Table 3 Dependent measure accuracies (i.e. comparison of test to reference alignment) for AMPS run on the Master data set at various gap-penalties sorted by Acc _SCR – Clustering was performed on SD score (from 100 randomisations) with the PAM250 matrix. Acc _SCR : Accuracy of Structurally Conserved Regions. AC _a : Accuracy average (pairwise). AC _w : Accuracy of whole alignment. PSE : Position Shift Error. NA in columns marked p indicates the highest accuracy in the preceding column, p gives the Wilcoxon Signed Rank test probability that the difference to the highest scoring row occured by chance.

From: OXBench: A benchmark for evaluation of protein multiple sequence alignment accuracy