Bioinformatics at
the Department of Informatics, University of Bergen

New page is found at http://www.ii.uib.no/forskningsgrupper/bio/

• What is bioinformatics?
• History of the group
• Teaching activities
• Research activities
• Selected publications
• Group members
• Externally funded projects
• Hovedfagsoppgaver (Thesis topics, M.S., in Norwegian)
• Bioinformatics links

Bioinformatics is the development and application
of methods in informatics for approaching problems in
molecular biology.

An article about bioinformatics in a local students paper is here (Norwegian text).

History of the group

The group also collaborates closely with researchers at

• the European Bioinformatics Institute in Cambridge,
• the Computer Science department at University of Helsinki, Finland,
• the Biocomputing Research Unit at University of Edinburgh,
• the Computer Science department at City University in London, and
• the National Institute of Medical Research in London.

Teaching Activities

The aim of I282 is to give the students insight into problems from molecular biology in which methods from informatics can be helpful. No prior knowledge in biology is required, but a solid informatics background including an advanced course in algorithms, is required. We emphasise the theoretical and biological foundation of the problems and methods, and in particular we focus on methods for assessing the statistical and biological signficance of the results produced by computational methods.

In collaboration with the Department of Molecular Biology, we are planning to make a study plan in bioinformatics (where each student may choose to put most emphasis on informatics or on molecular biology) which will include an introductory level bioinformatics course as well as more advanced graduate courses and seminars.

Seminars in bioinformatics:

• Seminar in Bioinformatics spring 97
• I289 Seminar in Bioinformatics spring 98

The focus of the research in the bioinformatics group is on methods for the analysis of biological sequences and structures and for structure prediction. Research has also been done in the area of approximate pattern matching and database design. Below we describe the most important areas in some more detail:

• Sequence Motif Discovery:
  We are working with search based methods for identifying patterns conserved in sets of related, unaligned protein or nucleotide sequences. Conserved patterns are useful both for classification purposes (e.g., PROSITE) and for understanding relations between sequence, structure and function of the proteins. We have developed a new tool Pratt for the identification of conserved patterns. We continue our work in developing new methods for finding conserved patterns that are difficult to find using existing methods.
• Protein Structure Prediction:
  We are developing new methods for the ab-initio prediction of the structures of proteins from their sequences. The work focuses on small zinc-containing proteins (zinc fingers) and exploits the extra constraints the structures need to satisfy to be able to bind zinc. So far, the developed methods have used genetic algorithms to search the space of possible structures.
• Structure Motif Discovery:
  Based on our experience in methods for sequence motif discovery, we develop new methods for the discovery of motifs in protein strucutures. A method named Spratt has been developed that is able to analyse one or several protein structures to find recurrring patterns (motifs) of local packing of residues.
• Constraint Based Languages for the Description of Biological Structures:
  In collaboration with City University, London, we have developed a new constraint based language for the description of structures in biological sequences. For instance, in this language we can describe repeats, stem loops, and pseudo-knots in RNA sequences. A prototype of a CSP solution to matching descriptions and sequences has been developed.
• Multiple Sequence Alignment:
  This work aims at developing better methods for multiple sequence alignment for discovering and visualising similarity between several sequences. Methods using parallel processing have been developed.

• PALM - a Pattern Language for Molecular Biology
  Carsten Helgesen and Peter R. Sibbald
  in Proceedings of the First International Conference on Intellignent Systems for Molecular Biology (ISMB-93), AAAI Press 1993, p 172-180.
• Methods for finding motifs in sets of related biosequences
  Alvis Brazma, Inge Jonassen, Ingvar Eidhammer, David Gilbert.
  Journal of Computational Biology 5, 279-305 (1998).
• Finding flexible patterns in unaligned protein sequences.
  Inge Jonassen, John F. Collins, Desmond Higgins.
  Protein Science 4, 1587-1595 (1995).
• Efficient discovery of conserved patterns using a pattern graph.

Inge Jonassen.
CABIOS 13, 509-522 (1997)
• Discovering patterns and subfamilies in biosequences

Alvis Brazma, Inge Jonassen, Esko Ukkonen, Jaak Vilo
in Proceedings of the Fourth International Conference on Intellignent Systems for Molecular Biology (ISMB-96), AAAI Press 1996, p 34-43.
• A Constraint Based Structure Description Language for Biosequence

Ingvar Eidhammer, David Gilbert, Inge Jonassen, Madu Ratnayake
Dept. of Informatics, Univ. of Bergen, Reports in Informatics no 133, May 1997,
• Structure Prediction of the Zinc-finger protein family by the use of a parallel genetic algorithm
  Kjell Petersen, Ingvar Eidhammer, Rein Aasland, Inge Jonassen
  Proceedings of NIK97 (Norwegian Informatics Conference), p 57-68, Tapir 1997.
• Discovering patterns conserved in sets of unaligned protein sequences
  Inge Jonassen
  Book chapter in "Predicting Protein Structure" edited by D.M.Webster, In the series "Methods in Molecular Biology" from Humana Press, In press (1998).
• Discovery of Local Packing Motifs in Protein Structures
  Inge Jonassen, Ingvar Eidhammer, William R. Taylor
  Proteins: Structure, Function, and Genetics 34, 206-219.
• Prediction of Regulatory Elements in Silico on a Genomic Scale
  Alvis Brazma, Inge Jonassen, Jaak Vilo, Esko Ukkonen
  Genome Research 8, 1202-1215 (1998)