Difference between revisions of "Module 4: How does feature selection impact integrative clustering analysis?"
 
Line 1: Line 1:
−
*  Title: "How to make valid prognostic models with gene expression signatures?"
+
*  Title: "How do "our expectations" confound the results of our analyses?"
  
−
* Paper to be examined: “The genomic and transcriptomic architecture of 2,000 breast tumours reveals novel subgroups”, Nature 486(7403):346-52 (2012)[http://www.nature.com/nature/journal/v486/n7403/full/nature10983.html]  
+
* Paper to be examined: “Mutual Exclusivity analysis identifies oncogenic network modules”, Genome Research 2012 22, 398-406 [http://genome.cshlp.org/content/early/2011/10/12/gr.125567.111.full.pdf]
 +
+ cancer genomic studies from The Cancer Genome Atlas (TCGA).
  
−
* Key claim of the paper: "We have generated a robust, population-based molecular subgrouping of breast cancer based on multiple genomic views. [...] The joint clustering of CNAs and gene expression profiles further resolves the considerable heterogeneity of the expression-only subgroups."  
+
* Key claim of the paper: "We introduce here a simple but effective method to evaluate the statistical significance of correlations between genomic events, that concurrently preserves both '''tumor selectivity''' and '''tumor heterogeneity'''"  
  
 
* Data and Code
 
* Data and Code
 
    
 
    
 
* Schedule:  
 
* Schedule:  
−
** H1: General introduction to the paper/motivation
+
** H1: General introduction to the cancer genomics concepts and motivation introduced in the paper
−
** H2: Write code to import the data and practice with the iClusterPlus R package with vignette example
+
** H2: The concept of random expectation (the null model)
−
** H3: Reproduce results from Figure 4 on subsample(s) of the data
+
** H3: How should we model cancer heterogeneity to estimate the significance of mutual exclusivity?
−
** H4-5: Write code to import second dataset and reproduce clustering results
+
** H4-5: Use R to load and analyze the first cancer genomics dataset (compare alteration distributions and permutation models)
−
** H6: Discussion: "What features discriminate the resulting clusters? Do we see the issue? How can we improve?"
+
** H6: Reproduce mutual exclusivity analyses proposed in the original paper.
−
** H7-8: Based on discussion, modify feature selection and redo the analyses on one (two) datasets
+
** H7-8: Examine TCGA cancer genomic studies to perform mutual exclusivity analyses under different permutation models
−
** H9: Summarize results (e.g. on this wiki)
+
** H9: Summarize and present the results.
  
 
* Key bioinformatics concept of this module:  
 
* Key bioinformatics concept of this module:  
−
** Feature selection (and its importance for cluster analyses)
+
** The importance of the null model (your expectations)
−
** integrative analysis
+
** Properly assessed Mutual Exclusivity inform on biological pathways altered in cancer
  
 
* back to [[UNIL MSc course: "Case studies in bioinformatics 2015"]]
 
* back to [[UNIL MSc course: "Case studies in bioinformatics 2015"]]

Latest revision as of 19:44, 23 November 2015

  • Title: "How do "our expectations" confound the results of our analyses?"
  • Paper to be examined: “Mutual Exclusivity analysis identifies oncogenic network modules”, Genome Research 2012 22, 398-406 [1]

+ cancer genomic studies from The Cancer Genome Atlas (TCGA).

  • Key claim of the paper: "We introduce here a simple but effective method to evaluate the statistical significance of correlations between genomic events, that concurrently preserves both tumor selectivity and tumor heterogeneity"
  • Data and Code
  • Schedule:
    • H1: General introduction to the cancer genomics concepts and motivation introduced in the paper
    • H2: The concept of random expectation (the null model)
    • H3: How should we model cancer heterogeneity to estimate the significance of mutual exclusivity?
    • H4-5: Use R to load and analyze the first cancer genomics dataset (compare alteration distributions and permutation models)
    • H6: Reproduce mutual exclusivity analyses proposed in the original paper.
    • H7-8: Examine TCGA cancer genomic studies to perform mutual exclusivity analyses under different permutation models
    • H9: Summarize and present the results.
  • Key bioinformatics concept of this module:
    • The importance of the null model (your expectations)
    • Properly assessed Mutual Exclusivity inform on biological pathways altered in cancer
Retrieved from "http://www2.unil.ch/cbg/index.php?title=Module_4:_How_does_feature_selection_impact_integrative_clustering_analysis%3F&oldid=4583"