Get-Together for String Algorithms Researchers

Tapahtuman tyyppi: 
13.10.2011 - 14:15 - 16:00
Exactum B120

Jarkko Toivonen will give a talk:

Title: Finding over-represented motifs and motif pairs in biological sequence data

Abstract: In gene regulation it is important to be able to describe the strength of binding between a transcription factor (TF, a protein) and a binding site (short piece of DNA). Our aim is to learn a model for this interaction based on a large dataset of short dna sequences that contain a binding site of a certain TF with high probability. There has been numerous attempts at learning models for TF binding sites but still the binding sites have been described only for a fraction of all human factors. The difficulty may be caused by the fact that transcription factors seldom act alone but rather as complexes of several transcription factors. We propose a model that describes both individual binding sites of a single TF and the change in binding strength caused by the presense of other factors. The model is learned from the over-represented subsequences and the pairs of these subsequences in the dataset using permutation testing and dynamic programming. Experiments have shown that two distinct TFs with the same binding site preference can still act differently when other factors are present.


11.10.2011 - 15:08 Leena Salmela
10.10.2011 - 16:29 Leena Salmela