Bioinformatics Faculty Appointment Form - University of Memphis

websterhissBiotechnology

Oct 1, 2013 (3 years and 8 months ago)

64 views





Title
:

Analysis of Gene Expression Data using Hierarchical Bayesian

and Gene Set Enrichment Procedures

Haibao Wan

University of Memphis

Date:

Friday, November

9
, 2007

Time:

1:30pm
,

Place:

Rm
351

Dunn Hall



Abstract:



A Bayesian hierarchical model

is used to detect differentially expressed genes between two
experimental conditions. The modeling is an integrated statistical approach which simultaneously
models intrinsic biological variability, systematic array effects and differential expression. A
windows
version of the software Bayesian inference Using Gibbs Sampling (WinBUGS) that performs Markov
Chain Monte Carlo simulations, is used to implement the model and obtain the samples of the model
parameters from their posterior distributions. Differen
tially expressed genes are selected by a rule,
given in Lewin et al. (Lewin et al., 2006), and false discovery rates are computed and used to control
the false positives. A subset of the differentially expressed genes is used in the ingenuity pathway
analy
sis to discover, visualize and explore the biological networks.

The identified networks of the biological pathways are then analyzed. As an alternative and
confirmatory analysis to the Bayesian model, a gene enrichment analysis (GSEA) procedure is
implem
ented. Gene set enrichment analysis (GSEA) is a newly developed strategy for analyzing
microarray data as gene sets instead of individual genes. Gene sets are defined based on prior
biological knowledge such as biochemical pathway. Compared to the single
-
g
ene methods, GSEA has
the advantage of easy interpretation for biological meaningfulness and can detect modest changes in
individual genes.