WGBS - Methylation state at CpG to peak calls

Question

[Nico Gerstner]

Hello everyone,

I have downloaded a WGBS file from ENCODE in bedMethyl format (https://www.encodeproject.org/experiments/ENCSR890UQO/). The content of this file is the methylation state per CpG, e.g.:

track name="SL1615 MspIRRBS" description="GM12878-16_Ximat" visibility=2 itemRgb="On"
chr1    100003153   100003154   ...[omited]      2  0
chr1    1000170     1000171     ...[omited]     22  0

What I want to have is a bed file containing larger regions, somewhat like peaks called from the file above. Is there a tool available that can perform this task?

(My ultimate goal is to use the peak information to state whether an certain region in the genome (e.g. a promoter) can be called "methylated".)

Thank you!

[mark.ziemann]

some tips here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4165320/ I have used methylkit and it appears to work well. http://bioconductor.org/packages/release/bioc/vignettes/methylKit/inst/doc/methylKit.html

[Nico Gerstner]

Thank you very much for your reply! It seems to me that in the paper and the doc they are calculating differential scores for regions, i.e. they assume I have (at least) two files. Is there a possibility to calculate scores for regions with only one file? (It does not matter if the tool calculates the regions as well, or calculates scores for regions which I provide.)

[mark.ziemann]

The methylKit documentation includes an example where methylation scores are calculated on tiles (1kbp)

[Nico Gerstner]

Thank you very much! This was what I was looking for.