I am very new to bioinfromatics and have never handled DNA methylation data before. I have Illumina EPIC array data and I am looking for specific CpG probes that are within 2.5 megabases from the start of specific genes of interest. I am most interested in CpG probes in the promoter regions of my genes of interest. I want to create a list of these CpG probes based on their location and only use them to test for differentially methylated probes between my two groups of patients.
I have no idea how to do this, so any help would be really helpful.
•
written
by
sswang25 •
You can start by listing the coordinates of the regions you are interested in on genome assembly hg19. Then you download the Illumina EPIC annotation file from bioconductor :
library(IlluminaHumanMethylationEPICanno.ilm10b4.hg19)
data("IlluminaHumanMethylationEPICanno.ilm10b4.hg19")
data(Locations)
locEPIC=data.frame(Locations)
Finally you just have to retrieve the coordinates of the CpGs that are located within the list of coordinates of genes of interest you identified
Hi thank you very much. So in essence I make a dataframe from genomic ranges of the areas of the genome I'm interested in. Using the dataframe I can get the CpG probe names in these regions from the Epic manifest?
Is there a way to find out which CpG probes are in Cpg Islands, promoter regions etc from the Epic manifest?
Thanks very much! I will try both these solutions out
Log in to answer this question.
I managed to get the chromosome number and the start position and end position of all of my genes of interest:
I have the Chromosome number and MAPINFO i.e. location of my CpG sites:
How do I go about finding out which CpG probes are situated within the start/end positions of the genes I have identified? I can't figure out how to code this selection in R as the chromosome column has to match and the MAPINFO value I assume has to between the start and end values.
Create two Granges objects corresponding to your gene coordinates and the probes : Beware, first you need to transform your chromosome columns to "chr1" "chr2" ... and not "1" "2" "3" After this,
Then, use the findOverlaps() function from GenomicAlignments package :
findOverlaps(probesRanges,genesRanges)