NA Purity values

[nesilin]

Hi!

I 'm running FACETS in a pediatric leukemia cohort and half of these samples have null purity values (NA). The BAM files come from a whole exome sequencing at x30 depth and have very few number of mutations per patient. I've run FACETS with different parameter's settings but I always get a NA purity for the same samples. Concretely, I have tried with different values (upper and lower from default) for snp.nbhd, cval, min.nhet but still the same samples failed to have a purity value.
This is an example of the many tests I have tried:

>snp-pileup -g 00-common_all.vcf -q15 -Q20 -P100 -r25,0 pileup_file normal.bam tumor.bam

>R
library("facets")
data = gunzip(paste(pileup_file , ".gz", sep=""))
rcmat = readSnpMatrix(data)
xx = preProcSample(rcmat, snp.nbhd = 700)
oo = procSample(xx, cval=150, min.nhet = 30)
fit = emcncf(oo) 

Can anyone help to know which is the appropriate way to run FACETS in my case? Which parameters should I change from default and how should I be changing them?

Thanks!

[veseshan]

NA for purity happens when the sample shows no discernible copy number changes. When there are no copy number changes the purity is unknowable from copy numbers.

There can be many reasons for seeing no copy number changes in the sample -

1. the tumor is driven by mutations, translocations or epigenetic changes
2. the coverage is low
3. the tumor sample is mostly normal cells

The coverage in your case is low 30x. One thing you can try is to get more snps with lower coverage from snp-pileup i.e. use -r15,0 instead of -r25,0. And use ndepth=15 in preProcSample. The results in such low coverage setting may not be reliable. But it's worth a try.

Hope this helps,
Venkat

[nesilin]

Thanks for the quick reply! I'll give it a try.