Deep template-based protein structure prediction
Fig 3
Alignment quality (TM-score and GDT) comparison between DRNF, CNFpred, NDThreader and DeepThreader on the CASP13 alignment test set.
Top: average alignment quality (TM-score and GDT) on the CASP13 alignment test set. GDT is scaled to [0, 1]. Bottom left: NDThreader vs. DeepThreader. Bottom right: DRNF vs. CNFpred. Each point represents the quality of two alignments generated by two competing methods for the same protein pair.