5. Promoter prediction II: reliability EMBL - UPF course 2001
   
 

Problems:

Too much false positives (specifity) due to:

  • Binding sites patterns are very short (5-15 bp)
  • Usage of TF databases: bias (new patterns)
  • Relationships between TFs are complex and degenerated
There is room for the improvement...

  • Genome-wide expression data from microarrays
  • Phylogenetic information from homologous genes
  • New research about epigenetic information, CpG islands, DNA bendability, modules (cooperative sites)

L A S T H O M E N E X T

Enrique Blanco, Sergi Castellano and Genis Parra © 2001