The DREAM challenges have provided a platform to understand network inference algorithm behaviors and experimental data characteristics, unlikely to be as productive if done by individual researchers or labs. This provides an opportunity to evaluate ones work more objectively, in addition to the peer-review system.
DREAM Challenges have greatly influenced the research activities in our lab. We studied the performance of our algorithms in DREAM3,4,5 for their inherent limitations. This led to the design of a novel algorithm in 2010 now called FunChisq that claimed a best performer in DREAM8 network inference challenge. Without DREAM challenges, it would have been unlikely for us to invent the algorithm and even more unlikely to appreciate its power on diverse experimental data.