How to map CpG ids to nearest genes?

Question

[preethibagopi5]

I am working in dna methylation data and I am having a .csv file containing cpg ids. I want to map the cpg sites to nearest gene. How to do that in R?

Thanks in advance

Answer 1

[Devon Ryan]

Load the relevant txdb database and use the nearest() command, presumably with the transcripts.

[Kangli Wang]

Hi, Devon. Could you give a detail? I also have a question. When two genes are near the CpG, how to decide the nearest one. Should the TSS, promoter and gene body be considered? @Devon Ryan

[Devon Ryan]

Honestly it's a bit arbitrary what one considers. Biologically speaking, it'd make sense to consider genes downstream (when considering the gene's strand) or containing a SNP, but I'd have to check and see if that's possible with nearest().

Answer 2

[nandobonf]

If you are working with Illumina 450k or similar, you can use the manifest table to annotate your probes. See here: https://bioconductor.org/packages/release/data/annotation/html/IlluminaHumanMethylation450kmanifest.html

Answer 3

[preethibagopi5]

I try to install IlluminaHumanMethylation450k.db package for annotation but I came across this error.

Error: package or namespace load failed for 'IlluminaHumanMethylation450k.db':
 .onLoad failed in loadNamespace() for 'IlluminaHumanMethylation450k.db', details:
  call: NULL
  error: 'fun' is defunct.
Use 'FDb.InfiniumMethylation.hg19' instead.
Use 'FDb.InfiniumMethylation.hg18' instead.
Use 'mapToGenome() function in minfi or methylumi' instead.
See help("Defunct")
Error: loading failed
Execution halted
ERROR: loading failed
* removing 'C:/R/R-3.4.1/library/IlluminaHumanMethylation450k.db'

[Devon Ryan]

Please note the "ADD COMMENT" button, which can be used to reply to specific posts.