Figure 4
From: A study of quality measures for protein threading models
Comparison between manual assessors ranking and average ranking from all other non-global measures using model based normalization. The groups that differ most between the manual ranking and the average ranking are shown. It should be noted that the official manual ranking used slightly different targets than we used here, therefore the manual ranking is not identical with the official CASP3 ranking this is to ensure that exactly the same targets were used in the comparison between the manual and automatic measures.