salmon quant error: Segmentation fault (core dumped)

[martynakgajos]

The bug is primarily related to salmon (bulk mode) Describe the bug

Version Info: This is the most recent version of salmon.
# salmon (alignment-based) v0.14.1
# [ program ] => salmon 
# [ command ] => quant 
# [ libType ] => { U }
# [ alignments ] => { Analysis/Quantification/OJ33.sam }
# [ targets ] => { Analysis/Pinfish/corrected_transcriptome_polished_collapsed.fas }
# [ output ] => { Analysis/Quantification/OJ33 }
# [ noErrorModel ] => { }
# [ writeUnmappedNames ] => { }
Logs will be written to Analysis/Quantification/OJ33/logs
[2019-08-13 11:13:43.563] [jointLog] [info] Fragment incompatibility prior below threshold.  Incompatible fragments will be ignored.
Library format { type:single end, relative orientation:none, strandedness:unstranded }
[2019-08-13 11:13:43.563] [jointLog] [info] numQuantThreads = 4
parseThreads = 4
Checking that provided alignment files have consistent headers . . . done
Populating targets from aln = "Analysis/Quantification/OJ33.sam", fasta = "Analysis/Pinfish/corrected_transcriptome_polished_collapsed.fas" . . .done
[2019-08-13 11:13:46.239] [jointLog] [info] replaced 1 non-ACGT nucleotides with random nucleotides

processed 2000000 reads in current roundSegmentation fault (core dumped)

To Reproduce

Which version of salmon was used? salmon 0.14.1 How was salmon installed (compiled, downloaded executable, through bioconda)? downloaded executable Which reference was used? custom, transcriptom derived from nanopore long-reads using Pinfish pipeline Which read files were used? SAM file returned by minimap2 (nanopore long reads) Which which program options were used? salmon quant -l U -a Analysis/Quantification/OJ33.sam -t Analysis/Pinfish/corrected_transcriptome_polished_collapsed.fas -o Analysis/Quantification/OJ33 --noErrorModel --writeUnmappedNames > Analysis/Quantification/OJ33.log Desktop (please complete the following information): x86_64 GNU/Linux

[martynakgajos]

I can also add the result from valgrind

[k3yavi]

Hi @martynakgajos ,

It's very hard to say without looking at the data. If you can share a minimal data on which we can replicate the error, that'd be great.