Hi CNAnorm users

I am using CNAnorm for identifying somatic CNV's from shallow WGS of FFPE samples. I have access to the tumor and the paired control samples. I have used CNAnorm on one of the pairs and managed to get the genome output. I see that the estimated ploidy for this sample is 5. I know these are tumor samples, but wondering if it would be possible to assume  a normal diploid genome and ask CNVnorm to analyze assuming a diploid genome. Any other recommendation of a different tool to process SCNA's from shallow WGS is highly appreciated. Many thanks for your help.

Here is the script I am using for CNAnorm


#################################

library(CNAnorm)
data <- read.table("1270_LCIS_mod", header=T, sep="\t")
CN <- dataFrame2object(data)
toSkip <- c("chrY", "chrM")
CN <- gcNorm(CN, exclude=toSkip)
CN <- addSmooth(CN, lambda=7)
CN <- peakPloidy(CN, exclude=toSkip, method="density")    # I used density method to get integer copy number levels from sample
plotPeaks(CN)


CN <- validation(CN)

seg_list <- list(alpha=0.01, min.width=5, undo.splits="sdundo", undo.SD=2)
CN <- addDNACopy(CN, DNAcopy.weight = "gaussian", DNAcopy.segment = seg_list)

CN <- discreteNorm(CN)

plotGenome(CN, superimpose='DNACopy', show.centromeres=FALSE)

##################################

I have tried the sugg.ploidy function, which actually gives me the ploidy as 0,1,5

Venu

Hi Venu.

You can try  method="closest" in peakPloidy to normalise it so that the peak closest to the meadian is considered "normal".

This way, whoever, not "absolute ploidy estimation" will be perfromed. What is now ploidy 5 will become ploidy 2 and the rests is normalised accordingly.

Alternatively, you can follow the instructions on the "Manual correction" chapter of the vignette and using validation, sugg.peaks, sugg.ploidy, you can override the number, location and ploidy of the peaks and re-do the genome wide normalisation. You could assign the peak that now has ploidy 5 to ploidy2, leave the peak with ploidy 0 at ploidy 0, and ignore the peak at ploidy 1.

I hope this is useful