Skip to main content

Table 4 Performance benchmark with the Remark 465 dataset under parameters at comparable high specificity level.

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

0.11

0.393

0.950

0.388

0.344

DisEMBL Rem465

1

0.316

0.958

0.338

0.274

IUPred short

0.55

0.328

0.947

0.318

0.275

IUPred long

0.59

0.285

0.948

0.277

0.233

DisEMBL Hotloops

3

0.204

0.950

0.196

0.153

SEG25

2.91;3.21

0.197

0.951

0.194

0.149

SEG12

2.29;2.59

0.188

0.943

0.162

0.130

SEG45

3.27;3.62

0.175

0.949

0.162

0.124

DisEMBL Coils

1.96

0.167

0.950

0.155

0.117

CAST

48

0.154

0.949

0.136

0.103

  1. Predictors were run and ranked as described in Table 3 but over the Remark 465 dataset which puts more emphasis on short disorder inside or flanking globular structures. For DisEMBL Remark 465 (in italic), parameter tuning only allowed a specificity of 0.958 as closest value to our criterion.
Back to article page

BMC Genomics

ISSN: 1471-2164

Contact us