| |
|
Univariable Models |
Multivariable Model |
| PC |
% Variation Explained in 2,095 CpG Sites |
β |
p-value |
R2LR x100 |
β |
p-value |
R2LR x100 |
| 1 |
50.65 |
-0.12 |
6.63E-06 |
12.72 |
-0.13 |
5.60E-08 |
|
| 2 |
9.53 |
0.15 |
0.014 |
10.34 |
0.16 |
3.34E-03 |
|
| 3 |
4.52 |
-0.69 |
8.84E-14 |
18.95 |
-0.72 |
1.03E-15 |
|
| 4 |
2.47 |
-0.41 |
4.79E-04 |
11.39 |
-0.43 |
7.45E-05 |
|
| 5 |
2.15 |
0.16 |
0.21 |
9.58 |
0.22 |
0.22 |
26.76 |
| 6-10 |
4.68 |
|
|
|
|
|
| Total |
74.00 |
|
|
|
|
36.54 |
|
| Model: Ageij=β0 + β1PCij + Wj+ εij (univariable model) |
| Table 3: Association between top 5 principal components (estimated from 2,095 site M-values significant with age, after Bonferroni correction for α=0.05) and age. |
