Centre for Molecular and Biomolecular Informatics Radboud University, Toernooiveld 1, P.O. Box 9010, 6500 GL Nijmegen, +31 (0)24-3653391, postmaster@cmbi.ru.nl

One of the goals of the post-genome era is to figure out how all the proteins that are encoded in the genomes interact with each other to produce protein complexes, pathways and phenotypes. We are developing techniques to predict such interactions from the genomes themselves. Using the "genomic associations" of genes (e.g. the genes are encoded in the same operon) we predict that the encoded proteins are involved in the same process. The goal of this project is to quantify these methods and to combine the various types of genomic association that have been proposed with each other to a single "genomic association" score. Research projects in this area exist either in testing of the quality of the existing methods (comparisong the results with experimental data), in using the existing methods to make new predictions (e.g. for genes that are involved in human diseases), and in combining the various methods to improve the quality of the predictions.

Methodology

The project involves comparing complete genomes with each other to detect homologous and orthologous relations, database mining of experimental "genomics" data, and creative usage of statistical techniques to develop a general "scoring" system for genomic association.

Required pre-training

The student should have a background in genetics and/or statistics, some experience in programming would be of great use.

Keywords: Comparative genome analysis, genome evolution, metabolic pathways, gene regulation, gene expression analysis, yeast 2-hydrid, masspec techniques.

Training period: 6-12 months

Information: Martijn Huynen, huynen@cmbi.kun.nl, tel 024-3653374.