Human 450K methylation data: Tumor vs Normal vs Control

Question

[arpit.singh1203]

Hello

I intend to use Infinium/Illumina 450K human methylation array data in cancer studies. I need to understand if data for normal sample is required? I am planning to download data from tcga (where normal data isn't available).

Also another basic doubt is "if the normal and tumor sample is mixed when loading onto the slide"?

Answer 1

[Sean Davis]

Tumor and normal are not hybridized on the same slide. Whether or not you need normal is not a question of bioinformatics, but of biology. In other words, you need to define what questions you are going to address. If those questions include comparison to normals, you will need normals.

[arpit.singh1203]

For studying methylation changes in a disease, comparison to "normal" is required. Is that right?

Also to find the differentially methylated regions, "normal" or "control" must be required. Please correct me if I am thinking in a wrong direction.

[Sean Davis]

What is the biological question?

[arpit.singh1203]

I get your point. Thank you.