Segmentation fault in full_alignment

[chris-cheshire]

Hey there, I am running the containerised version of clair3 with 32GB of mem. Its running on a small viral genome so shouldnt be too taxing. The seg fault happens and then the merged vcf is just the pileup vcf instead of the true merged. Any ideas?

[INFO] BAM FILE PATH: /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/FAY66992_BC23.primertrimmed.rg.sorted.bam
[INFO] REFERENCE FILE PATH: /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/SARS-CoV-2.fasta
[INFO] MODEL PATH: /opt/models/r941_prom_hac_g360+g422
[INFO] OUTPUT FOLDER: /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/.
[INFO] PLATFORM: ont
[INFO] THREADS: 10
[INFO] BED FILE PATH: EMPTY
[INFO] VCF FILE PATH: EMPTY
[INFO] CONTIGS: EMPTY
[INFO] CONDA PREFIX: 
[INFO] SAMTOOLS PATH: samtools
[INFO] PYTHON PATH: python3
[INFO] PYPY PATH: pypy3
[INFO] PARALLEL PATH: parallel
[INFO] WHATSHAP PATH: whatshap
[INFO] LONGPHASE PATH: EMPTY
[INFO] CHUNK SIZE: 800
[INFO] FULL ALIGN PROPORTION: 1.0
[INFO] FULL ALIGN REFERENCE PROPORTION: 1.0
[INFO] PHASING PROPORTION: 0.7
[INFO] MINIMUM MQ: 5
[INFO] MINIMUM COVERAGE: 2
[INFO] SNP AF THRESHOLD: 0.08
[INFO] INDEL AF THRESHOLD: 0.15
[INFO] BASE ERROR IN GVCF: 0.001
[INFO] GQ BIN SIZE IN GVCF: 5
[INFO] ENABLE FILEUP ONLY CALLING: False
[INFO] ENABLE FAST MODE CALLING: False
[INFO] ENABLE CALLING SNP CANDIDATES ONLY: False
[INFO] ENABLE PRINTING REFERENCE CALLS: False
[INFO] ENABLE OUTPUT GVCF: False
[INFO] ENABLE HAPLOID PRECISE MODE: True
[INFO] ENABLE HAPLOID SENSITIVE MODE: False
[INFO] ENABLE INCLUDE ALL CTGS CALLING: True
[INFO] ENABLE NO PHASING FOR FULL ALIGNMENT: True
[INFO] ENABLE REMOVING INTERMEDIATE FILES: False
[INFO] ENABLE LONGPHASE FOR INTERMEDIATE VCF PHASING: False
[INFO] ENABLE PHASING FINAL VCF OUTPUT USING WHATSHAP: False
[INFO] ENABLE PHASING FINAL VCF OUTPUT USING LONGPHASE: False
[INFO] ENABLE HAPLOTAGGING FINAL BAM: False
[INFO] ENABLE LONG INDEL CALLING: False
[INFO] ENABLE C_IMPLEMENT: True

�[33m[WARNING] Threads setting exceeds maximum available threads 4, set threads=4�[0m
+ /opt/bin/scripts/clair3_c_impl.sh --bam_fn /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/FAY66992_BC23.primertrimmed.rg.sorted.bam --ref_fn /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/SARS-CoV-2.fasta --threads 4 --model_path /opt/models/r941_prom_hac_g360+g422 --platform ont --output /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/. --bed_fn=EMPTY --vcf_fn=EMPTY --ctg_name=EMPTY --sample_name=FAY66992_BC23 --chunk_num=0 --chunk_size=800 --samtools=samtools --python=python3 --pypy=pypy3 --parallel=parallel --whatshap=whatshap --qual=2 --var_pct_full=1.0 --ref_pct_full=1.0 --var_pct_phasing=0.7 --snp_min_af=0.08 --indel_min_af=0.15 --min_mq=5 --min_coverage=2 --min_contig_size=0 --pileup_only=False --gvcf=False --base_err=0.001 --gq_bin_size=5 --fast_mode=False --call_snp_only=False --print_ref_calls=False --haploid_precise=True --haploid_sensitive=False --include_all_ctgs=True --no_phasing_for_fa=True --pileup_model_prefix=pileup --fa_model_prefix=full_alignment --remove_intermediate_dir=False --enable_phasing=False --enable_long_indel=False --keep_iupac_bases=False --use_gpu=False --longphase_for_phasing=False --longphase=EMPTY --use_whatshap_for_intermediate_phasing=True --use_longphase_for_intermediate_phasing=False --use_whatshap_for_final_output_phasing=False --use_longphase_for_final_output_phasing=False --use_whatshap_for_final_output_haplotagging=False

