Linking SNPs to DNA methylation calls

Question

[Dataminer]

Hi,

I am almost sure at least few of you have solved this problem or have some ideas on how to proceed.

I have two sets of data, one set is SNP calls and the other one is DNA methylation calls. I want to identify the SNPs that may have direct impact on DNA methylation and on gene expression (also available). I am kind of lost, may be I should first take the SNPs that are present within the DNAmethylation regions and look at gene expression or see if the site is hypo or hyper methylated.

Unfortunately, I am unable to see a concrete way forward. Could anyone of you who has dealt with such a problem in the past, be kind enough to shed some light and guide.

Thank you

Answer 1

[Sean Davis]

A common approach is to use regression of genotype (independent variable) on methylation values (dependent variable), typically using all SNPs within some genomic window of the methylation measurement. PLINK has functionality for doing this, and R can do this as well. A search of the literature for methQTL analysis will turn up some useful references.