Skip to main content
. 2019 Aug 20;87(12):1249–1262. doi: 10.1002/prot.25794

Figure 4.

Figure 4

A,B. The fraction of models improved by refinement according to (A) GDT_HA or (B) RMS_CA is comparable to results in the last two CASP rounds. Each panel shows the histogram of differences from the starting model for all submitted refinement models. C‐E. Violin plots indicate that the top three human groups in CASP13 (Baker, Feig, and Seok) have achieved quite consistent improvements in GDT_HA over the last four CASP rounds, with the exception of Baker in CASP10. For the Baker and Feig groups the median improvement was higher in CASP13 than in the previous three rounds