Combining logical formalisation and constraint programming to model wild-type and perturbed genetic networks

Genetic regulatory networks play a crucial role in the control of most developmental and physiological processes. Extensive genetic sequencing now provides us with almost complete sets of genes, whereas genetic and functional genomic analyses progressively uncover the numerous cross-regulatory interactions occurring between these genes and their products. However, at present, these analyses lead essentially to qualitative rather than to reliable quantitative data. This situation has motivated the development of a qualitative, logical approach for the modelling of these genetic networks [5]. Associating discrete variables, functions and parameters to each element of a network, this logical approach emphasise the roles of regulatory circuits. According to the parity of the number of negative interactions, feedback circuits are classified into positive (even) versus negative (odd) circuits, endowed with very different properties. Indeed, positive circuits are necessary to generate differentiation (multistability), whereas negative circuits generate stable oscillatory behaviour [6]. More recently, a constraint programming approach to solve large logical models has been proposed, enabling the location of all stable states or the selection of proper parameter sets generating specific combination of stable states [2]. This approach is implemented in a Java software dedicated to the logical modelling and analysis of genetic networks [3]. Here, we report how these approach and modelling platform have been extended to design virtual experiments, encompassing various types of perturbations, from loss of function and cis-regulatory mutations, to ectopic gene expression.