pVACseq 3.0.0 fails on "Creating aggregated report"

[riederd]

Running the following command results in failure:

pvacseq run \
    --iedb-install-directory /opt/iedb \
    -t 10 \
    -p test_vep_phased.vcf.gz \
    -e1 8,9,10,11 \
    -e2 15,16,17,18,19,20,21,22,23,24,25 \
    --normal-sample-name test_normal \
     \
    --netmhc-stab \
    --tumor-purity 0.8 \
    --binding-threshold 500 --percentile-threshold 2 --top-score-metric lowest --expn-val 2 --maximum-transcript-support-level 5 --normal-vaf 0.01 --trna-vaf 0.02 --tdna-vaf 0.02 \
    test_vep_somatic_gx.vcf.gz test_tumor HLA-B*15:01 NetMHCpan MHCflurry NetMHCIIpan ./

The error is:

Combining Parsed Prediction Files
Completed
Creating aggregated report
usage: pvacseq [-h]
               {run,binding_filter,coverage_filter,transcript_support_level_filter,top_score_filter,net_chop,netmhc_stab,calculate_reference_proteome_similarity,generate_protein_fasta,generate_aggregated_report,download_example_data,install_vep_plu
               ...

positional arguments:
  {run,binding_filter,coverage_filter,transcript_support_level_filter,top_score_filter,net_chop,netmhc_stab,calculate_reference_proteome_similarity,generate_protein_fasta,generate_aggregated_report,download_example_data,install_vep_plugin,valid_all
    run                 Run the pVACseq pipeline.
    binding_filter      Filter variants processed by IEDB by binding score.
    coverage_filter     Filter variants processed by IEDB by coverage, vaf,
                        and gene expression.
    transcript_support_level_filter
                        Filter variants processed by IEDB by transcript
                        support level.
    top_score_filter    Pick the best neoepitope for each variant.
    net_chop            Run NetChop on existing pVACseq output .tsv to predict
                        cleavage sites on the neoepitopes.
    netmhc_stab         Run NetMHCStabPan on existing pVACseq output .tsv to
                        add stability predictions to the neoepitopes.
    calculate_reference_proteome_similarity
                        Blast peptides against the reference proteome on
                        existing pVACseq output .tsv.
    generate_protein_fasta
                        Generate an annotated fasta file from a VCF with
                        protein sequences of mutations and matching wildtypes.
    generate_aggregated_report
                        Generate an aggregated report from a pVACseq
                        .all_epitopes.tsv report file.
    download_example_data
                        Download example input and output files.
    install_vep_plugin  Install the Wildtype and Frameshift VEP plugins into
                        your VEP_plugins directory.
    valid_alleles       Show a list of valid allele names.
    allele_specific_cutoffs
                        Show the allele specific cutoffs.

optional arguments:
  -h, --help            show this help message and exit
Error: No command specified

I'm running pVACseq 3.0.0 with python 3.8. I also see tons of pandas warning about deprecation:

/opt/conda/lib/python3.8/site-packages/pvactools/lib/prediction_class.py:350: FutureWarning: The frame.append method is deprecated and will be removed from pandas in a future version. Use pandas.concat instead.
  results = results.append(epitope_df)

[susannasiebert]

Hi Dietmar, Thank you for this error report. I believe that this issue is related to using a newer version of pandas. We've been using pandas 0.25.2 as per our requirements.txt file but I was able to reproduce this issue with pandas 1.4.1. I've made a bugfix (#780) to resolve this compatibility issue. As a stop gap until we release a hotfix, I believe you should be able to get your run to work by downgrading your pandas version to 0.25.2.

[riederd]

Hi Susanna, thanks so much for the quick reply and fix. I can also confirm, that downgrading to pandas 0.25.2 works.