NINJA-OPS v1.5.1

• ninja.py: Now automatically detects number of threads to use for aligner (thanks bhillmann!)
• ninja_parse_filtered: Rewrote parser for low memory use and slightly better I/O performance
• Databases: Added SILVA v123 ribosomal database
• Databases: Moved databases to external site, currently http://ninja-ops.ninja/db to dramatically reduce zip download size

NINJA-OPS v1.5.0

• New OTU compaction options (now default). Now breaks ties between equally good reference matches deterministically instead of at random (pools all equally good matches into single OTU id). Please consider carefully whether this behavior is desired for your statistical analyses. (Use -K to suppress OTU compaction)
• ninja.py: Taxonomy is now optional; the wrapper can detect whether the taxonomy is present in a database folder
• ninja.py: Default speed option is now -m max (more sensitive)
• ninja_prep: Now takes a prefix to name all output files.
• ninja_prep: Now generates unspaced concatesomes by default
• ninja_prep: Now generates compressed concatesome (compressasomes, TCF format) by default
• ninja_compact: New program. Takes samfile produced by aligner and coalesces OTUs (deterministic tie-breaking) losslessly.
• Databases: regenerated, UNITE database now is spaced, TCF file added to all databases for use with ninja_compact

NINJA-OPS v1.4.1

• Documentation: Fix outdated instructions for taxonomy sorting, similarity as %, etc
• ninja.py: New option to enable both forward and reverse strand matching -z (--both_strands)
• ninja_filter: Removes extraneous carriage returns from input sequence headers
• ninja_parse: Adds newline to end of otu_map
• Databases: New SILVA r119 database

Thanks to Jenya Kopylov (ekopylova) for the database pull request (and for other suggestions/reports in 1.4.x)

NINJA-OPS v1.4.0

• ninja_prep: now builds databases from a FASTA with string headers, line splits, comments, blank space, etc
• ninja_prep: removed unnecessary variable region trimming and sorts
• ninja_prep: now inserts spacers flanking marker genes; disable with --no-spacers
  • spacers allow for better chimera handling if ambigs penalized, or better overhang support if not (ITS; UNITE)
• ninja_filter: new compilation directives controlling k-mer denoising stringency
• ninja_filter: report what line a failure occurs on during read-in
• ninja_filter: removed some extraneous tree functions
• ninja_parse_filtered: now outputs QIIME OTU map file in LOG mode (-F full output in wrapper)
• ninja_parse_filtered: no longer requires sorted taxonomy
• ninja_parse_filtered: no longer requires all taxonomic labels to match references
• ninja_parse_filtered: newline after legacy table, BIOM 1.0 format now outputs fixed-digit date
• Rewrote database format
• Cosmetic tweaks for consistency and readability
• UNITE 30-1-16 DB now included
• SILVA V4 database removed temporarily as new version is prepared (target: v1.4.1)
• Re-compiled Greengenes DB
• Reworked directory structure and updated various documentation

NINJA-OPS v1.3.2

• Fixed a bug in -m max (wrapper only)
• New feature: auto-detect whether input fasta has as many samples as sequences; if it does, treat as one sample and warn user
• New feature: LOG mode in parser now outputs what all matched sequences aligned to (ninja_pass.log). In wrapper, simply use -F to enable full output.

NINJA-OPS v1.3.1

• Explicit paired-end support (in wrapper; optional distance parameter)
• Filtered file shows reasons for filtering
• Alternative ambiguous base handling schemes (discard or convert A)
• New commandline options and documentation
• Fix Mac crash in parser in certain situations
• Streamlined BIOM 1.0 format
• Re-added relative path support if binaries aren't in PATH
• Python script tweaks for handling drop-in bowtie2
• Renamed Linux binaries to be compatible with script calls

NINJA-OPS 1.2

• removed relative paths
• ninja executable is now in the bin directory
• cleaned up python code
• now properly prints errors from spawned subprocesses
• writes to both stdout and log
• precompiled linux binary includes static libraries

NINJA-OPS 1.1

• Much faster support for -F full output mode.
• Now outputs a list of all IDs of sequences that failed to pass the filter step, and a separate list of all IDs of sequences that failed to match the reference database.
• Added support for legacy output.