search

Illumina / GTCtoVCF

Script to convert GTC/BPM files to VCF
Apache License 2.0
41 stars 31 forks source link

Bus error when converting GTC to VCF #67

Closed khzhu closed 3 years ago

khzhu commented 3 years ago

Hi Illumina team, we got an "Bus error" when trying to convert a GTC file to VCF. The GTC file was generated from the Infinium LCG Assay idat files using iaap-cli gencall tool on linux.

  1. error message 2020-10-21 12:32:35,934 - GTC converter - WARNING - Reference allele is not queried for locus: rs876659253_mnv 2020-10-21 12:32:35,941 - GTC converter - WARNING - Reference allele is not queried for locus: ilmnseq_rs878853738_mnv 2020-10-21 12:32:35,957 - GTC converter - WARNING - Reference allele is not queried for locus: rs886040595_mnv Bus error
  2. command used ~/molecpathlab/bin/GTCtoVCF/gtc_to_vcf.py --gtc-paths /gpfs/data/molecpathlab/development/Infinium_LCG/test --manifest-file /gpfs/data/molecpathlab/data/Infinium/GDA-8v1-0_A1.csv --genome-fasta-file /gpfs/data/molecpathlab/data/Infinium/genome.fa --output-vcf-path /gpfs/data/molecpathlab/development/Infinium_LCG/test --log-file /gpfs/data/molecpathlab/development/Infinium_LCG/test/Infinium_LCG_test.log --skip-indels --disable-genome-cache
jjzieve commented 3 years ago

Hi @khzhu, this seems like a lower-level python error I have not encountered before. Can you post more details on your environment (python version, exact OS, memory, CPU)? Also can you try not providing a --log-file?

khzhu commented 3 years ago

Hi @jjzieve , not sure if the log file helps but it's attached for your reference. Infinium_LCG_test.log below is our environment:

  1. python : Python 3.6.5
  2. OS: Red Hat Enterprise Linux 7.4
  3. Cores: 39
  4. RAM: 395GB
jjzieve commented 3 years ago

ok seems like its not a hardware issue then, you have more than enough resources. Can you send us one of you GTCs so we can try to reproduce?

khzhu commented 3 years ago

@jjzieve , I've already sent the file to the Illumina technical support. It is really patient data and cannot share it on Github.

jjzieve commented 3 years ago

@khzhu I wasn't able to reproduce the issue. techsupport can point you to my log and the created VCF file. Is this reproducible on another machine you have access to (or after a reboot)? Are there more logs if you run python -v (i.e. with tracing). This SO post suggests a Bus error could be an actual hardware problem after all? It's difficult to determine the exact cause since I can't reproduce on my end.

khzhu commented 3 years ago

@jjzieve , I got the issue resolved. The error was caused by missing PyVCF and pysam two modules. After pip install dependencies, I was able to convert my GTCs to VCFs. It will be helpful to have a requirements.txt included in GTCtoVCF git branch.

jjzieve commented 3 years ago

@khzhu that's good to hear. I realize https://github.com/Illumina/GTCtoVCF#requirements is not as straightforward as a requirements.txt. I referenced this issue in a previous semi related issue (#17) to track creating a requirements.txt.