28-29 September 2009: Workshop at Roskilde University

The workshop takes place in room 42.2-02 (bldg. 42, level 2, room 02: ISL/PLIS-lab); participation only.

Preliminary schedule

The following time table is flexible, which means that if we start an interesting discussion during one presentation, we may postpone the next sessions, even until the following day. And we can decide to swop presentation.

Only lunch breaks and the workshop dinner are fixed.

Monday, September 28

9:00
Welcome, Presentation

9:10-9:40
Overview of the LoSt project
- what have been achieved so far, goals and activities for the remaining project period
Henning Christiansen

9:40-10:20
Searching for hard-to-find genes, or: What we may (hope) to do better than the others
Ole Skovgaard

10:20-10:30 Short break

10:30-11:30
Bayesian and non-Bayesian parameter estimation for PRISM programs
James Cussens

11:30-11:50 Break

11:50-12:30
Layered model of analysis
Ole Torp Lassen

12:30 Lunch

13:30-14:10
Bayesian coordinating HMMs for sequence analysis
Henning Christiansen

14:10-14:50
Probabilistic Logic Models of mRNA sequence and secondary structure
Søren Mørk

14:50-15:00 Short break

15:00-15:40
HMM approximation of CFGs
Christian Theil Have (joint work with Ole Torp Lassen)

15:40-16:10 Break

16:10-16:50
Constrained HMM for Easygene
Matthieu Petit

16:50-17:30 (Perhaps re-schedule one of the two talks scheduled for Tuesday; below)

17:30-18:30 (at latest)
Discussion: open problems, where to put our attention
Conclude: Topics for detailed work in the rest of the workshop, what, who, when.

18:30
Departure for the Workshop Dinner

Tuesday, September 19

9:00-9:50
Manfred Jaeger
Relational information gain - with applications to bio-databases

9:50-10:00 Short break

10:00 Starting detailed work in (perhaps) smaller groups

11:15-11:30 break

11:30-12:30 More detailed work 
10:30 Short break and detailed work, until

12:30-13:30 Lunch

13:30-14:20
EM clustering with GO scoring
Nicos Angelopoulos

14:20-14:40 Break

14:40-16:00 Detailed work

16:00-17:00 Reports from the working groups, suggestions for future collaborations

Host

prof. Henning Christiansen, henning@ruc.dk, PLIS research group, CBIT Institute, at Roskilde University and the LoSt project,

For directions on how to get to Roskilde University, go here and also here

4 March 2009: Mini Workshop at Roskilde University

Announcement

Dear LoSt participants, external partners and other interested parties,First of all, welcome to Christian Theil Have as new PhD student in the project, starting officially March 1.

On Wednesday, March 4, next week we will organize a mini workshop starting 15:30 in the afternoon with different contributions in probabilistic-logic modelling. Workshop program; (timing include discussions, changing computers, etc.)

Program

15:30	Welcome and overview of the LoSt project Henning Christiansen
15:35	A focus on Overlapping Genes: The problem and current investigations Ana Capatana
	Recent progress in the LoSt project
15:55	Optimization by decomposition: Chopping and evaluation Ole Torp Lassen
16:15	PRISM programs with annotation: Optimzation by program analysis and transformation Henning Christiansen
16:35	Coffee, tea, fruit, ...
17:00	Coordinating HMMs and other sorts of Constrained HMMs Matthieu Petit
	In depth with constraints and failure
17:20	The implementation of an EM algorithm for Stochastic Logic Programs MCMC over distributional logic programs clustering via EM with mclust package of R Nicos Angelopoulos
17:50	Stochastic Definite Clause Grammars Christian Theil Have
18:10	Final discussion: Most pressing problems and possible new directions
18:40	The end

Host

prof. Henning Christiansen, henning@ruc.dk, PLIS research group, CBIT Institute, at Roskilde University and the LoSt project,

For directions on how to get to Roskilde University, go here and also here

5 May 2008: Research seminar at Roskilde University

Announcement

A methodology for Bayesian machine learning and its application to ligand discovery

by Nicos Angelopoulos, Edinburgh University

Roskilde University, room 42-2.37,
Monday May 5, 2008, 12.15-13.00

Abstract

The talk describes a probabilistic logic programming language in the context of Bayesian knowledge representation. An MCMC algorithm operating on model structures defined in the language and its implementation are also described. The system has recently been used to provide a Bayesian analysis of high-dimensional descriptor data for the pyruvate kinase protein.

Classification trees have been learnt from experimentally determined binding affinity data. The algorithm has performed better than two Neural Network approaches. In addition, it can be more easily run on a distributed environment and lends itself naturally to classifying sets with missing data. The approach has been used to select and rank potential biologically active compounds and could provide a powerful tool in compound testing.

Host

prof. Henning Christiansen, henning@ruc.dk, PLIS research group, CBIT Institute, at Roskilde University and the LoSt project,

For directions on how to get to Roskilde University, go here and also here

5-6 February 2008: Workshop at Roskilde University

Announcement

This is first internal workshop in LoSt project; LoSt is the acronym for the project "Logic-statistic analysis of biological sequence data".

Place: Roskilde University, CBIT Institute, Rooms 42-2.37 and 42-1.37

Participation: By invitation only.

Workshop program (preliminary)

	Tuesday 5 Feb 2008 (Room 42-2.37)
10:00-10:30	Henning Christiansen, Roskilde University: Presentation of project and workshop participants Overall project goals with an emphasis on computational issues
10:30-11:10	Ole Torp Lassen, Roskilde University: Combining HMM and SCFG in PRISM, and approaching complexity problems by preprocessing with an additional HMM.
11:20-12:00	Bjarne Knudsen, Peter Johannes Steffensen, CLC bio: How a producer of competitive bioinformatics software may contribute to and benefit from the LoSt project. Integration of PRISM-like methods and CLC bio's software.
12:00-13:00	Lunch (the foyer of bldg. 0)
13:00-14:00	Taisuke Sato, Japan Institute of Technology: Background and intentions of PRISM, history and development, applications; perspectives for future developments and applications.
14:15-15:00	Yoshitaka Kameya, Japan Institute of Technology: Algorithms and data structures in PRISM with emphasis on complexity and bottlenecks for complex problems.
15:00-18:00	Coffee, General discussion Working groups and PRISM hands-on
18:00 approx.	Departure by taxi to ...
18:30-	Workshop dinner Restaurant Mumm, Roskilde

	Wednesday 6 Feb 2008 (Room 42-1.37)
9:00-9:45	Ole Skovgaard, Roskilde University: On Chr. Hansen's role in the project Introduction to gene finding
9:50-10:25	Søren Mørk, Roskilde University: Overview of PhD project
10:30-11:15	John Gallagher, Roskilde University: Applications of program analysis and transformation for optimization of logic-statistic programs.
11:30-12:15	Manfred Jaeger, Aalborg University: An overview of Logic-Statistic methods and systems
12:15-13:15	Lunch (the foyer of bldg. 0)
13:15-13:30	Bjarne Knudsen, CLC bio: A smart little box containing the Smith-Waterman algorithm
13:30-15:00	Working groups (?)
15:10-15:25	Taisuke Sato, Japan Institute of Technology: Summary of the workshop and recommendation for the project
15:25-16:00	Panel discussion
	Concluding remarks