Skip to main content

Advertisement

Table 2 The performance of the global quality predictions of our three tools and the other four methods in terms of average correlation, overall correlation, average real GDT-TS score of top 1 models ranked by each method, and average loss of top 1 models ranked by each method, evaluated on 84 CASP9 single-domain targets

From: SMOQ: a tool for predicting the absolute residue-specific quality of a single protein model with support vector machines

  Avg. correlation Over. correlation Avg. top 1 Avg. loss
Basic (85 targets) 0.737 0.737 0.588 0.082
Profile 0.708 0.658 0.589 0.080
Profile+SOV 0.696 0.681 0.594 0.075
ModelEvaluator 0.636 0.767 0.597 0.073
ProQ 0.494 0.707 0.563 0.110
ProQ2 0.662 0.787 0.607 0.066
QMEAN 0.733 0.803 0.594 0.078
  1. Basic, profile, and profile + SOV are the three single-model local QA tools (SMOQ) presented in this manuscript.
  2. The other four QA predictors are ModelEvaluator (predictor name in CASP9: MULTICOM-NOVEL), ProQ, ProQ2, and QMEAN. Top 3 QA predictors’ performances according to each metric were bolded.