Skip to main content

Table 5 Performance benchmark with the SL dataset under parameters at default values.

From: Parameterization of disorder predictors for large-scale applications requiring high specificity by using an extended benchmark dataset

Method

threshold

sensitivity

specificity

MCC

PE

DISOPRED2

0.05

0.645

0.897

0.567

0.541

IUPred long

0.5

0.596

0.924

0.560

0.520

IUPred short

0.5

0.513

0.942

0.515

0.454

CAST

40

0.448

0.951

0.474

0.399

SEG45

3.40;3.75

0.527

0.880

0.441

0.407

DisEMBL Rem465

1.2

0.314

0.979

0.407

0.293

SEG25

3.00;3.30

0.396

0.917

0.374

0.313

SEG12

2.20;2.50

0.213

0.972

0.293

0.184

DisEMBL Hotloops

1.4

0.456

0.801

0.275

0.257

DisEMBL Coils

1.2

0.750

0.464

0.220

0.214

  1. All predictors were run over the SL set using their respective default settings and ranked by the MCC. These settings produce results of varying levels of specificity which makes their ranking more dependent on the used overall performance measure (e.g. MCC or PE).
Back to article page