6.0.3

What's Changed

• Remove allele no longer supported by MHCflurry. by @susannasiebert in #1342

Full Changelog: v6.0.2...v6.0.3

6.0.2

What's Changed

This is a bugfix release. It fixes the following problem(s):

• Define transcript set in the pVACsplice top score filter on Junction, not Index, in order to be consistent with the aggregate report. by @susannasiebert in #1340
• Fix some bugs with the pvacseq create_peptide_ordering_form command by using the Index to construct the 51mer ID instead of modifying the AA change. by @susannasiebert in #1341

Full Changelog: v6.0.1...v6.0.2

6.0.1

What's Changed

This is a bugfix release. It fixes the following problem(s):

• Ensure that the top_score_metric2 parameter is passed through to UpdateTiers so that the aggregate report gets sorted correctly. by @susannasiebert in #1335

Full Changelog: v6.0.0...v6.0.1

6.0.0

What's Changed

This is a major version release. Please note that pVACtools 6.0 is not guaranteed to be backwards-compatible and certain changes could break old workflows. It also removes support for Python 3.7 and 3.8.

New Tools

• A new tool, pVACcompare allows users to visualize differences between output files from different pVACtools versions. This tool may be useful when investigating changes to predicted neoantigens between versions in controlled experiments where enumerating and explaining such differences may be crucial, e.g. in clinical trials. The tool can be run using the pvactools compare command. For more information please see the Comparison Tool documentation.
• A new standalone command pvacseq create_peptide_ordering_form generates peptide ordering files (FASTA, annotated ordering Excel spreadsheet, and review template Excel spreadsheet) to streamline preparation of peptides for synthesis and review. For more information please see the documentation

New Features

• pVACseq and pVACsplice now take into account the MANE Select and Canonical status of a transcript for filtering, prioritizing, and tiering a neoantigen candidate. MANE Select, Canonical, and TSL status is evaluated according to the new --transcript-prioritization-strategy parameter. This parameter allows users to list one or more criteria (mane_select, canonical, tsl) to take into consideration. As part of this update the “transcript support level filter” has been renamed to “transcript filter” including the standalone command which is now pvacseq|pvacsplice transcript_filter instead of pvacseq|pvacsplice transcript_support_level_filter.
• The aggregate report tiers have been updated to add four new tier:
  • PoorBinder: candidate fails the binding criteria but passes all other critieria.
  • RefMatch: candidate has a reference match but passes all other criteria.
  • ProbPos: candidate has a problematic amino acid but passes all other criteria.
  • PoorTranscript: a candidate’s best transcript is neither MANE Select, Canonical, nor meets the TSL cutoff (depending on the specified --transcript-prioritization-strategy; only available in pVACseq and pVACsplice).
• By default, transcripts where the FLAGS VEP annotation is set will now be filtered out before processing by pVACseq and pVACsplice. This field identifies transcripts where the CDS 5’ or 3’ is incomplete. A new parameter --allow-incomplete-transcripts has been added which can be used to replicate the previous behavior where such transcripts were included.
• The pVACsplice logic for aggregating variants in the aggregate report creation has been updated to aggregate neoantigen candidates from the same Junction together instead of using the Index.
• Output file names of the reports have been updated to include MHC_I/MHC_II/Combined prefixes for easier identification of the type of output file.

Bugfixes

• This release fixes a bug with sorting of pVACfuse output files.

