The work of Rene Thomas [1] on qualitative modelling of regulatory circuits 
is a well known approach to gene network modelling and has been successfully applied
to modelling of known biological systems. However, despite its usefulness in
modelling known networks, it is not suitable for automated network inference
from experimental data. We examine the consequences of introducing the stochastic factor 
in the form of white noise into the differential equations representing regulatory interactions. 
We will present the class of Markov Chains appearing as a result of this extension 
and propose an Expectation-Maximization technique of estimating such models from time-series
data.

References

[1] R.Thomas and R.D'Ari, Biological Feedback, CRC Press.