Got Error When Using Exomecnv To Call Loh

Question

[Conan]

Hi all, I'm trying to use ExomeCNV for my paired normal-tumor exome sequencing data, to infer if there's any significant LOH in my tumor sample and if does, it'll be reasonable to use SNP array to do further LOH analysis. I followed user guide (https://secure.genome.ucla.edu/index.php/ExomeCNV_User_Guide#LOH_Calling_Procedure) and created BAF files for samples, and called LOH on each heterozygous position. An error occured when following the step "Combine multiple positions into LOH segments":

the.loh = multi.LOH.analyze(normal, tumor, all.loh.ls=list(eLOH),test.alpha=0.001, method="variance.f", sdundo=c(0,0), alpha=c(0.5,0.1))

Analyzing: Sample.1 Error in sort(abs(diff(genomdat)))[1:n.keep] :

only 0's may be mixed with negative subscripts

so it means n.keep is negative? It's actually accomplished by package DNAcopy in this step, I've read the code's part where exactly stuck on website: http://search.bioconductor.jp/codes/11539 but unfortunately still have no idea what should I do, is it correlated with BAF file which might be abnormal? Any advice would be really appreciated, thanks.

Answer 1

[Pascal]

Hi Conan,

I got the same error. After removing extra chromosomes (I had GLXXXX included) everything worked fine.

Pascal

Answer 2

[ysh]

found same error when using DNAcopy package recently, found out just some column is empty in input data.(table header is fine, but row below is not complete.