DSHMM

by Michael Seifert, Marc Strickert, Alexander Schliep, and Ivo Grosse

Description

Motivation

Changes in gene expression levels play a central role in tumors. Information about the distribution of gene expression levels and distances between adjacent genes on chromosomes should be integrated into the analysis of tumor expression profiles.

Results

We use a Hidden Markov Model with scaled transition matrices (SHMM) to incorporate chromosomal distances of adjacent genes on chromosomes into the identification of differentially expressed genes in breast cancer. We train the SHMM by integrating information about potential distributions of expression levels of differentially expressed and unchanged genes in tumor. We find that the usage of these information and the modeling of distances between adjacent genes lead to a substantial improvement of the identification of differentially expressed genes. In comparison to existing methods, we find that the SHMM identifies differentially expressed genes with higher accuracy than related methods for analyzing comparative genomic hybridization data. The performance benefit is further supported by the observation that the SHMM predicts genes well-known to be associated with breast cancer. This suggests applications of SHMMs for screening of other tumor expression profiles.

Paper

The paper Exploiting prior knowledge and gene distances in the analysis of tumor expression profiles by extended Hidden Markov Models has been submitted to Bioinformatics.

Download

• SHMMs will be available soon in Jstacs.
• The data sets will be available soon.