mergeSegments.pl varscan confusion !

Question

[poisonAlien]

Hi all,

I know this question has been asked so many times on this forum. But still none of the answers seem to help solve the problem with mergeSegments perl script. The --ref-arm-sizes - it says that it should be of 2 column files with chromosome name and sizes but here, on the discussion forum, author says it should be 4 column file with arm sizes. But, neither of the file seems to work for me. It constantly gives error saying

Argument "NA" isn't numeric in numeric ge (>=) at mergeSegments.pl line 149, <GEN1> line 8294.
Argument "NA" isn't numeric in addition (+) at mergeSegments.pl line 363, <GEN1> line 8295.
Argument "NA" isn't numeric in numeric ge (>=) at mergeSegments.pl line 149, <GEN1> line 8570.
Argument "NA" isn't numeric in numeric ge (>=) at mergeSegments.pl line 149, <GEN1> line 8596.
Use of uninitialized value in split at mergeSegments.pl line 255.
Use of uninitialized value $chrom_size in division (/) at mergeSegments.pl line 271.
Illegal division by zero at mergeSegments.pl line 271.

Can any one help with this. And, on the side, I have one more question - is it appropriate to use varScan2 to check for CNAs on a targetted capture sequencing data (around 100 genes)?

Thanks.

[ingrid.schulman]

Has there been any progress on this?

[mgreenwood]

I believe it is appropriate to use for targeted capture, especially if paired normal is available. I think removing NA lines from DNAcopy output might help with NA error. I'm having the 255 and 271 errors and can't figure out how to get by them. I'll include some perl remove statements with the caveat that I'm still stuck.

possibly remove NA with: perl -i.bak -ne'print unless /NA/;' DNAcopy.p_value.file

or (if you have other non-interesting CNV calls you can remove them too):

perl -i.bak -ne'print unless /_/ or /chrM/ or /NA/;' DNAcopy.p_value.file