the multiple testing problem is harder in network inference than in other contexts because the involved statistical tests are typically not independent.