Clustering-based approach to identify solutions for the inference of regulatory networks