Skip to main content
. 2012 Mar 9;28(9):1216–1222. doi: 10.1093/bioinformatics/bts110

Table 2.

AUC on beta-barrel superfamilies

HMMER RAPTOR HHPred SMURF-Lite 1 SMURF-Lite 1, SimEv SMURF-Lite 2 SMURF-Lite 2, SimEv
SMURFLite performs best
    Translation proteins 0.66 0.93 0.92 0.93 0.93
    Barwin-like endoglucanases 0.75 0.77 0.63
    Cyclophilin-like 0.67 0.61 0.7 0.82 0.85 0.82 0.83
    Sm-like ribonucleoproteins 0.73 0.71 0.77 0.76 0.71 0.76 0.85
    Prokaryotic SH3-related domain 0.81 0.83 0.82 0.83 0.83
    Tudor/PWWP/MBT 0.78 0.74 0.67 0.83 0.77 0.83 0.79
    Nucleic acid-binding proteins 0.75 0.67 0.76 0.89 0.76 0.92
HHPred performs best
    Translation proteins SH3-like 0.83 0.81 0.86 0.62 0.62
RAPTOR performs best
    PDZ domain-like 0.96 1.0 0.99 0.97 0.97 0.97 0.97
    FMN-binding split barrel 0.62 0.82 0.61
HMMER performs best
    Electron Transport accessory proteins 0.84 0.77 0.63 0.63 0.66

Note: for SmurfLite, the number (1,2) indicates the interleave threshold, and SimEv is simulated evolution. A dash (‘–’) in a result entry indicates the method failed on these structures, i.e. an AUC of <0.6.