Skip to main content
. 2025 Aug 26;26(4):bbaf391. doi: 10.1093/bib/bbaf391

Figure 5.

A multi-panel figure evaluating ProDualNet’s modeling strategies and ablation analyses. (a) Scatter plots showing the performance analysis of different ProDualNet’s modeling strategies, plotting AlphaFold2 structure confidence scores and TM-Score. (b) A bar graph comparing the performance of the pretrained ProDualNet model with ProteinMPNN. (c) A bar graph comparing the fine-tuning results of ProteinMPNN_mean and ProDualNet on training set and test set 1. (d) A bar graph analyzing the impact of different ratios of mixed augmented data in the NoiseMix strategy. (e) A bar graph evaluating the effect of the ESM2 language model and the recycling strategy on model performance.

Evaluation of ProDualNet’s modeling strategies and ablation analysis. (a) Performance analysis of ProDualNet’s modeling strategies, using AlphaFold2 structure confidence scores and TM-score. (b) Comparative analysis between the ProDualNet pretrained model and ProteinMPNN. (c) Comparison of fine-tuning ProteinMPNN_mean and ProDualNet on training and test set 1. (d) Analysis of the mixed augmented data ratio in the NoiseMix strategy. (e) Evaluation of the effect of the ESM2 language model and validation of the recycling strategy.