I want to get the genotype recall from idat files. I have few thousands samples, and I'm not sure how to do that efficiently. I used the next set of commands:
library(crlmm) library(ff) library(snow) library(doSNOW) cl <- makeCluster(90, type="SOCK") registerDoSNOW(cl) options("krlmm.cores" = 90) ocProbesets(150e3) ocSamples(5000) dir_name = ‘set1’ # Directory contains 500 samples g = genotype.Illumina(cdfName='humanomni25quadv1b', path=dir_name, call.method='krlmm', verbose=TRUE)
Can someone please help me in fine tuning the different parameters:
ocProbesets, ocSamples, krlmm.cores, snow, ff
Are all needed or are they masking each-other?
Can someone give me rough estimation of runtime? Should I expect hours, days or weeks?
I have access to different machine with up to 100 cpus and up to 1T RAM, how can I take advantage of such?