v2.0.0

What's New

Major Changes

• Base functions have been rewritten in Python and now follow a standardized format: bin/vaper_*.py.
• The CDC IRMA assembler is no longer supported due to suspected issues with iterative reference refinement when multiple species or subtypes are present, which could lead to assembly errors.
• The fast reference selection mode has been removed. This method was not found to be faster or more accurate than the default accurate mode.
  • The former accurate mode is now referred to as standard. The accurate mode name will still work for backward compatibility.

New Modes and Behavior

• Added support for:
  • Consensus modes: standard, mixed
  • Reference selection modes: standard, sensitive, kitchen-sink
• Reference selection is now more sensitive and reproducible due to:
  • Deterministic processing at final selection steps
  • Improved reference refinement using Sourmash and DBSCAN
• Read alignment is now performed against all references simultaneously, improving efficiency and reducing spurious alignments, especially for closely related species.

Reference and Database Updates

• Introduced a new reference set containing all species targeted by the Illumina VSP v2.0 panel, plus hundreds of additional species (generated using EPITOME v2.1).
• Reference sets are now accepted in CSV or JSONL format (JSONL is standard).
• Reference selection supports include and exclude filtering via the samplesheet or command-line interface.
• Updated viral Sourmash database used for metagenomic analysis. This now requires git large file storage if cloning the repo to access the DB files. These files should still be staged automatically when the workflow is run.

Assembly and Output Improvements

• FASTQ files are no longer required to be gzip-compressed issue #6.
• Assemblies can optionally be output in a "tidy" format that:
  • Removes terminal N characters
  • Excludes mixed IUPAC ambiguity codes
  • Is suitable for GenBank submissions
• A detailed JSON summary is now generated in addition to the standard CSV summary.

Privacy and Compliance

• Reads can optionally be scrubbed of human sequences, facilitating submission to public repositories such as SRA.

Parameter Changes

• The default allele ratio threshold used by iVar has been changed from 0.6 to 0.
  • This reflects the true behavior of the parameter, despite current documentation.
  • See: consensus with ambiguities · Issue #51 · andersen-lab/ivar

New documentation page available at https://doh-jdj0303.github.io/vaper-docs/docs/v2.0/