Jacques van Helden Jacques'van'Heldenulb'ac'be

1
Correspondence Analysis

• Statistics Applied to Bioinformatics

2
A multitude of biological data

• With the advent of high-throughput methods, we
  dispose of several genome-scale datasets,
  measuring different properties of genes and
  proteins
• Hundreds of genomes.
• DNA microarrays to measure transcriptome.
• Mass spectrometry to detect protein complexes.
• Two-hybrid to detect protein-protein
  interactions.
• ChIP (Chromatin Immunoprecipitation) on chip to
  detect protein-DNA binding.
• Mutant phenotypes.
• Genotypic variations (SNP).
• Catalogs of annotations
• Regulons (TRANSFAC, RegulonDB, FootPrints,
  Jaspar)
• Metabolic pathways (MetaCyc, KEGG, aMAZE)
• Processes, cellular locations and lolecular
  functions (Gene Ontology project).
• How can we combine the pieces of information
  coming from these different sources ?

3
Study case 1 - Gasch (2000) expression data
4
Study case 2 - Harbison (2004) ChIP on chip
5
Correlation analysis
6
Associations
7
Correspondence analysis