Bioinformatics, Genetics and Stochastic Computation: Bridging the Gap (07w5079)

Arriving in Banff, Alberta Sunday, July 1 and departing Friday July 6, 2007


Arnaud Doucet (University of British Columbia)

(Fred Hutchinson Cancer Research Center)

(Universite Paris Dauphine)


On the one hand, there is an explosion of complex statistical models appearing
in genetics and bioinformatics. These typically highly structured
systems are unfortunately very difficult to fit. On the other hand, there
has been recently significant advances on MC methods but most of these
methods are unknown to the applied community and/or formulated in ways
that are too theoretical for direct application. By providing scientists with
improved inferential methods it would allow them to consider richer models,
which are more realistic than those dictated by computational constraints.
The exchanges between the applied and methodological communities remain
surprisingly limited.

We believe that this workshop would be an ideal place to
(1) Gather people from different research communities and foster links
between these communities (applied, methodological and theoretical).
(2) Expose the applied community to novel statistical methodologies and
advanced MC methods, and expose the MCMC community to the specifics
of the complex modeling problems met in bioinformatics and genetics.
(3) Classify topologies of computational problems met bioinformatics and
genetics, and equip all participants to the workshop with benchmark problems,
if possible before the workshop.

We anticipate this workshop to create an exceptional opportunity for
exchanging ideas between the communities; and help to shape the future
of stochastic computation within bioinformatics and genetics. Input from
statisticians working in bioinformatics and genetics is absolutely crucial for
development of appropriate statistical methodologies. We will encourage
researchers to bring data from their own field, which could be used to implement
methods and try new algorithms. We have targeted areas of stochastic
computation that are of great interest to practitioners such as automatic
algorithms, computational issues and parallel implementations. Bioinformatics
and genetics are relatively new fields that enjoy a high representation
of young talents, including many women. This workshop would be a great
learning/training environment for these new talents.