[INFO] Check environment variables
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/log
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/tmp/pileup_output
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/tmp/merge_output
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/tmp/phase_output
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/tmp/gvcf_tmp_output
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/tmp/full_alignment_output
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/tmp/phase_output/phase_vcf
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/tmp/phase_output/phase_bam
[INFO] Create folder /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/tmp/full_alignment_output/candidate_bed
Warning: cannot find your CPU L2 cache size in /proc/cpuinfo
[INFO] --include_all_ctgs enabled
�[93m[WARNING] For efficiency, we use a maximum 30% reference candidates for full-alignment calling�[0m
[INFO] Call variant in contigs: NC_045512.2
[INFO] Chunk number for each contig: 38
[INFO] 1/7 Call variants using pileup model
Calling variants ...
Total processed positions in NC_045512.2 (chunk 1/38) : 100
Total time elapsed: 1.77 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 3/38) : 72
Total time elapsed: 1.73 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 2/38) : 112
Total time elapsed: 2.10 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 4/38) : 63
Total time elapsed: 1.71 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 5/38) : 73
Total time elapsed: 1.70 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 6/38) : 82
Total time elapsed: 1.74 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 8/38) : 60
Total time elapsed: 1.69 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 7/38) : 62
Total time elapsed: 1.68 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 9/38) : 76
Total time elapsed: 1.66 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 10/38) : 69
Total time elapsed: 1.68 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 11/38) : 47
Total time elapsed: 1.67 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 12/38) : 48
Total time elapsed: 1.61 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 14/38) : 44
Total time elapsed: 1.62 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 13/38) : 67
Total time elapsed: 1.71 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 15/38) : 57
Total time elapsed: 1.68 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 16/38) : 51
Total time elapsed: 1.70 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 17/38) : 61
Total time elapsed: 1.68 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 18/38) : 58
Total time elapsed: 1.63 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 19/38) : 51
Total time elapsed: 1.64 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 21/38) : 63
Total time elapsed: 1.67 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 20/38) : 78
Total time elapsed: 2.02 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 22/38) : 61
Total time elapsed: 1.67 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 23/38) : 56
Total time elapsed: 1.67 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 24/38) : 54
Total time elapsed: 1.72 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 25/38) : 56
Total time elapsed: 1.71 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 26/38) : 55
Total time elapsed: 1.71 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 27/38) : 61
Total time elapsed: 1.70 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 28/38) : 74
Total time elapsed: 1.76 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 29/38) : 100
Total time elapsed: 1.74 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 30/38) : 78
Total time elapsed: 1.68 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 31/38) : 61
Total time elapsed: 1.70 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 32/38) : 67
Total time elapsed: 1.65 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 33/38) : 71
Total time elapsed: 1.64 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 34/38) : 72
Total time elapsed: 1.73 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 35/38) : 57
Total time elapsed: 1.74 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 36/38) : 68
Total time elapsed: 1.73 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 37/38) : 81
Total time elapsed: 1.59 s
Calling variants ...
Total processed positions in NC_045512.2 (chunk 38/38) : 69
Total time elapsed: 1.57 s

real	0m46.009s
user	2m2.892s
sys	0m4.457s
Warning: cannot find your CPU L2 cache size in /proc/cpuinfo
[INFO] 2/7 No phasing for full alignment calling

[INFO] 5/7 Select candidates for full-alignment calling
Warning: cannot find your CPU L2 cache size in /proc/cpuinfo
[INFO] Set variants quality cutoff 28.0
[INFO] Set reference calls quality cutoff 32.0
Warning: cannot find your CPU L2 cache size in /proc/cpuinfo
[INFO] Low quality reference calls to be processed in NC_045512.2: 2268
[INFO] Low quality variants to be processed in NC_045512.2: 156

real	0m0.945s
user	0m0.893s
sys	0m0.058s

[INFO] 6/7 Call low-quality variants using full-alignment model
Calling variants ...
Segmentation fault

real	0m7.052s
user	0m6.251s
sys	0m0.473s
Warning: cannot find your CPU L2 cache size in /proc/cpuinfo
�[93m[WARNING] No vcf file found, output empty vcf file�[0m
�[93m[WARNING] Copying pileup.vcf.gz to /Users/cheshic/dev/repos/pipelines-tech/wic-viral-assembler/work/9c/7bfa116a33775c60e7dc6c3b2a0065/./merge_output.vcf.gz�[0m
[INFO] Exit in full-alignment variant calling

real	0m57.367s
user	2m13.145s
sys	0m5.249s

[chris-cheshire]

I should also say I am running on an ARM M3 mac and that I have 2 samples, one of which runs fine and the other produces this error

[aquaskyline]

It is also what we have observed on our side. Most of the cases run to the end with Apple Silicon. Some samples failed with 'illegal instructions' or 'segmentation fault', which might probably be fixed in the future version of docker on mac.

[chris-cheshire]

Thanks for the reply, do you have any idea whats causing it?

[aquaskyline]

It's at the virtualization layer in Docker so it's complicated. Is it possible if you could try your samples in a Linux setup as well?

[aquaskyline]

there is another way, could you please try adding the option --disable_c_impl. it switches back to using the python implementation, which might cause fewer problems with the virtualization.

[chris-cheshire]

Ah that worked thank you!