|
- Third International Conference of the
Canadian Proteomics Initiative -
May 23rd to 27th, 2003 - University of British Columbia, Vancouver, BC
by Chris
Dambrowitz, Genome Prairie (dambro@genomeprairie.ca)
This
meeting was sponsored by PENCE (the Canadian Protein Engineering National Centre
of Excellence), Genome BC, and several other Canadian research organizations.
A dominant subtext of the event involved the future of proteomics research in
Canada - and what initiatives might be responsible for shaping that future.
This report seeks to provide a view of the themes and undercurrents of the
conference, rather than an all-inclusive review of the scientific presentations.
For readers interested in the latter, the abstracts posted at
http://www.cpi03.org provide such a reference.
In her
opening address, Martha Piper, President of the University of British Columbia (UBC),
announced that UBC would be adding ten new faculty positions in proteomics and
bioinformatics. These researchers will be part of the Biotechnology
Laboratory, based in the 7,400-square metre Michael Smith Building currently
under construction and slated for completion in late 2004.
The
first scientific presentations were by researchers from Seattle’s Institute for
Systems Biology. Ruedi Aebersold discussed novel applications of ICAT
technology for quantitative proteome analysis - particularly focusing on
comparative and quantitative analysis of protein complexes. John Aitchison
provided an overview of the complexity and data-richness involved in a
systems-biology approach to understanding the dynamics of a single organelle,
the peroxisome. Both of these ISB-based presenters have strong research
roots and current collaborations in Canada. Not only did the scope and success
of their presented research set the bar for subsequent speakers at this meeting,
but also highlighted the challenge before Canadian proteomics researchers as the
future is contemplated. Proteomics research on a grand scale is
demonstrably feasible in Canada - do we also have the scientific vision,
operational clarity, and political will to capitalize on that potential?
Day 2
of CPI ’03 began with Fiona Brinkman’s (Simon Fraser University) description of
the BAE-watch database for potentially identifying patterns in the evolution of
microbial virulence - and an outline of strategies for microbial control derived
from Sun Tzu’s Art of War. Mitsu Ikura’s (University of Toronto)
discussion of calcium-induced conformational changes in calmodulin and related
protein domains highlighted the perspective that large-scale proteomic methods
may still require support from targeted studies on each individual protein
class, in a manner that may not be amenable to high-throughput techniques.
The tension between the classical protein biochemistry “one-protein-one lab-one
lifetime” approach, and the massively-parallel paradigm of proteomics again
became a predominant theme of the day, as echoed through presentations, audience
questions, and coffee-break discussions.
Presentations that
focused on structural proteomics followed, by Natalie Strynadka (UBC) regarding
high-throughput structural studies of membrane proteins by X-ray
crystallography, and by Lewis Kay (U of T) on multi-dimensional NMR analysis of
high-MW proteins and multi-protein complexes. The first of these two talks
particularly emphasized that the answer to a case-by-case approach to resolving
protein structures lies in the judicious selection of experimental algorithms
for batch processing that permit high-throughput structural analysis.
McGill’s John Bergeron presented a view of single-organelle proteomic analysis
efforts by LC/MS/MS, with emphasis on the study of the endoplasmic reticulum.
A practical note, highlighting the bottleneck presented by available
bioinformatic techniques, was his estimate that 5% of lab time is spent actually
performing wet-lab work (such as organelle isolation, sample prep, etc.), while
the other 95% is dedicated to matching his group’s MS/MS data to protein
databases. Tony Pawson (SLRI/U of T), in his discussion of approaches to
assembling multiprotein regulatory complexes from analysis of protein domain
interactions, drew chuckles by commenting that his emphasis might be described
as “one-DOMAIN-one-lab-one-lifetime”. Bob Olafson described the evolution
of the UVictoria-Genome BC Proteomics Centre from a protein-chemistry core
facility in the early ‘80s to its present incarnation as a high-throughput
facility providing LC/MS/MS and ICAT services (among others) to academic,
industrial and governmental research groups. During the Q&A period,
vigorous discussion focused on the accessibility of such services to Canadian
academic groups, and also on the problem of obtaining support for large-scale
“discovery science” proposals within the confines of the traditional Canadian
research funding paradigm.
On the
heels of this discussion, U of Toronto’s Andrew Emili brought into the open key
issues that had been hinted at for much of the conference. In presenting his
group’s approach to large-scale analysis of proteome dynamics during myoblast
differentiation, Emili’s key theme was that algorithms, not MS methodologies,
identify proteins. Automation and ultrahigh throughput techniques are absolutely
essential to the success of discovery-science projects. In a succinct
summary of one of the meeting’s key undercurrents (and a contentious one, based
on the resultant rise in tension in the room), Emili asserted that what
proteomics research is really all about is “Go Big, or Go Home”.
Proteometricians should leave (traditional) biology to the biologists - and
should instead focus on proteomics, providing a holistic view of biology at the
systems level. The challenges that this paradigm shift presents to the
Canadian research community, and the divisions within that community on this
issue, were evident at this meeting.
One of
the most anticipated sessions of CPI ’03 was a forum on the Shape of Proteomics
in Canada. The session’s organizing committee had selected several
speakers to provide context for discussion. David Griller gave an
excellent presentation of
SECOR survey results indicating that Canadian proteomics must support bold
initiatives having strong socioeconomic impact, and maintain a global
orientation in commercialization in order to capture economic benefits for
Canada. Subsequent presenters, including Aled Edwards (U of T), John Bergeron
(under the HUPO banner), Francis Ouellette (UBC), Michel Chretien (U Ottawa),
Roger Foxall (Genome BC), Steve Armstrong (Genome Atlantic), and Peter Freeman (ANPI),
described various regional or strategic initiatives relating to
proteomics-related networks. Finally, a plan for a Canadian Proteomics
Network was outlined that might contribute to a national infrastructure for
research in proteomics and protein engineering. A steering committee,
potentially composed of PENCE’s Steve Withers, Joel Weiner, Peter Lewis, John
Bergeron and the NRC’s Bob Boyd, would take responsibility for determining the
mandate, objectives, strategy and budget of such a national network.
Unfortunately, time constraints prevented an assessment of grassroots response
to the outlined initiative, leaving issues raised through this session open for
future discussion.
The
final day of speaker presentations (leaving two days for tutorial sessions on
various topics) was similar in tone to the first - several excellent talks that
highlight the scope attainable through proteomic studies. Mike Moran of
MDS Proteomics reviewed MDS.P’s technology for measuring changes in protein
expression, organellar proteome, protein-protein interactions, and phospho-protein
profiles in response to disease and to drug treatments, including MDS.P’s new
software and hybrid FT-ICR mass spectrometry instrumentation. NIST’s John Moult
reviewed his group’s study of the relationship between mild destabilizing
effects on canonical protein structures caused by single nucleotide variations
and disease, shedding light on the system-engineering processes that underlie
the robustness of protein networks. Finally, Chris Dobson (Cambridge
University) demonstrated the effective combination of experimental data and
simulations to identify surface features related to protein folding and mis-folding,
and the relationship between protein folding, self-assembly, and structural
evolution. (Unfortunately, this reviewer missed the opportunity to attend
the latter two talks, and so is relying on secondhand information for their
description.)
The
next Canadian Proteomics Initiative Conference is currently slated for May 2004,
in Montreal QC.
|