Seminar Details
| Date |
20-12-2006 |
| Time |
15:00 |
| Room/Location |
DISI, Sala conferenze 322, 3° piano |
| Title |
Probabilistic modelling of transcriptional networks: from genome-wide to single input motifs |
| Speaker |
Dr Guido Sanguinetti |
| Affiliation |
Department of Computer Science, University of Sheffield UK |
| Link |
https://www.disi.unige.it/index.php?eventsandseminars/seminars
|
| Abstract |
In this talk I will describe some recent work in collaboration with Neil Lawrence and Magnus Rattray. We propose a framework to integrate microarray and location data to infer active transcription factor (TF) protein concentrations and the sensitivity with which target genes respond to TF's binding. We model the dynamics of the protein concentrations using linear state-space models, and devise efficient algorithms that make the inference task
tractable on a genome-wide scale. The results on real data sets capture many known biological relations. Furthermore, a spectral analysis on the response sensitivities allows us to identify small approximately independent submodules of the regulatory network that could be modelled
at a more fine grained level. Finally, I will describe some preliminary results on inference in single input motif networks, where we assume the dynamic of the process to be described by a linear differential equation and we model the TF concentration as a continuous Gaussian process.
|
|
|
 |
