Gibbs sampling for microarray data analysis
1. Introduction 1.1 Goal of the study
Draw backs of conventional clustering algorithms The need of integration of different data sources
Biclustering algorithm 1.2 Achievements
2. Microarray 2.1 The technology
2.2 Characteristics of microarray data
2.3 An important tool for bioinformatics, but can’t explain everything – the need to integrate the other sources
3. Clustering
3.1 First generation algorithms 3.2 Second generation algorithms 3.3 Biclustering algorithms
2.4 Gibbs sampling: the basic idea, the introduction of Gibbs sampling method to bioinformatics (i.e. its application to motif-finding problems), why Gibbs sampling for the biclustering problem
4. Bayesian models for microarray data analysis and Gibbs sampling 4.1 Basic ideas of Bayesian methods: posterior likelihood * prior 4.2 Applying Bayesian models on microarray data
4.2.1 Advantages
4.2.2 Which distributions to choose: taking into the account of the noisy characteristics of microarray data
4.2.3 The preprocessing of microarray data: (in order to apply the specified models) 4.3 Gibbs sampling
5. Biclustering for patient data 5.1 The discrete model
5.2 The Gibbs sampling strategy 5.3 Query-driven biclusteing
Construction of the priors:
5.4 Influence of the parameters
5.5 Case study
6. Biclustering of genes 6.1 The continuous model 6.2 Model for time-sequence data 6.3 The Gibbs sampling strategy 6.4 Query-driven biclustering
Construction of the prior 6.5 Influence of the parameter 6.6 Case study
7. Conclusions
7.1 Comparison of the Gibbs biclustering strategy to other methods:
Comparison between Bayesian and non-Bayesian methods Parameterization: Gibbs sampling vs. EM
8. Future work: