Date(s) : 25/09/2017 iCal
13 h 00 min - 14 h 00 min
Multilevel gene regulatory networks can be encoded as maps on a subset of the Boolean state space {0,1}^n, for n sufficiently large.
Extending these maps to the full state space can lead to distortions of the asymptotic behaviour and the regulatory graph.
We introduce a method for extending the maps to the entire state space that guarantees the preservation of attractors and does not introduce new circuits in the regulatory structure.
The presence of regulatory circuits has been connected to specific asymptotic behaviours; in particular, global negative circuits are necessary for sustained oscillations.
We show that the presence of a local negative circuit is not necessary for the presence of cyclic attractors, in the Boolean setting, for n>=6, by converting a known example for the multilevel case.
We then give an overview of a constraint programming approach that enables the analysis of the problem for small values of n.
Catégories