Merging Segments From Varscan2 Output...

Question

[sousuffer]

I am a little new to this so any help would be appreciated - I ran a variation of the standard example workflow for VarScan2 copy number variant caller (using my exome sequencing data). I visualized the results of the segmentation (Step 4 - first few lines of output shown below) and it looks good, so I am now trying to implement Step 5: Merge adjacent segments of similar copy number and classify events by size (large-scale or focal).

chrom loc.start loc.end num.mark seg.mean
chr1 126901 336305 20 -0.2205
chr1 566191 569388 3 1.2087
chr1 653310 2338123 1505 -0.1378
chr1 2338223 16057066 5079 -0.19
chr1 16057166 16258373 168 -0.1205
chr1 16258673 16258973 4 -1.2245
chr1 16259073 17012168 579 -0.1526
chr1 17012268 17273478 48 0.0673
chr1 17282515 26109165 4152 -0.1866

I am trying to do this using the mergeSegments.pl package and am getting the following message: ../../apps/mergeSegments.pl out.file

Use of uninitialized value $input in <handle> at ../../apps/mergeSegments.pl line 446.
readline() on unopened filehandle at ../../apps/mergeSegments.pl line 446.
Can't use an undefined value as a symbol reference at ../../apps/mergeSegments.pl line 456.

I am trying to figure out what I am doing wrong. Initially, I figured that my file was in the wrong format. The mergeSegments specifies that the input should be:

"A segments file with p-values from the DNAcopy library. This should be tab- or space-delimited with a header and the following columns: ID, chrom, loc.start, loc.end, num.mark, seg.mean, bstat, pval, lcl, ucl."

My file doesn't have an ID, bstat, pval, lcl, or ucl and I'm not even sure how I would generate these if indeed this was the issue. Any help would be greatly appreciated. Thank you very much.

Answer 1

[Chris Miller]

I believe this answer is what you need to generate those additional columns: Varscan Copy Number Calling Workflow: Using Mergesegments.Pl After Cbc

Answer 2

[sousuffer]

Thanks for the advice...I generated the correct columns and the mergeSegments.pl still fails:

perl ../../apps/mergeSegments.pl outp.file
Use of uninitialized value $input in <handle> at ../../apps/mergeSegments.pl line 446.
readline() on unopened filehandle at ../../apps/mergeSegments.pl line 446.
Can't use an undefined value as a symbol reference at ../../apps/mergeSegments.pl line 456.

This is the file format: ID chrom loc.start loc.end num.mark seg.mean bstat pval lcl ucl
Sample.1 chr1 126901 336305 20 -0.2205 7.35726563517599 3.50306367498813e-12 336305 336305
Sample.1 chr1 566191 569388 3 1.2087 8.08753912963223 1.21687750240143e-13 569388 569388
Sample.1 chr1 653310 2338123 1505 -0.1378 6.06563083831769 2.6260014877838e-07 2234624 2446703
Sample.1 chr1 2338223 16057066 5079 -0.19 3.1268048900692 0.149143434568713 15904118 16174570

Any help would be greatly appreciated. Thanks.

[cris.giu]

IDs need to be unique. Sample.1 Sample.2 ... and so on.

Answer 3

[sousuffer]

Can anyone help with this?

Answer 4

[aheikkila]

There are some discrepancies in the documentation and in the code. Does this help you: A: Question about mergeSegment.pl

Answer 5

[dankoboldt]

I believe this was an issue with the ref-arm-sizes file and has been resolved.

Answer 6

[lmesrob]

Hi,

I've got same errors and here is the solution:

1. Add $stats{'num_merged_events'} in line 109:

$stats{'num_merged_events'} = $stats{'num_variants'} = $stats{'num_fail_pos'} = $stats{'num_fail_strand'} = $stats{'num_fail_varcount'} = $stats{'num_fail_varfreq'} = $stats{'num_fail_mmqs'} = $stats{'num_fail_var_mmqs'} = $stats{'num_fail_mapqual'} = $stats{'num_fail_readlen'} = $stats{'num_fail_dist3'} = $stats{'num_pass_filter'} = 0;

2. Remove $sample in line 134:

my ($id, $chrom, $chr_start, $chr_stop, $num_mark, $seg_mean, $bstat, $p_value, $lcl, $ucl) = split(/\s+/, $line);

3. By the way, assure that the $ref_arm_sizes file is tab delimited and contains "chr" in the first column if you have "chr" before the chromosomes.

Hope this helps,

Lilia