1.1.0

This version adds a host of new features to pVACtools:

• pVACseq is now able to parse VAF, depth, and expression information directly from the VCF. This makes the --additional-input-file-list option obsolete. The --additional-input-file-list option is now deprecated and will be removed in an upcoming release. For more information on how to annotate your VCF with readcount and expression information, see the Input File Preparation page.
• pVACseq is now able to handle proximal germline and somatic variants. In order to incorporate those into the epitope predictions, you will need to provide a phased variants VCF to your pVACseq run using the --phased-proximal-variants-vcf option. For more information on how to create this file, see the Input File Preparation page.
• We added support to pVACseq for filtering on transcript support levels. This requires the input VCF to be annotated with the TSL field by VEP. Be default, any transcripts with a TSL above 1 will be filtered out.
• The binding filter of pVACseq and pVACfuse can now be run with flexible, allele-specific binding-thresholds. This feature can be enabled using the --allele-specific-binding-thresholds flag. The thresholds used are taken from the IEDB recommendations.
• pVACseq now supports a --pass-only flag that will result in any VCF entries with a FILTER to be skipped. Using this flag, only VCF entries with a FILTER of PASS or . will be processed.
• We added support for the MHCflurry and MHCnuggets prediction algorithms. These can be used by listing MHCflurry, MHCnuggetsI (for MHC Class I alleles), and/or MHCnuggetsII (for MHC Class II alleles) as the prediction algorithms in your run commands.
• The default --tdna-vaf and --trna-vaf cutoff values have been updated from 0.4 to 0.25. This is the minimum VAF threshold that an epitope candidate must meet in order to pass the coverage filter.
• We now offer a graphical user interface, pVACviz, to run pVACseq as an alernative to using the command line. pVACviz, can also be used to plot and filter your pVACseq results.

1.0.8

This is a hotfix release. It fixes the following issues:

• The log directories were accidentially included with the pVACseq example data. They are now removed.
• Some users were reporting mixed type warnings for pandas when running pVACseq. We added some options to avoid this warning.

1.0.7

This is a hotfix release. It fixes the following issues:

• VEP82 and higher supports selenocysteine modicfications (amino acid “U”), which is not supported by downstream IEDB prediction algorithms. pVACtools now skips sequences containing this amino acid with a warning.

1.0.6

This is a hotfix release. It fixes the following issues:

• There was a bug in how alternate alleles were resolved when matching VEP consequence fields to an entry which resulted in certain indels to be skipped. This has now been fixed.

1.0.5

This is a hotfix release. It fixes the following issues:

• IEDB changed the format of combinatorial class II alleles to use / as a delimiter instead of -. DP alleles were previously fixed in pull request #85 but this failed to address DQ alleles. This version fixes this oversight.

1.0.4

This is a hotfix release. It fixes the following issues:

• We discovered a couple more cases of mutations involving stop codons that would result in errors. These are amino acid changes (VEP Amino_acids field) for large indels that would span exon boundaries (multiple * in the Amino_acids field), or amino acid changes involving the transcript stop codon (ending in X). These cases are now handled.

1.0.3

This is a hotfix release. It fixes the following issues:

• Stop-gain mutation were previously not handled correctly. If a mutation had a * (stop gain) in the VEP Amino_acids field, pVACseq would throw an error. We now ensure that those cases are handled. pVACseq will also skip stop-gain mutations if the resulting mutant peptide sequence is not novel.
• pVACseq would previously throw an error if multiple mutations resulted in the same consequence. This is now handled by assigning a unique identifier to each mutation.
• We added a better warning messages if the chosen prediction algorithms and alleles MHC classes are mutually exclusive, e.g., if only class I prediction algorithms were chosen with only class II alleles. Previously, pVACseq would simply finish without producing any output or errors.

1.0.2

This is a hotfix release. It fixes the following issues:

• The epitope length used for generating the peptide fasta when running with multiple epitope lengths was incorrect. This would potentially result in including fasta sequences that were shorter than the largest epitope length which would cause an error during calls to IEDB.
• pVACseq would fail with a nondescript error message if the input VCF was not annotated with VEP before running. A more descriptive error message has been added.
• IEDB changed the format of class II IEDB alleles which would cause an error when running with those alleles. pVACtools will now handle transposing the affected alleles into the new format.
• The standalone binding filters had a few bugs that would result in syntax errors during runtime.
• The indexes created for each fusion entry with pVACfuse had the potential to not be unique which would result in parsing errors downstream.
• pVACseq had the potential to use the incorrect VEP allele for positions with multiple alternate alleles which would result in the incorrect CSQ entry getting used for some of those alternate alleles.
• pVACseq would throw an error if the chosen peptide sequence length exceeds the wildtype protein sequence length of a transcript.

1.0.1

This is a hotfix release. It fixes the following issues:

• Additional data, like example data and VEP plugins were not included in the package correctly so the commands to download these files would fail. This has been corrected.
• Class II predictions would fail if the protein sequences used for binding predictions in IEDB were shorter than 15 peptide sequences. This has been fixed.

1.0.0

This is the initial release of pVACtools, a cancer immunotherapy suite consisting of the following tools:

pVACseq

A cancer immunotherapy pipeline for identifying and prioritizing neoantigens from a list of tumor mutations.

pVACfuse

A tool for detecting neoantigens resulting from gene fusions.

pVACvector

A tool designed to aid specifically in the construction of DNA vector-based cancer vaccines.