Fig. 6From: Constructing effective energy functions for protein structure prediction through broadening attraction-basin and reverse Monte Carlo samplingQuality of predicted structures using the original weights and the optimized weights of energy terms on 101 benchmark proteins of the Test101 dataset. For each benchmark protein, we predicted structures using both original weights and optimized weights of energy terms, and showed RMSD of the predicted structures as x-axis and y-axis, respectively. Most proteins fall below the diagonal line, suggesting that when using the optimized weights of energy terms, the predicted structures usually have smaller RMSDBack to article page