Skip to main content

Table 6 Comparison of performance on BM-Set1

From: Predicting peptides binding to MHC class II molecules using multi-objective evolutionary algorithms

Dataset

AUC

  

†SVRMHC

Gibbs

ARB

TEPITOPE

MOEA

Original

set1

0.711

0.799

0.666

0.760

0.760

 

set2

0.652

0.766

0.653

0.736

0.765

 

set3a

0.626

0.740

0.652

0.730

0.733

 

set3b

0.618

0.751

0.666

0.750

0.752

 

set4a

0.706

0.788

0.668

0.748

0.748

 

set4b

0.664

0.770

0.661

0.748

0.770

 

set5a

0.553

0.604

0.539

0.653

0.777

 

set5b

0.606

0.621

0.579

0.679

0.748

 

Southwood

0.912

0.862

0.514

0.490

0.784

 

Geluk

0.697

0.723

0.682

0.710

0.786

NR

set1

0.619

0.673

0.572

0.594

0.587

 

set2

0.581

0.665

0.640

0.653

0.685

 

set3a

0.578

0.598

0.600

0.598

0.660

 

set3b

0.577

0.692

0.669

0.699

0.713

 

set4a

0.597

0.671

0.575

0.573

0.599

 

set4b

0.577

0.669

0.651

0.655

0.690

 

set5a

0.544

0.601

0.536

0.646

0.790

 

set5b

0.593

0.610

0.572

0.671

0.743

 

Southwood

0.917

0.850

0.671

0.505

0.770

 

Geluk

0.655

0.697

0.510

0.670

0.768

  1. Comparison of AUC values of the BM-Set1 (DRB1*0401). †These values are based on smaller dataset sizes as SVRMHC didn't predict values for some of the peptides. The values from the Gibbs sampler were estimated from the matrix provided by the authors in [32].