PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

Module extraction in autoregressive models : application to gene regulatory networks inference.
Nicolas J-B. Brunel, Yousri Slaoui and Florence d'Alché-Buc
In: Second International Workshop on Machine Learning in Systems Biology, 13-14 septembre 2008, Bruxelles.

Abstract

Complex regulatory mechanisms at work in the cell are assumed to involve different subsets of genes, mRNA and proteins that behave more or less independently, implementing different biological functions. UNder this biological assumption, models of gene regulatory networks should incorporate the notion of subnetworks or modules. We propose two new algorithms of modules extraction in first order autoregressive models estimated from the data? The first one consists in two steps : thresholding the estimated transitions matrix using cross-validation and then searching for an appropriate permutation and the empirical matrix. The second one consists in 3 STEPS We successfully tested the 2 methods on simulated data and on kinetics of gene expression measure form specific skin cellsduring the switch between proliferation and differentaition . Comparison between these two-methods and with related methods sowh that first appraoch outperfoms significantly the others.

EPrint Type:Conference or Workshop Item (Poster)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
Learning/Statistics & Optimisation
Theory & Algorithms
ID Code:5223
Deposited By:Nicolas Brunel
Deposited On:24 March 2009