Table 1 All five lead variants from the GWAS analyses of natural log-transformed CFB of LDL-C, LDL-C percent change and natural log-transformed baseline LDL-C based on five statistical models: M1 (baseline adjustment, 1-df test and 2-step regression), M3 (baseline un-adjustment, 1-df test and 2-step regression), M5 (baseline adjustment and 2-df test), M6 (baseline un-adjustment and 2-df test) and M7 (baseline association only).

From: A statistical perspective on baseline adjustment in pharmacogenomic genome-wide association studies of quantitative change

Gene

SNP

CHR

BP

MA

MAF

Modela

\(\beta _{{{{\boldsymbol{y}}}}_0}\)b

\(\beta _{{{\boldsymbol{G}}}}\) or \(\beta _{{{{\boldsymbol{G}}}}_0}\)c

P_G

\(\beta _{{{{\boldsymbol{GT}}}}}\)d

P_GTd

P_2dfd

CELSR2/PSRC1/SORT1

rs599839

1

109822166

G

0.240

M1 (BJ-1dfT-2SR)

−0.490

−0.172

2.92E−14

   

M3 (BuJ-1dfT-2SR)

 

−0.156

6.19E−12

   

M5 (BJ-2dfT)

−0.493

−0.052

1.12E−12

0.026

8.12E−02

2.22E−12

M6 (BuJ-2dfT)

 

−0.049

1.74E−10

0.029

6.04E−02

2.44E−10

M7 (Baseline only)

 

−0.008

6.58E−02e

   

STAG1/SLC35G2/NCK1

rs79929954

3

136623748

G

0.015

M1 (BJ-1dfT-2SR)

−0.490

0.215

7.81E−03e

   

M3 (BuJ-1dfT-2SR)

 

0.214

7.91E−03e

   

M5 (BJ-2dfT)

−0.486

0.079

2.68E−03

−0.244

3.04E−06

2.05E−07e

M6 (BuJ-2dfT)

 

0.080

3.23E−03

−0.285

1.78E−07

1.58E−08

M7 (Baseline only)

 

−0.002

8.92E−01e

   

LPA

rs10455872

6

161010118

G

0.080

M1 (BJ-1dfT-2SR)

−0.490

0.287

5.94E−16

   

M3 (BuJ-1dfT-2SR)

 

0.240

1.38E−11

   

M5 (BJ-2dfT)

−0.499

0.090

6.82E−15

−0.050

3.06E−02

6.53E−15

M6 (BuJ-2dfT)

 

0.078

1.13E−10

−0.049

4.28E−02

1.20E−10

M7 (Baseline only)

 

0.024

3.68E−04e

   

SLCO1B1

rs4149056

12

21331549

C

0.162

M1 (BJ-1dfT-2SR)

−0.490

0.129

6.92E−07e

   

M3 (BuJ-1dfT-2SR)

 

0.141

5.20E−08e

   

M5 (BJ-2dfT)

−0.485

0.043

3.76E−07

−0.028

9.10E−02

5.95E−07e

M6 (BuJ-2dfT)

 

0.050

1.33E−08

−0.036

3.75E−02

1.12E−08

M7 (Baseline only)

 

−0.013

9.64E−03e

   

APOE

rs1065853

19

45413233

T

0.089

M1 (BJ-1dfT-2SR)

−0.490

−0.491

2.13E−48

   

M3 (BuJ-1dfT-2SR)

 

−0.388

1.32E−30

   

M5 (BJ-2dfT)

−0.535

−0.164

3.91E−50

0.081

1.59E−04

5.87E−52

M6 (BuJ-2dfT)

 

−0.130

7.99E−30

0.082

2.73E−04

1.52E−31

M7 (Baseline only)

 

−0.064

2.35E−23

   
  1. SNP single nucleotide polymorphism, CHR chromosome, BP base pair, MA minor allele, MAF minor allele frequency, \(\beta _{y_0}\) effect size of baseline variable, βG effect size of G (genotype) on CFB (from M1, M3, M5 and M6), \(\beta _{G_0}\) effect size of G (genotype) on baseline (from M7), P_G p value of G (genotype), βGT effect size of G*T (genotype by treatment interaction), P_GT p value of G*T (genotype by treatment interaction), P_2df p value of 2df test (joint test of genotype and genotype*treatment interaction).
  2. aBJ: Baseline-adjusted; BuJ: Baseline-unadjusted; 1dfT: 1 degree of freedom test; 2dfT: 2 degree of freedom test or joint test of genotype and genotype*treatment interaction; 1SR: 1-step regression; 2SR: 2-step regression. In M1, M3, M5 and M6, difference of natural log-transformed Simvastatin and Ezetimibe/Simvastatin on low-density lipoprotein cholesterol levels were used for analysis.
  3. bEffects calculated for the nature log-transformed baseline LDL-C.
  4. cEffects calculated with respect to the minor allele. A negative value indicates more intense drug (Simvastatin and Ezetimibe/Simvastatin) LDL-C lowering.
  5. dResults were only available in the 2df test model M5 and M6, which also tests the genotype*treatment interaction and joint test of genotype and genotype*treatment interaction.
  6. eNot reaching genome-wide significance (p < 5E−08). For 2df test methods, p values from the 2df test (P_2df) were used.
Back to article page