Age and DNA methylation data: HELP NEEDED

#1
I have data on patient Age as well as change in % DNA methylation levels for promotor regions. I seek to understand the impact of Age on the % change in DNA methylation levels. What is the best way to do this?

Paired t-test of Age vs % change in DNA methylation data yields ridiculously high p-values (indicating this is an inappropriate way of measurement).

Clearly, a solution to this problem is to create an age-dummy (i.e. age = 1 if age > 40, and age = 0 if age < 40) and perform a standard OLS-regression (linear regression). However, it seems as if our data does not fit a linear model. Is there any statistical method (preferable paired t-test) whereby I can test the impact of age on % change in DNA methylation levels, without creating a dummy variable or performing a linear regression?

If important in any respect, the programming is made in "R".

Reply much appreciated.

Best
 

Dason

Ambassador to the humans
#2
You mentioned you did a paired t-test? That doesn't seem right at all - that would tell if you if the average difference between age and % methylation is statistically significant - which isn't the question you want to ask at all.

Why are you creating a dummy variable instead of just doing linear regression directly?