Table 1 Thirteen representative eGenes (9 most significant genes and 4 GWAS loci) that correlate with BMI in METSIM and TwinsUK (for the full data on all 54 genes, see Supplementary Table 6)

From: Integration of human adipocyte chromosomal interactions with adipose gene expression prioritizes obesity-related genes from GWAS

Pearson

Linear regression

Ranka

Gene

Chrf

METSIMc

METSIMd

TwinsUKe

Effect size (r)

p-value

Effect size (β)

SE

p-value

Effect size (β)

SE

p-value

1

ADH1B

4

−0.45

7.40 × 10−18

−0.21

0.02

1.68 × 10−20

−0.58

0.03

4.47 × 10−71

2

ORMDL3 b

17

−0.45

8.57 × 10−18

−0.16

0.02

2.06 × 10−20

−0.58

0.03

2.65 × 10−70

3

AKR1C3

10

0.33

4.78 × 10−10

0.13

0.02

2.95 × 10−11

0.49

0.03

5.19 × 10−54

4

CMTM3

16

0.41

4.32 × 10−15

0.087

0.01

3.84 × 10−17

0.50

0.03

6.64 × 10−52

5

LPIN1

2

−0.38

1.49 × 10−13

−0.14

0.02

2.27 × 10−15

−0.47

0.03

2.38 × 10−44

6

RNF157

17

−0.29

5.19 × 10−8

−0.096

0.02

5.87 × 10−9

−0.47

0.03

8.86 × 10−42

7

MYOF

10

0.32

1.07 × 10−9

0.086

0.01

7.37 × 10−11

0.46

0.03

2.59 × 10−40

8

NAA40

11

0.28

1.81 × 10−7

0.052

0.009

2.67 × 10−8

0.46

0.03

4.00 × 10−40

9

TMEM165

4

0.33

2.45 × 10−9

0.045

0.007

1.84 × 10−10

0.45

0.03

3.52 × 10−37

10

RFFL

11

0.27

1.02 × 10−6

0.035

0.006

1.84 × 10−7

0.43

0.03

5.67 × 10−37

28

ACADS b

12

−0.37

2.91 × 10−12

−0.085

0.01

7.12 × 10−14

−0.24

0.03

6.65 × 10−19

31

LACTB b

15

0.30

1.67 × 10−8

0.069

0.01

1.40 × 10−9

0.32

0.04

4.94 × 10−18

34

MAP2K5 b

15

−0.25

7.83 × 10−6

−0.039

0.01

1.90 × 10−6

−0.21

0.03

3.81 × 10−10

  1. a Thirteen representative eGenes, including 4 GWAS loci, ranked by their p-value in the TwinsUK cohort dataset
  2. b GWAS gene
  3. c Effect size (r, Pearson rho) and p-value calculated from Pearson correlation between gene expression and BMI (see Methods)
  4. d Effect size, standard error (SE), and p-value calculated using a linear regression model with BMI and age, age² and the 14 technical factors as covariates when compared to a null model without BMI. These models were compared using an F-test (see Methods)
  5. e Effect size, standard error (SE), and p-value calculated from linear mixed effects model. A full model including BMI was compared to a null model in which the same model was fitted, but with BMI omitted. These models were compared using an F-test (see Methods)
  6. f Chr is an abbreviation for chromosome