|
Regulatory
and functional RNAs: computational, genomic, and structural
approaches
Report
Scientific
content of the event
As it was expected, in this meeting we discussed probably
all the mayor research
fronts in the field of computational approaches for RNA. The
structure of
the meeting included two talks a day (open ending, with most
of them extending
for almost 2 hours) plus additional more informal seminars
that were included
at the participants request. There was also a large amount
of person–to–person
interactions which were encouraged by the relaxed atmosphere,
the relatively
large amount of free time between the morning and the evening
sessions, and
the good facilities that the Benasque center provided.
All
talks characterized for being very participative, with great
in–depth
mathematical discussions when it was so requested by the participants.
I particularly
remember two of them, one in which people working on Mfold
and the
ViennaRNA package realized that their implementation on how
to calculate the
partition function although assumed to be the same was in
fact di erent. (The
partition function is the sum for a given RNA sequence of
all possible structures
which is necessary to make more robust estimation of RNA folding.)
The
other one was a very detail discussion of all the di erent
approaches to make
the Sanko algorithm (how to fold and align RNA sequences simultaneously)
work in a time–and–memory efficient way. I think
that amongst the participants
we had probably all the players who are currently working
on the Sanko
algorithm, which is a very important issue in order to make
RNA folding more
reliable based on alignment information.
There
were also many discussions on how to model the evolutionary
processes
observed in RNA molecules. Traditionally, the evolutionary
models used in
probabilistic phylogeny consider the evolution of one residue
(a column in a
multiple alignment) at the time. This is clearly insufficient
to describe RNA
evolution. In the workshop, we discussed methods able to consider
the join
evolution of pairs of columns subject to base pairing. Relevant
as base-pairing
is for RNA secondary structure, it was also pointed out at
the workshop that this
pair–evolution approach is still clearly insufficient.
The reality of the evolution
of three–dimensional architectures shows that whole structural
domains can
appear or disappear throughout the course of evolution without
a ecting the
functionality of the RNA molecule. Models to describe more
realistic RNA
evolutionary processes still remain as an important open problem.
In
summary the topics we discussed during the meeting included:
Formalisms
for RNA detection (stochastic context free grammars), improvements
on RNA 2D structure prediction, such as improvements on the
thermodynamic
parameters, use of probabilistic based parameters, posterior
information on the
more reliable base pairs, use of alignment information (the
Sanko algorithm
and its many variants). RNA 3D structure and the identification
of 3D motifs.
How to incorporate evolution in RNA models, and its applications.
RNA
genefinding, and in particular genefinders for snoRNAs. How
to model the
kinetics of RNA folding. We also discussed bioinformatic topics
such as the
latest developments on databases for RNA sequences and RNA
motifs, and the
convenience or not of a common XML–based language for
RNA display.
Next in this report we include a list of the mayor talks that
took place during
the workshop. We concluded the workshop with a last session
of open questions,
a summary of which (collected by George Hartzell) is also
included in this report.
2. Programme
Monday,
July 28
9:00 h, General Meeting
11:30 h, Sean EDDY Stochastic context-free grammars
18:30 h, Organizational Meeting
Tuesday, July 29
11:30 h, Francois MAJOR, Daniel GAUTHERET Structural Motif
Discovery
18:30 h, Francois MAJOR, Michael ZUKER, Fabrice JOSSINET RNAML
Wednesday, July 30
11:30 h, Robert GIEGERICH Comparison of RNA 2D structures
18:30 h, Steven BRENNER RNA processing, splicing, degradation
and regulation
Thursday, July 31
9:00 h, Eric WESTHOF Tutorial on 3D structure prediction
11:30h, Charles LAWRENCE, Ye DING How good is a sample of
size one, even
if it is optimal?
18:30h, Michael ZUKER ”Pseudo free energies” for
base pair stacking and 2x2
loops.
Friday, August 1
11:30h, David MATHEWS, Jan GORODKIN Finding a secondary structure
common to multiple sequences
17:00h, Workshop on applications of Sanko algorithm
18:30h, Ivo HOFACKER, Gerhard STEGER, Elena RIVAS Predicting
consen-
sus structure from sequence alignment
Monday, August 4
11:30h, Mark YANDELL Genome annotation
16:00h, Robert GIEGERICH The algebraic dynamic programming
(in)formalism
18:30h, Todd LOWE, Christine GASPIN, Peter SCHATTNER Gene
prediction
Tuesday, August 5
9:30 h, Eric WESTHOF Tutorial on 3D structure prediction (Cont.)
11:30 h, Steven BRENNER, Sam GRIFFITHS-JONES RNA Databases
(SCOP
and Rfam)
16:00 h, Peter CLOTE Time warping and functional genomics
Wednesday, August 6
9:30h, David MATHEWS Thermodynamic parameters for RNA folding
11:30 h, Peter STADLER Kinetics of RNA folding
18:30 h, Ian HOLMES, Evolutionary SCFGs
Thursday, August 7
11:30 h, Ian HOLMES, Bjarne KNUDSEN Evolution of RNA
18:30 h, Open problems
Future
directions and open problems
Feature requests (enhancements, small changes, things w/out
much
”research” mileage):
Better packaging, documentation, and shareable code
Better outreach to biologists/collaborators (Marketing/outreach)
Hardware specific implementations of core algorithms/applications
|
