Skip to main content

Table 4 Blind assessment results of standard predictions’ 20,000 decoys against SS-based predictions’ 2000 decoys. Although our approach shows slightly better outcomes, there is no significant difference between the two sets of results

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

 First modelBest modelAverage of the Best 5 models
20,000 Decoys (Std. Predictions)
vs
2000 Decoys (SS-based. Predictions)
12/24 (50%) + 0.4%13/24 (54%) + 4.6%10/24 (42%) + 0.3%