Issues in FACETS Tumor Purity Estimation with Nanopore Sequencing Data

[Fenne-FU]

Hello.

I'm trying to run FACETS version 0.6.2 to extract information about tumor purity from the BAM file created by merging HCC1395 normal and tumor samples.

These BAM files were generated from nanopore sequencing.

As described in the document,

I ran SNP-pileup using the following command:

./snp-pileup -g -q15 -Q20 -P100 -r25,0 00-common_all.vcf.gz \
HCC1395_75x_merged_40pc_snp-pileup.gz \
alignment-sort-hcc1395bl.bam \
HCC1395_merged_tumor_purity_40pc_sorted.bam

and successfully obtained the output of SNP-pileup.

Then, I attempted to run the following commands in R:

library(facets)
set.seed(1234)
datafile = "HCC1395_75x_merged_80pc_snp-pileup.gz"
rcmat = readSnpMatrix(datafile)
xx = preProcSample(rcmat)
oo=procSample(xx,cval=150)

However, I encountered the following error:

Error in abs(lorclust[[jhi]]$valor) : 
  non-numeric argument to mathematical function

I also tested using the stomach.gz file from GitHub, the BAM file created by merging the normal and tumor samples from the HCC1395 dataset released by Google DeepSomatic, and the merged BAM file of COLO829.
Both stomach.gz and the HCC1395 merged BAM file successfully completed the purity estimation using FACETS, but I encountered the same error when analyzing the COLO829 merged BAM file.

Could these errors be related to the format or content of the BAM file?

Thank you!
Fenne

[veseshan]

From the error it seems like it is hitting an empty vector but I can't tell why.
I do not have any experience with nanopore data. Can you provide a snapshot of it?