Full Changelog: v5.5.4...v6.0.0

5.5.4

What's Changed

This is a bugfix release. It fixes the following problem(s):

• Fix an issue where the output parsing of MHCflurryEL in pVACfuse and pVACsplice where only the smallest epitope length would have results. by @susannasiebert in #1320
• Fix per-allele counts in the aggregated report. by @susannasiebert in #1316
• Fix a few bugs where standalone commands weren’t working with all_epitopes/filtered TSVs. by @susannasiebert in #1324
• Update --top-score-metric2 parameter help text for clarity and remove unnecessary usage. by @susannasiebert in #1321

Full Changelog: v5.5.3...v5.5.4

v5.5.3

What's Changed

This is a bugfix release. It fixes the following problem(s):

• Fix a bug in the reference proteome similarity step that may result in an IndexError: string index out of range error. by @susannasiebert in #1313
• Improve handling of inframe indels in repetitive region to upstream filter out variants that don’t result in novel neoantigen candidates and to better find a matched wildtype epitope. by @susannasiebert in #1314

Full Changelog: v5.5.2...v5.5.3

5.5.2

What's Changed

This is a bugfix release. It fixes the following problem(s):

• Remove unused --top-score-metric2 parameter from pVACvector. by @ldhtnp in #1304
• Add --biotypes parameter to pVACvector when running with an input VCF. by @ldhtnp in #1297
• Speedup processing of AGFusion exon files. by @susannasiebert in #1301

Full Changelog: v5.5.1...v5.5.2

5.5.1

What's Changed

This is a bugfix release. It fixes the following problem(s):

• Fix a couple of issues with the new --top-score-metric2 by @susannasiebert in #1291

• When adding the --top-score-metric2 option, the logic for determining the included candidates during aggregate report creation was amended to compare either the IC50 or percentile to the aggregate inclusion binding threshold. This logic should not have been changed and instead only the IC50 should be compared to the aggregate inclusion binding threshold, no matter which --top-score-metric2 was selected. This specific change has been reverted

• In order to achieve deterministic results when using the percentile --top-score-metric2 option, a peptides.sort() call was used. This release replaces this with a better way of finding the best peptide by using the IC50 as a secondary sort criteria.

Full Changelog: v5.5.0...v5.5.1

5.5.0

What's Changed

This is a minor feature release. It adds the following features:

• Add a new parameter --top-score-metric2 that allows users to set whether to prioritize ic50 or percentile scores when creating the filtered and aggregated reports. by @Jaz2021 in #1267
• If an unsupported amino acid is the last/first amino acid in the peptide sequence, clip it so that the sequence can still be included in the run. by @susannasiebert in #1266

It also fixes the following pVACsplice bugs:

• Remove option for DA and N anchors in pVACsplice. Splice sites with the DA anchor (known donor and acceptor) don’t result in any novel epitopes (the splice site peptide sequence predicted by pVACsplice is identical to the WT sequence). For the N anchor, there isn’t enough information to construct a splice site peptide sequence with our current approach since both the donor and acceptor are not observed in the reference. by @susannasiebert in #1283
• Add a missing import statement and fix an issue in get_mutated_peptide_with_flanking_sequence where, for certain variants, the mutated peptide was not being determined correctly. by @susannasiebert in #1282
• Fix an issue in the pVACsplice calculate reference proteome similarity step where, for certain variants, the start position of the peptide to query was not being determined correctly. by @susannasiebert in #1278
• Add better handling of pVACsplice frameshifts where the downstream sequence may not have been fully included in the query sequence for the calculate reference proteome similarity step. by @susannasiebert in #1285
• Handle cases in pVACsplice where no processable neoepitopes were found so that the run exits gracefully with better stdout messages. by @susannasiebert in #1286
• Fix documentation of the input fasta parameter in pvacsplice calculate_reference_proteome_similarity to point to the correct location. by @susannasiebert in #1288

New Contributors

• @Jaz2021 made their first contribution in #1267

Full Changelog: v5.4.3...v5.5.0

5.4.3

What's Changed

This is a bugfix release. It fixes the following problem(s):

• Fixed IC50/%ile WT sorting to use numerical sorting instead of string sorting. by @ldhtnp in #1265
• Bugfixes for the pvacsplice generate_protein_fasta command to handle sequences shorter than the flanking sequence length and sequences with unsupported amino acids. by @susannasiebert in #1268

Full Changelog: v5.4.2...v5.4.3