Sorry for the grammar in my last post folks- fast fingers and spell check +
on the run = gibberish.
I may be able to finally add some value to the discussion, and am happy to
assist with papers, demo, or anywhere else if I can add value. I don't do
much coding these days, but I am an expert (if anyone is-questionable) on
e-commerce, e-learning, economic modeling and adoption with several notable
firsts coming out of our small shop since '95 (don't be first- martyrdom is
tough on loved ones). I just filed a patent amendment to our Kyield system
so it will hopefully be published soon for anyone with deeper interest-
brings together over a decade of work, much of which is relative to this
thread.
Regarding sources- While I review a fair volume in my VC work, I don't
consider myself an expert in LS R&D, much less any of the dozens of
specialties within the broad area of study- they should be the primary voice
on sources needed. I have received correspondence however that suggests talk
and some pressure recently on the FDA in making more of their data available
(from academia and corp.), which apparently many believe would accelerate
prevention and therapy if not cures on many fronts. So that would be one
huge example in the public sector. But we can expect that as has been the
case all along, when it comes to intelligent data representing knowledge,
much of it will remain private. The medium threatens many industries
(whether real or imagined the result is similar), and the battles have been
brutal at times.
We have two primary interests and goals here from my perspective:
1) Immediate- demonstrating the acceleration of R&D within the broad and
highly complex environment known as life science. It's pretty clear to me
and most engaged here (apparently) that ontological languages are already
playing an essential role in this goal, but the combination of lack of
sources (Alan Brain Institute addressed as shown in your example- $100
million), trained technicians, and adoption of standards remain challenges.
For drug development and genomics among others, we know that high volume
data manipulation and collaboration can and do lead to accelerated
discovery. A big part of the problem is one of culture, which is the most
important aspect of knowledge systems in my and others' view. On one hand we
have sub-cultures that will share anything and destroy organizations- and
even cost the lives of large numbers of people, while on the other extreme
we have sub-cultures that are so intent on protecting agency turf that it
creates the same result. Both of these extremes have been evident within the
same broader U.S. Government over the past few years as partial causes of
major disasters- an excellent body of case history. I have made the argument
with members of congress, agency heads, CIOs, etc. that our Kyield system
could have prevented same, simply due to embracing more intelligent systems
as discussed here in combination with equally intelligent organizational
design. Lots of talk- little action.
So I argue in enterprise architecture to manage culture intelligently first
and always. Otherwise the tools won't be used properly if used at all, and
once cultures deteriorate sufficiently then a negative spiral of decision
making can occur- and that can be catastrophic- in any type of organization.
My guess is that even private life science companies could share far more
data safely than many believe, within the right structure of course. But as
my old friend and colleague Russell Borland said a decade ago in one of our
GWIN forums on the topic- free only works if everything is free. Sacrifice
and investment need compensation as bills need to be paid. Structure
matters.
One of the main themes currently in life sciences in the VC world is an
attempt to show the giants that collaboration with smaller firms would be a
more intelligent path. I'm thinking that this specific target would be ripe
for this type of demo perhaps in showing how a similar collaboration between
entities can actually lead to a hypothetical discovery.
This WG is I think doing an excellent job of helping to demonstrate the
potential for acceleration of solutions with a more intelligent web, whether
public, private, or some combination thereof. And I think that life science
is among the best communities to target the broader technology for adoption.
The industry cluster has serious problems that affect us all, not least of
which is that it requires $1 billion to bring the average drug to market,
meaning that very few can even begin the journey, and also very likely that
most affective therapies die long before potential value can be known.
2) Mid-term- demonstrating to other clusters, including multi-disciplinary,
the value of semantic technologies to life science that can then be applied
to their industry or discipline. Based on my own discussions with a couple
of other industry leaders, I'd say we have a lot left to be done to show the
value relative to their perceived needs.
One additional comment- the cost of entry is not what matters to those of us
with battle scars in IT investment. What matters is total cost of ownership,
scale, and sustainability in order to achieve the objective.
- .02, MM
----- Original Message -----
From: "Alan Ruttenberg" <[EMAIL PROTECTED]>
To: "Mark Montgomery" <[EMAIL PROTECTED]>
Cc: "William Bug" <[EMAIL PROTECTED]>;
<public-semweb-lifesci@w3.org>
Sent: Sunday, May 13, 2007 10:24 AM
Subject: Re: Banff demo
On May 12, 2007, at 9:05 AM, Mark Montgomery wrote:
Nice work folks on the slide presentation folks. Haven't been able to
access the demo, but the goal is obvious. I agree that it's a good
foundation to build upon, although it would indeed be nice to have
additional sources to work with. Access to sources is likely to be a
perm challenge.
Well, we've got quite a lot to do in the way of making available the
existing public resources before we are data starved. However, I agree
that specific resources that we would like may be hard to acquire due to
licensing issues. Do you any specific sources in mind?
Of course the communications on value with some will be more
challenging than expressed here- I can almost hear the typical CIO's
cynical reply after years of reduced budgets, 70-80% of which is
required for sustaining legacy-living little if any for innovation, but
R&D divisions I would hope be more receptive.
I would. I think we're at the stage of starting to show the possibility
of doing things that are not available in any single public resource. Not
too long from now I hope we'll be able to do things that haven't ever
been done in easily deployable systems.
In terms of value, I think that the first message is that this technology
is, at least to deploy, very inexpensive. Hardware to run the thing at
this scale is ~3K. All the data is free, and there is an open source
version of Virtuoso. On the other hand, cost of development of new
resources is still high, and we are starved for people who know the
technology well. Hopefully we'll help accelerate that process as we put
more of the documentation about how we converted each resource up on the
wiki, a project for the next few weeks.
I do hope we can more explicitly develop the value proposition to
companies in the future. From my perspective, for the moment the the
challenge is showing value.
Congrats- a good step. - MM
Thanks, and for your comments,
Regards,
Alan
----- Original Message -----
From: William Bug
To: Alan Ruttenberg
Cc: public-semweb-lifesci@w3.org
Sent: Saturday, May 12, 2007 1:51 AM
Subject: Re: Banff demo
Vunderbar!
Thanks to all who worked hard to pull this off!
It's hard to get the full impact just perusing the slides, but it looks
to me that you pulled together a very compelling demo that examined
several questions of biological relevance to neuroscientists studying
neurodegenerative disease (AD in particular).
I also really like the list breakdown of tools to target different
aspects of the overall development - e.g., Pellet, Jena, Perfuse, etc.
I think this will be a fantastic base to build off for ISMB (and SfN).
Kudos!
Cheers,
Bill
On May 11, 2007, at 2:18 PM, Alan Ruttenberg wrote:
I have updated the page http://esw.w3.org/topic/HCLS/Banff2007Demo with
slides, pointers to the triple store etc.
-Alan
Bill Bug
Senior Research 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]