Enter the body of text here

Hi , I am interested to use EpiDISH analysis for the tissue of origin. It seems input data should have cpG region. Unfortunately, I have genomic regions, and the data looks like the following. How could I use EpiDISH on this data set.

seqnames start end c1 c2 c3 c4 c5 c6 c7 c8 c9 c10

1 chr1 788801 789000 0.38229376 0.28131635 0.54229935 0.25362319 0.3808256 0.36501080 0.35294118 0.39784946 0.36206897 0.42260062 2 chr1 789801 790000 0.34383202 0.38427948 0.48932039 0.43383948 0.4440344 0.42606516 0.42679558 0.43659044 0.47068146 0.44897959 3 chr1 1044401 1044600 0.95161290 0.92400000 0.92070485 0.92771084 0.9447514 0.94852941 0.94603175 0.88372093 0.95561358 0.93406593 4 chr1 1110801 1111000 0.95098039 0.96840149 0.96956522 0.96167247 0.9367089 0.94240838 0.92800000 0.95681063 0.93939394 0.93822394 5 chr1 1165001 1165200 0.81048387 0.86111111 0.75247525 0.90990991 0.9180000 0.84615385 0.85294118 0.86186186 0.88679245 0.70812183 6 chr1 1206401 1206600 0.02846975 0.01733102 0.03088803 0.02459016 0.0156250 0.03529412 0.02230483 0.04790419 0.03585657 0.05777778 Thanks

Code should be placed in three backticks as shown below

# include your problematic code here with any corresponding output 
# please also include the results of running the following in an R session 

sessionInfo( )

Hi Bony,

EpiDISH needs CpG methylation beta value to estimate the cell type proportions. If you want to use it on the region data, you could potentially map the CpG site in your region and take the region methylation as a proxy for that CpG site. The rationale behind is that the methylation levels of the neighboring CpGs are similar. Whether this trick works or not depends on the region length. I cannot guarantee it works well on your data. You could also consider other tools that use differentially methylated regions to estimate cell type proportion.