The performance of our method when using different thresholds of the cooperativity score. To check the robustness of our method against different thresholds of cooperativity scores, we tested four different thresholds (125, 120, 110 and 90). The figure shows that no matter which threshold is used, the performance of our method is always the same (i.e. superior to 10 out of 11 existing methods) on the performance index 1. This suggests that our method is robust against different thresholds of the cooperativity score.