Agent-based model of the IFNB1-DDX58 feedback loop