Effect of parameter variations on the inference of context-sensitive Probabilistic Boolean Networks

This paper presents the results of an investigation into the effect of parameter variation on model inference from gene expression data. The models in question are context sensitive Probabilistic Boolean Networks. It is usually necessary to observe a large number of sample points in order to infer the model parameters accurately. This is because the data can become trapped in some fixed point attractor cycles for long time periods. To tackle this problem, a novel sampling strategy for model inference also has been introduced in the paper.