Skip to main content

Table 2 Blind assessment showing improvements of SS-Rosetta’s performance against standard predictions’ in terms of number of better structures (out of 24) and GDT

From: Enhancing fragment-based protein structure prediction by customising fragment cardinality according to local secondary structure

Std vs SS- Rosetta

First model

Best model

Average of the Best 5 models

20,000 Decoys

15/24 (63%) + 6.3%

12/24 (50%) + 5.0%

18/24 (75%) + 3.1%

2000 Decoys

16/24 (67%) + 24.2%

18/24 (75%) + 11.5%

18/24 (75%) + 8.3%