Completing SBGN-AF Networks by Logic-Based Hypothesis Finding

This study considers formal methods for finding unknown interactions of incomplete molecular networks using microarray profiles. In systems biology, a challenging problem lies in the growing scale and complexity of molecular networks. Along with high-throughput experimental tools, it is not straightforward to reconstruct huge and complicated networks using observed data by hand. Thus, we address the completion problem of our target networks represented by a standard markup language, called SBGN (in particular, Activity Flow). Our proposed method is based on logic-based hypothesis finding techniques; given an input SBGN network and its profile data, missing interactions can be logically generated as hypotheses by the proposed method. In this paper, we also show empirical results that demonstrate how the proposed method works with a real network involved in the glucose repression of S. cerevisiae.