diffBind dba.count error for a peaks file selected

Question

[iridha]

Hello,

I am trying to count the peaks based on CpG island. first I imported the ed file for CpG island and run the dba.count on it. I get the error of Error in mcfork() : unable to fork, possible reason: Cannot allocate memory.

am running it on cluster and basically should not get memory issue, any idea on that?

library(rtracklayer) CpG_sites_import <- import("GRCh38_fullChr_CpGSites.bed", format="bed") marks_obj_CpG_island <- dba.count(marks_peaks, peaks=CpG_sites_import)

Answer 1

[Rory Stark]

The issue could be related to running this in the default parallel mode. How many samples do you have? By default dba.count() will grab all the cores on your machine and attempt to count each sample in parallel. If you have a lot of "peaks" (all CpGs!), this could take a lot of memory for each parallel process.

You could try running this with bParallel=FALSE to run serially, or set marks_peaks$config$cores to a lower number to control how many processes get spawned (prior to calling dba.count()).