Genome MuSiC calcRoiCovg crashing

Question

[E.Earley]

I'm having trouble getting calcRoiCovg to work on some paired exome alignments.

I did get calcRoiCovg to finish completely and successfully on a test ROI file with 6 rows

genome music bmr calc-covg --bam-list test_bam_input.tsv --output-dir test_genome_music_out/ --reference-sequence /opt/indexes/DNA/h_sapiens/hg19.fa --roi-file test.annot

However, when I provide the full merged and sorted ROI file for the full exome with this command:

genome music bmr calc-covg --bam-list test_bam_input.tsv --output-dir test_genome_music_out/ --reference-sequence /opt/indexes/DNA/h_sapiens/hg19.fa --roi-file sorted_exome_v5.annot

This error is returned:

*** stack smashing detected ***: calcRoiCovg terminated
Failed to execute: calcRoiCovg /home/eearley/somatic_vars/bams/DNA-Blood-2535_L1.SSXT13.sort.targetedfor95.bam /home/eearley/somatic_vars/bams/DNA-2535_L1.SSXT9.sort.targetedfor95.bam sorted_exome_v5.annot /opt/indexes/DNA/h_sapiens/hg19.fa test_genome_music_out/roi_covgs/DNA-2535.covg 6 8 20
Failed to execute: 'gmt music bmr calc-covg-helper --normal-tumor-bam-pair "DNA-2535    /home/eearley/somatic_vars/bams/DNA-Blood-2535_L1.SSXT13.sort.targetedfor95.bam        /home/eearley/somatic_vars/bams/DNA-2535_L1.SSXT9.sort.targetedfor95.bam" --roi-file "sorted_exome_v5.annot" --reference-sequence "/opt/indexes/DNA/h_sapiens/hg19.fa" --output-file "test_genome_music_out/roi_covgs/DNA-2535.covg"'

I then try to run the last command manually:

calcRoiCovg /home/eearley/somatic_vars/bams/DNA-Blood-2535_L1.SSXT13.sort.targetedfor95.bam /home/eearley/somatic_vars/bams/DNA-2535_L1.SSXT9.sort.targetedfor95.bam sorted_exome_v5.annot /opt/indexes/DNA/h_sapiens/hg19.fa test_genome_music_out/roi_covgs/DNA-2535.covg 6 8 20

returns:

*** stack smashing detected ***: calcRoiCovg terminated
Segmentation fault (core dumped)

Here's what I've tried / noticed:

• made sure ROI is 1-based and that headers in reference match those in the bams and the ROI file
• checked that I have full memory permissions, and I've tracked usage via top, but I'm not seeing any RAM use spikes
• I notice partial output generation. There's nothing in gene_covgs/ but I do have roi_covgs/DNA-2535.covg. Here's a tail of that file:

VAMP7   chrY:59274533-59274622  90      0       0       0       0
VAMP7   chrY:59274633-59274844  212     0       0       0       0
IL9R    chrY:59330338-59330512  175     0       0       0       0
IL9R    chrY:59335598-59335688  91      0       0       0       0
IL9R    chrY:59336200-59336459  260     0       0       0       0
IL9R    chrY:59336471-59336570  100     0       0       0       0
IL9R    chrY:59337046-59337145  100     0       0       0       0
IL9R    chrY:59338018-59338189  172     0       0       0       0
IL9R    chrY:59340274-59340374  101     0       0       0       0
#NonOverlappingTotals                   49592462        24120883        21478451        2442237

So it seems like there's a breakdown after the ROI coverage results and the translation into gene results.

Any help is greatly appreciated! Thanks

Answer 1

[Cyriac Kandoth]

"stack smashing" errors are related to when a running process (calcRoiCovg) tries to read data outside of the bounds of the memory addressing space that it has been allocated (a bitmask across a chromosome). My best guess is that your ROI file makes a reference to a region that is outside the bounds of anything in the reference FASTA. Write a quick script to make sure that your ROIs are all within the bounds of what's listed in the FASTA index (fa.fai).