Fig. 2: Accuracy of SARST2 and several state-of-the-art alignment algorithms.
From: SARST2 high-throughput and resource-efficient protein structure alignment against massive databases
a Results of information retrieval with Qry400 and SCOP-2.07 as the query and target datasets. Every query protein had a known number of family-level structural homologs in the target database. Based on the output hit list, precision measured how early and compactly homologs were retrieved (see “Methods”). Recall monitored the proportion of known homologs retrieved. The vertical axis indicates the average precision over 400 query proteins (n = 400). b Raw IR data. The average precision of the eleven recall levels is a standard accuracy measure for database searching22,32. SARST2 outperformed other methods in accuracy. The data of panel (a) are provided as a Source data file. clusteredDB clustered database, standardDB standard database.
