Skip to main content

Table 4 Obtained accuracy of the consensus predictions, in the test set of 20 outer membrane proteins

From: Evaluation of methods for predicting the topology of β-barrel outer membrane proteins and a consensus prediction method

METHOD

TYPE

SEQUENCE

Qβ

Cβ

SOV

Correctly predicted topologies

Correctly predicted barrel size

PRED-TMBB, ProfTMB, HMM-B2TMR

CONSENSUS

barrel

0.771

0.596

0.877

17

19

  

precursor

0.818

0.628

0.86

15

18

PRED-TMBB, ProfTMB, HMM-B2TMR, B2TMPRED

CONSENSUS

barrel

0.790

0.616

0.896

17

19

  

precursor

0.832

0.641

0.865

15

18

PRED-TMBB, ProfTMB, HMM-B2TMR, TBBPred-nn

CONSENSUS

barrel

0.809

0.635

0.917

18

20

  

precursor

0.839

0.653

0.867

15

18

PRED-TMBB, ProfTMB, HMM-B2TMR, TBBPred-svm

CONSENSUS

barrel

0.809

0.629

0.906

15

19

  

precursor

0.847

0.658

0.882

15

18

PRED-TMBB, ProfTMB, HMM-B2TMR, TBBPred-comb

CONSENSUS

barrel

0.791

0.607

0.894

17

20

  

precursor

0.833

0.648

0.859

15

18

PRED-TMBB, ProfTMB, HMM-B2TMR, TBBPred-nn/svm

CONSENSUS

barrel

0.824

0.638

0.92

17

19

  

precursor

0.85

0.647

0.871

13

17

PRED-TMBB, ProfTMB, HMM-B2TMR, B2TMPRED, TBBPred-nn/svm

CONSENSUS

barrel

0.825

0.637

0.927

17

18

  

precursor

0.854

0.652

0.876

15

17

PRED-TMBB, ProfTMB, HMM-B2TMR, B2TMPRED, TBBPred-nn

CONSENSUS

barrel

0.819

0.641

0.924

18

20

  

precursor

0.849

0.660

0.874

15

18

PRED-TMBB, ProfTMB, HMM-B2TMR, B2TMPRED, TBBPred-comb

CONSENSUS

barrel

0.807

0.625

0.907

17

19

  

precursor

0.845

0.658

0.868

15

18

PRED-TMBB, ProfTMB, HMM-B2TMR, B2TMPRED, TBBPred-svm

CONSENSUS

barrel

0.819

0.637

0.910

15

19

  

precursor

0.853

0.659

0.880

14

18

PRED-TMBB, ProfTMB, B2TMPRED, TBBPred-svm/nn

CONSENSUS

barrel

0.829

0.642

0.923

17

18

  

precursor

0.851

0.648

0.861

15

16

PRED-TMBB, ProfTMB, B2TMPRED, TBBPred-svm, TBBPred-nn, HMM-B2TMR, TMBETA-NET, PSI-PRED, BETA-TM

CONSENSUS

barrel

0.808

0.582

0.851

11

13

  

precursor

0.844

0.604

0.841

12

13

  1. We report the consensus of all the available methods, and the ones that were obtained using the 3 top-scoring HMMs combined in various ways with some of the top-scoring NN/SVM methods. The best results are highlighted with bold. For abbreviations see also Table 1.