Response to the comment on: Exploring the impact of design criteria for reference sets on performance evaluation of signal detection algorithms: The case of drug-drug interactions.