Estimation of Possible Reaction States in Metabolic Pathways Using Inductive Logic Programming

This paper provides a method for estimating states of enzyme reactions in metabolic pathways. We first introduce a new model based on the logical viewpoint of enzyme function. The proposed model logically represents causal relations between concentration changes of metabolites and enzyme activities. When we observe the concentration changes of metabolites, we can assume which enzyme reactions are accelerated under the model. This computation can be realized using inductive logic programming (ILP), which can find a hypothesis that accounts for given observations with a background theory. In this paper, we use CF-induction, which is an ILP technique and has a unique feature that can integrate inductive and abductive inferences. CF-induction can compute not only possible reaction states that can explain the observations, but also general rules that are missing in the current background theory. Then, we perform experiments with simple metabolic pathways, and confirm that CF-induction can both estimate states of reactions and discover new causal rules.