Reaction Networks to Boolean Networks

Emilie Allart (1, 2) , Joachim Niehren (2, 3) , Cristian Versari (1, 2)

1 Université de Lille
2 BioComputing
3 LINKS - Linking Dynamic Data

We study the problem of how to convert reaction networks into boolean networks. We start from the sign abstraction of the ODE semantics of reaction network, and show that it can be captured by a quasi-boolean network, a generalization of boolean networks that we introduce. We then show that any quasi-boolean networks can be converted into a boolean networks by using John’s overapproximation. We then apply a recent rewriting algorithm base on elementary modes to make the linear equation system exact with respect to the boolean abstraction, for improving the precision of the conversion of quasi-boolean to boolean networks, also in the general nonlinear case. Thereby we improve our compiler from reaction networks to boolean networks as well.

Mots clés: abstract interpretation; Linear equation systems; Systems biology; Sign abstraction; Reaction networks; Boolean networks

Domaines: Bio-informatique [q-bio.QM]; Informatique [cs]

Emilie Allart, Joachim Niehren, Cristian Versari. Reaction Networks to Boolean Networks. 2021. ⟨hal-02279942v6⟩