ARHMM: Difference between revisions
(New project page) |
mNo edit summary |
||
| Line 19: | Line 19: | ||
* [[SHMM]]: utilizing gene-pair orientations for improved analysis of ChIP-chip promoter array data | * [[SHMM]]: utilizing gene-pair orientations for improved analysis of ChIP-chip promoter array data | ||
* [[PHHMM]]: improved analysis of Array-CGH data | * [[PHHMM]]: improved analysis of Array-CGH data | ||
* [[MeDIP-HMM]]: HMM-based analysis of DNA methylation profiles | |||
* [https://sites.google.com/site/michaelseiferthmm/hmm-book HMM Book]: Hidden Markov Models with Applications in Computational Biology | * [https://sites.google.com/site/michaelseiferthmm/hmm-book HMM Book]: Hidden Markov Models with Applications in Computational Biology | ||
== Follow Me == | == Follow Me == | ||
* [https://sites.google.com/site/michaelseiferthmm/home Personal Homepage] | * [https://sites.google.com/site/michaelseiferthmm/home Personal Homepage] | ||
Revision as of 18:40, 3 August 2013
by Michael Seifert, Khalil Abou-El-Aradat, Barbara Klink, and Andreas Deutsch
Description
Motivation
Gene expression profiles of tumors are locally biased by deletions and duplications of chromosomal regions. Such local chromosomal dependencies affect expression levels of genes in close chromosomal proximity and should be utilized to improve the analysis of individual tumor expression profiles.
Results
We develop autoregressive higher-order Hidden Markov Models (ARHMMs) that exploit data-dependent local chromosomal dependencies to improve the identification of differentially expressed genes in tumors. ARHMMs combine enhanced local chromosomal modeling capabilities of higher-order state-transitions with direct modeling of chromosomal dependencies between successive gene expression levels using autoregressive emissions. We train ARHMMs using a specifically designed Bayesian Baum-Welch algorithm enabling the integration of prior knowledge on the distribution of differentially expressed genes. We apply ARHMMs to the analysis of breast cancer and brain tumor gene expression data and find that these models substantially improve the identification of differentially expressed genes with underlying copy number changes in comparison to existing methods. This is further supported by the identification of well-known and previously unreported hotspots of differential expression in glioblastoma multiforme highlighting the efficacy of ARHMMs for the analysis of individual tumor expression profiles.
Paper
The paper Exploiting local chromosomal dependencies in the analysis of tumor expression profiles by autoregressive higher-order Hidden Markov Models has been submitted to Bioinformatics.
Download
- ARHMM: A ZIP file containing a JAR file for analyzing tumor gene expression data sets by ARHMMs. This file also contains the breast cancer and the glioma gene expression data sets considered in our study.
Related Projects
- DSHMM: exploiting prior knowledge and gene distances in the analysis of tumor expression profiles
- SHMM: utilizing gene-pair orientations for improved analysis of ChIP-chip promoter array data
- PHHMM: improved analysis of Array-CGH data
- MeDIP-HMM: HMM-based analysis of DNA methylation profiles
- HMM Book: Hidden Markov Models with Applications in Computational Biology
