM, kc28 wrote:
Hi Bill,
You really can write faster than I can read :-). Actually, we have
discussed in a previous telconf about how to outreach to the
neuroscience community. I think this represents a good opportunity
to try to get people like Doug Bowden involved, as we are interested
in converting Neuronames into RDF/OWL. I wonder if it's possible to
invite neuroscientists like Doug Bowden and Gordon Shepherd (and
possibly more) to talk about their work in our future
BioRDF/Ontology telconf. This will foster more interaction between
the semantic web community and neuroscience community. I wonder how
this sounds to other semantic web folks.
Cheers,
-Kei
William Bug wrote:
Dear Matthias,
I would strongly recommend you contact Doug Bowden and colleagues
at NeuroNames before you undertake this task - or at least take a
look at the NeuroNames specifics I list in my previous email. I'd
be glad to answer any questions you may have about statements I
made. Doug and his collaborators are extremely collegial and make
a very sincere effort to work with those interested in making
effective - or novel - use of NN.
The other person you should contact is Daniel Rubin at NCBO, who,
for all I know, is lurking on this thread. Others in the thread
appeared to be addressing Daniel. This is a topic actively under
investigation both by NCBO and by the BIRN.
As I mentioned in my post to this thread, Doug & colleagues have
been working for the last year with Jack Park of SRI to express NN
in XTM format. A lot of effort needs to go into vetting this
"remapping" to make certain none of the assertions in the
hierarchy - explicit or implicit - are invalidated - as well as
ensuring no new assertions are unwittingly introduced. You may
want to work from this version of NN to create an RDF/OWL
version. As I mentioned in the previous post, there has been some
substantive effort to examine the differences and similarities
between XTM & RDF - and there may even be translators or XSL
instances that can get you most of the way.
Doug also distributes the entirety of NN on CD with all of the
latest work they've done in the past year to incorporate rat &
mouse neuroanatomical terminologies - an added dimension
absolutely critical to those of us interested in collating
microarray, in situ & IHC expression studies in mouse brain with
neuroimaging data sets and 3D digital brain atlases.
There is definitely a need for an open source, RDF/OWL version of
NeuroNames (and the neuroanatomical portion of RadLex for that
matter - http://www.rsna.org/RadLex/ - if you are interested in
human, radiological imaging of the brain).
I believe we must do our best to work with the curators/developers
on these various knowledge resource projects, given the
biological complexity embedded in these resources.
As far as the licensing goes, Doug realizes this is a thorny
issue. The initial license was merely put in place to avoid
others downloading this highly curated knowledge resource,
modifying it, then repackaging it as "NeuroNames." As I
mentioned, this was not a paranoid fear. The license was imposed
in response to someone actually having done this with NN.
Knowledge resources like this - even when they are just
terminologies - require careful curation, and uncontrolled
dissemination and modification can ultimately degrade the
usefulness of the resource.
Of course, closed, proprietary licensing can also degrade its
usefulness, so there is a delicate balance that must be struck.
This is an issue I believe NCBO can help us all to resolve. They
won't have all the answers, but may be able to sponsor a means to
derive an effective solution to this problem.
My recommendation is a statement be sent by the W3CSW HCLSIG -
maybe the BioRDF & BIOONT groups collectively - informing Doug of
the need as they see it. He will not be surprised by the nature
of your request, but will be very surprised and pleased to see
this need emerging from the semantic web community. I don't
believe he reads this list. I know he will be happy to work with
participants on the W3CSW HCLSIG to get us what we have all
identified as essential - an open source, unified neuroanatomical
terminological (and in association with FMA - as Neuro-FMA -
ontological) resource all formal annotation efforts can make
shared and productive use of.
Just my $0.02 on the topic.
Cheers,
Bill
On Jun 6, 2006, at 3:38 PM, Matthias Samwald wrote:
Hi Kei,
I am under the impression that the neuronames ontology available
on their website (as an Excel file...) is different from the
version that is licensed as part of the UMLS. I guess the version
that is online is a newer version of the one incorporated in
UMLS. However, this might be seen as a derivative work, so it
might still be restricted. In that case, it would seem like
people of the neuronames group are violating the licence
restrictions themselves (by making it available on the internet).
I will write them and ask about that.
kind regards,
Matthias
Hi Matthias,
Thanks for doing that, but do we still have the licensing issue as
stated by Olivier?
Cheers,
-Kei
Matthias Samwald wrote:
I will convert the neuronames - ontology to SKOS (an OWL ontology
used for the representation of taxonomies / theasauri). It will
be added to the extension of the bio-zen ontologies framework
[1]. I will keep you updated.
kind regards,
Matthias Samwald
[1] http://neuroscientific.net/index.php?id=download
On Mon, 05 Jun 2006 21:17:55 -0400, kc28 wrote:
For more up-to-date information about neuronames and related
tools, please visit: http://braininfo.rprc.washington.edu/.
While building our own open neural anatomy is one option,
getting the neuroscientist (e.g., braininfo people) involved if
possible may be another option (outreach to the neuroscience
community?).
Bill Bug
Senior Analyst/Ontological Engineer
Laboratory for Bioimaging & Anatomical Informatics
www.neuroterrain.org
Department of Neurobiology & Anatomy
Drexel University College of Medicine
2900 Queen Lane
Philadelphia, PA 19129
215 991 8430 (ph)
610 457 0443 (mobile)
215 843 9367 (fax)
Please Note: I now have a new email - [EMAIL PROTECTED]
This email and any accompany attachments are confidential. This
information is intended solely for the use of the individual to
whom it is addressed. Any review, disclosure, copying,
distribution, or use of this email communication by others is
strictly prohibited. If you are not the intended recipient please
notify us immediately by returning this message to the sender and
delete all copies. Thank you for your cooperation.
Bill Bug
Senior Analyst/Ontological Engineer
Laboratory for Bioimaging & Anatomical Informatics
www.neuroterrain.org
Department of Neurobiology & Anatomy
Drexel University College of Medicine
2900 Queen Lane
Philadelphia, PA 19129
215 991 8430 (ph)
610 457 0443 (mobile)
215 843 9367 (fax)
Please Note: I now have a new email - [EMAIL PROTECTED]
This email and any accompany attachments are confidential. This
information is intended solely for the use of the individual to whom
it is addressed. Any review, disclosure, copying, distribution, or use
of this email communication by others is strictly prohibited. If you
are not the intended recipient please notify us immediately by
returning this message to the sender and delete all copies. Thank you
for your cooperation.
--
..................................................................
John Wilbanks
Executive Director
Science Commons
http://sciencecommons.org
[EMAIL PROTECTED]
..................................................................