Errors when writing XLSX file

[code12412f]

I'm running Vaxrank and I'm getting two errors. I'm trying to get a XLSX neoepitope report as output:
vaxrank --vcf 303Aligned.sortedByCoord.out.bam_dup.bam.vcf --bam 303Aligned.sortedByCoord.out.bam_dup.bam --output-neoepitope-report test.csv --genome GRCm38 --mhc-predictor netmhc4 --mhc-alleles H2-Kb,H2-Db

with traceback: Traceback (most recent call last):
File "/home/kam9898/.local/lib/python3.6/site-packages/vaxrank/epitope_prediction.py", line 110, in predict_epitopes
{protein_fragment.gene_name: protein_fragment.amino_acids})
File "/home/kam9898/.local/lib/python3.6/site-packages/mhctools/base_predictor.py", line 208, in predict_subsequences
binding_predictions = self.predict_peptides(peptide_list)
File "/home/kam9898/.local/lib/python3.6/site-packages/mhctools/base_commandline_predictor.py", line 326, in predict_peptides
temp_dir_list=dirs)
File "/home/kam9898/.local/lib/python3.6/site-packages/mhctools/base_commandline_predictor.py", line 271, in _run_commands_and_collect_predictions
process_limit=self.process_limit)
File "/home/kam9898/.local/lib/python3.6/site-packages/mhctools/process_helpers.py", line 114, in run_multiple_commands_redirect_stdout
assert len(multiple_args_dict) > 0
AssertionError

Traceback (most recent call last):
File "/home/kam9898/.local/bin/vaxrank", line 10, in
sys.exit(main())
File "/home/kam9898/.local/lib/python3.6/site-packages/vaxrank/cli.py", line 407, in main
excel_report_path=args.output_neoepitope_report)
File "/home/kam9898/.local/lib/python3.6/site-packages/vaxrank/report.py", line 506, in make_minimal_neoepitope_report
'%.2f nM' % epitope_prediction.wt_ic50),
TypeError: must be real number, not NoneType

[iskandr]

Hi Kim / @code12412f!

Sorry no one else caught this on here...can you help me recreate the error?

I am able to run this on my machine:

mhctools --mhc-predictor netmhc4 --mhc-alleles H2-Kb,H2-Db --sequence SIINFEKL

What happens when you run the same command on yours?

[code12412f]

2019-07-03 09:18:51,494 - mhctools.cli.args - INFO - Building MHC binding prediction type for alleles ['H-2-Kb', 'H-2-Db'] and epitope lengths None
2019-07-03 09:18:53,592 - mhctools.base_commandline_predictor - INFO - Skipping allele BoLA-D18.4: The suffix '.4' of 'D18.4' was not parsed
2019-07-03 09:18:53,593 - mhctools.base_commandline_predictor - INFO - Skipping allele BoLA-JSP.1: The suffix 'SP.1' of 'JSP.1' was not parsed
2019-07-03 09:18:53,593 - mhctools.base_commandline_predictor - INFO - Skipping allele BoLA-T2C: The suffix 'C' of 'T2C' was not parsed
2019-07-03 09:18:53,593 - mhctools.base_commandline_predictor - INFO - Skipping allele BoLA-T2a: The suffix 'a' of 'T2a' was not parsed
2019-07-03 09:18:53,593 - mhctools.base_commandline_predictor - INFO - Skipping allele BoLA-T2b: The suffix 'b' of 'T2b' was not parsed
2019-07-03 09:18:56,455 - mhctools.process_helpers - INFO - Ran 2 commands in 2.8485 seconds
2019-07-03 09:18:56,467 - mhctools.cli.script - INFO -
source_sequence_name offset peptide allele affinity percentile_rank prediction_method_name length
0 PEPLIST 0 SIINFEKL H-2-Kb 19.37 0.06 netMHC 8
1 PEPLIST 0 SIINFEKL H-2-Db 3220.94 0.90 netMHC 8

[iskandr]

OK, so the MHC binding predictor works.

I tried running on some small mouse data locally and this command succeeded:

vaxrank --vcf b16.expressed.vcf --bam b16.combined.bam --output-neoepitope-report test.xlsx --genome GRCm38 --mhc-predictor netmhc4 --mhc-alleles H2-Kb,H2-Db

So...it must be something data dependent.

What happens if you run varcode --vcf 303Aligned.sortedByCoord.out.bam_dup.bam.vcf --one-per-variant --only-coding? How many coding effects are predicted?

Alternatively, if you want to upload the data somewhere I can access, I'll try running Vaxrank with it.

[code12412f]

I emailed you the VCF file

Varcode
Version: 0.9.1
Path: /home/kam9898/.local/lib/python3.6/site-packages/varcode
Traceback (most recent call last):
File "/home/kam9898/.local/bin/varcode", line 10, in
sys.exit(main())
File "/home/kam9898/.local/lib/python3.6/site-packages/varcode/cli/effects_script.py", line 66, in main
variants = variant_collection_from_args(args)
File "/home/kam9898/.local/lib/python3.6/site-packages/varcode/cli/variant_args.py", line 131, in variant_collection_from_args
variant_collections.append(load_vcf(vcf_path, genome=genome))
File "/home/kam9898/.local/lib/python3.6/site-packages/varcode/vcf.py", line 141, in load_vcf
reference_vcf_key)
File "/home/kam9898/.local/lib/python3.6/site-packages/varcode/vcf.py", line 483, in infer_genome_from_vcf
return infer_genome(reference_path)
File "/home/kam9898/.local/lib/python3.6/site-packages/varcode/reference.py", line 119, in infer_genome
reference_name = infer_reference_name(genome_object_string_or_int)
File "/home/kam9898/.local/lib/python3.6/site-packages/varcode/reference.py", line 98, in infer_reference_name
"Failed to infer genome assembly name for %s" % reference_name_or_path)
ValueError: Failed to infer genome assembly name for /projects/b1093/ezh2/mouse.fa

[iskandr]

If the reference isn't obvious in the VCF then you need to add --genome GRCm38 to the arguments for Varcode.

Can you also send me a link to the BAM? I won't be able to run Vaxrank without both.

[code12412f]

2041