Skip to main content

Table 3 The list of predictors included in the final binary (BNN) and 3-stage hair greying classification (MNN) models

From: Exploring the possibility of predicting human head hair greying from DNA using whole-exome and targeted NGS data

BNN model (greying vs. no greying)

MNN model (no greying vs. mild greying vs. severe greying)

Rank

SNP_ID

Chr position

GRCh38

Gene/Locus

mRMRe Score

AUC

Rank

SNP_ID

Chr position

GRCh38

Gene/Locus

mRMRe Score

AUC

No greying

Mild greying

Severe greying

1

Age

–

–

0.2429273

0.863

1

Age

–

–

0.2555938

0.859

0.788

0.892

2

rs59733750

2:240780193

KIF1A

0.0026937

0.864

2

Sex

–

–

0.0071533

0.861

0.803

0.891

3

Sex

–

–

0.0026841

0.866

3

rs7680591

4:80276795

FGF5

0.0026000

0.864

0.8

0.893

4

rs68088846

21:34835870

RUNX1

0.0026483

0.869

4

rs59733750

2:240780193

KIF1A

0.0024376

0.867

0.802

0.901

5

rs1005241

22:47291868

TBC1D22A

0.0023004

0.869

5

rs10928235

2:144920547

TEX41

0.0024138

0.869

0.803

0.901

6

rs7680591

4:80276795

FGF5

0.0022084

0.878

6

rs68088846

21:34835870

RUNX1

0.0023223

0.867

0.804

0.899

7

rs2361506

2:233830694

MROH2A

0.0020803

0.878

7

rs2361506

2:233830694

MROH2A

0.0023167

0.87

0.808

0.897

8

rs12203592

6:396321

IRF4

0.0016857

0.880

8

rs45483393

9:89378809

SEMA4D

0.0020403

0.87

0.806

0.899

9

rs45483393

9:89378809

SEMA4D

0.0015962

0.887

9

rs12203592

6:396321

IRF4

0.0020377

0.875

0.81

0.901

10

rs2416699

9:119434462

BRINP1

0.0015622

0.886

10

rs1005241

22:47291868

TBC1D22A

0.0016952

0.875

0.804

0.890

11

rs164741

16:89625890

DPEP1

0.0014107

0.888

11

rs2416699

9:119434462

BRINP1

0.0016313

0.878

0.812

0.896

12

rs1683723

12:128415460

TMEM132C

0.0010814

0.900

12

rs2814331

10:86233584

GRID1

0.0009898

0.879

0.821

0.903

–

–

–

–

–

–

13

rs164741

16:89625890

DPEP1

0.0008983

0.881

0.819

0.899

–

–

–

–

–

–

14

rs1127228

16:27226789

NSMCE1

0.0005584

0.894

0.836

0.904

  1. mRMRe score and the impact on prediction accuracy measured by AUC values in a 849-sample cohort was presented for particular predictors