I ran following sets of commands for getting differential methylation profiles & CNVs :

library(MEDIPS)
library(BSgenome.Mmusculus.UCSC.mm9)

BSgenome="BSgenome.Mmusculus.UCSC.mm9"

uniq=1e-3

extend=300

shift=0

ws=100

WT_Input = MEDIPS.createSet(file = "/Medip-seq_project/Alignment_BAMs_and_BEDs/3-WT-Input_S3.bam", BSgenome = BSgenome, extend = extend, shift = shift, uniq = uniq, window_size = ws, chr.select=NULL, paired = T)

WT_5mC = MEDIPS.createSet(file = "/Medip-seq_project/Alignment_BAMs_and_BEDs/1-WT-5mC_S1.bam", BSgenome = BSgenome, extend = extend, shift = shift, uniq = uniq, window_size = ws, chr.select=NULL, paired = T)

LshKO_Input = MEDIPS.createSet(file = "/Medip-seq_project/Alignment_BAMs_and_BEDs/4-LshKO-Input_S4.bam", BSgenome = BSgenome, extend = extend, shift = shift, uniq = uniq, window_size = ws, chr.select=NULL, paired = T)

LshKO_5mC = MEDIPS.createSet(file = "/Medip-seq_project/Alignment_BAMs_and_BEDs/2-LshKO-5mC_S2.bam", BSgenome = BSgenome, extend = extend, shift = shift, uniq = uniq, window_size = ws, chr.select=NULL, paired = T)  #chr.select=NULL doesn't affect further steps.

CS1 = MEDIPS.couplingVector(pattern = "CG", refObj = WT_Input)

mr.edgeR = MEDIPS.meth(MSet1 = WT_5mC, MSet2 = LshKO_5mC, CSet = CS1, ISet1 = WT_Input , ISet2 = LshKO_Input, p.adj = "bonferroni", diff.method = "edgeR", MeDIP = T, CNV = T, minRowSum = 10)

I got this error :

Creating results table...
Adding differential coverage results...
Adding CNV results...
Error in cnv.combined[i, 1]:cnv.combined[i, 2] : NA/NaN argument
In addition: Warning messages:
1: In MEDIPS.rms(MSet1[[i]], CSet, ccObj = ccObj) : NaNs produced
2: In MEDIPS.rms(MSet2[[i]], CSet, ccObj = ccObj) : NaNs produced

What does this error means? If I run function MEDIPS.meth with "CNV = F" jsut to get differntial methylation profiles then it's coming fine. But how to get CNVs?

Please help me out. Thank you :)

Hi Anupriya Verma,

MEDIPS.createSet calculates CNVs based on the window size selected for the primary MeDIP seq analysis. In most cases, the input coverage is not sufficient for CNV detection at this resolution. Maybe, that this somehow leads to the error message you got. The suggested way would be to set CNV=F in this step, and then add CNV information using MEDIPS.addCNV. This function allows the user to specify a different window size. See also the notes on the CNV parameter at the MEDIPS.meth help page

> ?MEDIPS.meth

Furthermore, the error might be somehow related to the nonchoromosomal conigs in the mm9 reference. You could restrict the analysis on the chromosomes, e.g. by setting

> chr.select=paste0("chr",c(1:19,"X",Y"))

By the way, our new analysis package "QSEA" provides far more options to estimate and deal with CNV (among other benefits). In particular, it allows using the CNV within the normalization, and thereby considers CNV information in the detection of differentially methylated regions, rather than just providing this information in the output table. For MeDIP seq analysis it can completely replace the MEDIPS package -except from some QC metrics, that are not implemented in QSEA.

Best, Matthias