Johannes Hüsing1, Michael Zeschnigk, Tanja Boes, Karl-Heinz Jöckel. 1. Institut für Medizinische Informatik, Biometrie und Epidemiologie, Universitätsklinikum Essen, Hufelandstrasse 55, D-45122 Essen, Germany. johannes.huesing@medizin.uni-essen.de
Abstract
MOTIVATION: Loss of chromosomal material is often observed in cancer cells. In this situation the expression of genes is related to their position on the genome. Epigenetic phenomena may also silence several genes in the same region of a chromosome. While cytogenetic or other molecular genetic methods spot changes of DNA copy number, they cannot detect other causes of silencing. RESULTS: We propose a method that utilizes the link from expression information gained from high-density DNA microarrays to the gene locus according to current databases. Statistical methods adequate to spot conspicuous runs of non-expressed genes are introduced and compared to one another by merit of their power and robustness against false positives. AVAILABILITY: Code for the formulae can be obtained (R code) via http://www.panix.com/~derwisch/hannes/longrun
MOTIVATION: Loss of chromosomal material is often observed in cancer cells. In this situation the expression of genes is related to their position on the genome. Epigenetic phenomena may also silence several genes in the same region of a chromosome. While cytogenetic or other molecular genetic methods spot changes of DNA copy number, they cannot detect other causes of silencing. RESULTS: We propose a method that utilizes the link from expression information gained from high-density DNA microarrays to the gene locus according to current databases. Statistical methods adequate to spot conspicuous runs of non-expressed genes are introduced and compared to one another by merit of their power and robustness against false positives. AVAILABILITY: Code for the formulae can be obtained (R code) via http://www.panix.com/~derwisch/hannes/longrun