search

Kingsford-Group / squid

SQUID detects both fusion-gene and non-fusion-gene structural variations from RNA-seq data
BSD 3-Clause "New" or "Revised" License
40 stars 22 forks source link

squid running errors #17

Open cauyrd opened 5 years ago

cauyrd commented 5 years ago

Hello, when I analyze my BWA mapped BAM files, I got the following error:

squid: src/SegmentGraph.cpp:2489: void SegmentGraph_t::CompressNode(): Assertion `LinkedNode.s ize()!=0' failed. /var/spool/torque/mom_priv/jobs/10249993.mesabim3.msi.umn.edu.SC: line 9: 6359 Aborted (core dumped) squid -b lncap.sv_build.bam -o lncap_squid --bwa

The stand output is below: Reference name 1 --> 0 Reference name 10 --> 9 Reference name 11 --> 10 Reference name 12 --> 11 Reference name 13 --> 12 Reference name 14 --> 13 Reference name 15 --> 14 Reference name 16 --> 15 Reference name 17 --> 16 Reference name 18 --> 17 Reference name 19 --> 18 Reference name 2 --> 1 Reference name 20 --> 19 Reference name 21 --> 20 Reference name 22 --> 21 Reference name 3 --> 2 Reference name 4 --> 3 Reference name 5 --> 4 Reference name 6 --> 5 Reference name 7 --> 6 Reference name 8 --> 7 Reference name 9 --> 8 Reference name MT --> 24 Reference name X --> 22 Reference name Y --> 23 [Fri Jan 4 10:03:53 2019] Starting reading bam file. [Fri Jan 4 11:18:49 2019] Building nodes, finish seeding. [Fri Jan 4 11:18:51 2019] Building nodes, finish expanding to whole genome. [Fri Jan 4 11:18:57 2019] Finish calculating reads per node. [Fri Jan 4 11:18:57 2019] Starting building edges. [Fri Jan 4 11:33:39 2019] Finish raw edges. [Fri Jan 4 11:33:39 2019] Finish filtering edges from multi-aligned reads. [Fri Jan 4 11:33:39 2019] Finish adding partial aligned reads. [Fri Jan 4 11:33:39 2019] Finish building edges. Error: 0 nodes are connected by edges.

What is the reason to get such error? memory issue? I run squid with 12 cores and 30gb total memory. My BAM file is 7.5gb. My command is: squid -b lncap.sv_build.bam -o lncap_squid --bwa

Congm12 commented 5 years ago

The error indicates that no edges are constructed. It shouldn't be because of memory. Can you make sure that your BAM alignments are sorted? And can you check that chimeric alignments exist in the BAM by selecting sam flag not to be 14 and count alignments:

samtools view -F14 <bam file> | wc -l

If the BAM is sorted and contains chimeric alignments, is it possible for you to share the BAM file for us to further look into the error?

Browco commented 4 years ago

Hi, I know that this issue is 1 year old but I had the same problem and I found a/the solution. It may help some people coming here with the same problem. Based on @Congm12 answer, I saw that no chimera were found in my alignment where I was supposed to (experimental control). I changed my bwa algorithm command from bwa aln to bwa mem and it worked. (Based on : http://bio-bwa.sourceforge.net/bwa.shtml ) NB: I have Illumina reads