view article

Figure 5
Comparison of DAQ-refine with five other existing methods. For each of the 13 targets the model with the highest score was selected from the models generated by each method. Blue diamonds, Rosetta relax; orange triangles, MDFF; green crosses, phenix.real_space_refine; red circles, phenix.dock_and_rebuild; purple triangles, AF2 + Rosetta relax. The phenix.dock_and_rebuild and AF2 + Rosetta relax protocols start by predicting the structure of the target protein using AF2. AF2 + Rosetta relax uses the initial protein model as a template protein structure in AF2. The other refinement methods started from the initial protein model. For our DAQ-refine protocol, the model with the highest DAQ(AA) score was selected. (a) Comparison is made in terms of r.m.s.d. to the native structure. Two models generated by phenix.dock_and_rebuild and AF2 + Rosetta relax have large r.m.s.d.s. of 25.6 and 24.8 Å and were not included in this plot. (b) Comparison is made in terms of GDT-HA.

Journal logoSTRUCTURAL
BIOLOGY
ISSN: 2059-7983
Follow Acta Cryst. D
Sign up for e-alerts
Follow Acta Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds