Skip to main content

Table 2 Genes selected by elastic-net regularized multinomial logistic regression for the association between the genes and the dominant genotypes of M.tb strains

From: Strain-specific behavior of Mycobacterium tuberculosis in A549 lung cancer cell line

 

Gene Symbol

Elastic net

Fold Regulation (L3-CAS1 vs. Con.)

Adjusted P-value (L3-CAS1 vs. Con.)

Fold Regulation (L4.5 vs. Con.)

Adjusted p value (L4.5 vs. Con.)

Fold Regulation (L4.5 vs. L3-CAS1)

Adjusted p value (L4.5 vs. L3-CAS1)

1

CFLAR

100%

 − 1.544

0.082

1.376

0.117

2.125

0.044

2

TIFA

93.9%

 − 1.392

0.199

 − 1.206

0.399

1.154

0.256

3

HRAS

91.8%

3.279

0.0001

6.681

0.0001

2.038

0.001

4

IRAK1

82.36%

1.395

0.082

1.136

0.503

 − 1.228

0.256

5

CD14

77.93%

4.724

0.0001

 − 1.217

0.171

 − 5.749

0.0001

6

BTK

67.28%

 − 4.258

0.055

 − 440.602

0.001

 − 103.476

0.0001

7

MAPK8IP3

59.63%

 − 13.117

0.0002

 − 12.323

0.0003

1.064

0.114

8

IFNG

55.51%

 − 24.028

0.0001

 − 9.426

0.0002

2.549

0.001

9

JUN

41.91%

3.146

0.0001

12.582

0.0002

3.999

0.001

10

IL1A

39.77%

2.204

0.0001

3.160

0.0001

1.434

0.081

11

CD80

33.64%

 − 3.630

0.0001

 − 5.856

0.0002

 − 1.613

0.044

12

PSIP1

31.03%

 − 3.972

0.045

 − 2.075

0.126

1.914

0.011

13

PTGS2

29.73%

40.880

0.0001

84.449

0.0001

2.066

0.001

14

BCL3

26.79%

 − 12.098

0.035

 − 1.617

0.188

7.482

0.0001

15

HSPA1A

26.63%

10.581

0.0001

4.009

0.0002

 − 2.639

0.001

16

RHOA

26.27%

 − 1.245

0.055

1.526

0.003

1.900

0.011

17

CCL2

26.23%

2.023

0.045

5.993

0.0002

2.962

0.001

18

CXCL8

24.92%

6.364

0.0001

20.966

0.0001

3.294

0.0001

19

HSPD1

23.79%

7.345

0.0001

9.759

0.0002

1.329

0.114

20

TLR2

22.52%

 − 8.980

0.0001

1.651

0.0017

14.826

0.0001

21

EGFR

21.83%

 − 3.547

0.035

1.286

0.433

4.561

0.0001

Misclassification error rate

 

14.8%

lambda

 

0.00849

Alpha

 

0.50

  1. Lambda is the tuning parameter and it is for controlling the overall strength of the penalty. Alpha is a controlled parameter between LASSO (L1) and ridge (L2) penalties. Alpha = 0.5 corresponds to elastic-net regularization. Importance value for each selected gene was reported, p values adjusted by Benjamini-Hochberg-FDR correction at α = 0.05