Figure 1From: Designing and evaluating the MULTICOM protein local and global model quality prediction methods in the CASP10 experiment The per-target correlation scores of each target against the average real quality of the largest model cluster divided by the average real quality of all models of this target on Stage 2. Back to article page