Mash v2.3

• Improvements to Screen function:
  • Up to 4x speed improvement due to Robin Hood hashing
  • Additional taxonomy-aware taxscreen command, for use with taxmash, due to Florian Breitweiser
• Sketching optimization due to Torsten Seemann
• Manpages due to Fabian Klötzl

Mash v2.2

• screen p-value update
• fix for screen values in bounds

Mash v2.1.1

• screen p-values fixed for amino acid queries
• sketching optimizations

Mash v2.1

Features

• Triangular matrix - A new triangle command computes a lower triangular
  distance matrix in relaxed Phylip format. This streamlines all-pairs distance
  commands and avoids computational redundancy.
• Custom IDs - The ID and Comment fields of a sketch can now
  be set with -I and -C. Only applies to the first sketch for multi-sketch files.
• Read pooling - If multiple input files are given in read mode (-r), e.g.
  paired ends as in mash sketch -r read1.fq read2.fq, they will now pool to the same sketch, avoiding the need for concatenation.

Mash v2.0

Mash's first major version increment focuses on a new top level command, screen, which estimates containment within (rather than distance to) a read set for many sketches simultaneously.

Features

• Screen - A new command that estimates how well sketches are contained
  within a set of reads.
• Hash seed parameter - The seed of the hash function can now be set with -S. Note that if it is changed from the default (42), any sketches created will not work with older versions of Mash (they will appear to old versions to have no sketches, causing an error or empty output).

Fixes

• Fix for large sketch files (Issue #16)
• Fix for multi-threaded sketching (Issue #41)

Mash v1.1.1

Features

• JSON sketch dumps - Sketches can now be converted to text in JSON format
  for interoperability with other tools. Metadata, such as k-mer and hash function
  information, are included with the hashes themselves, which are represented as
  unsigned integers.

Fixes

• Fix for stdin sketch input (Issue #32)

Mash v1.1

Features

• Read sketching
  • Minimum k-mer copy number (-m) for more precise and flexible filtering
    than Bloom filter.
  • Genome size and coverage estimation for improved p-values and optional
    termination at sufficient coverage (-c).
• Parallelism
  • Parallel sketching (-p), if more than one sketch is being created.
  • Parallel distance for all comparisons, not just multiple files.
• Alphabets
  • Amino acid (-a) and arbitrary alphabets (-z).
  • Case sensitivity option (-Z), which allows lowercase masking.
• Information
  • A new bounds command for printing expected accuracy for various
    parameters and distances.
  • K-mer copy number histogram available from mash info (-c) for
    sketches made with this version.
  • Tabular mode (-t) and more header information (-H) for mash info.

Fixes

• Large sketch file crash (Issue #16)
• Large pairwise comparison crash (Issue #18)

Mash v1.0.2

Fix for false k-mer size warnings (issue #17)

Mash v1.0.1

• fix for buffer overflow that causes "stack smashing" error with recent GCC versions (issue #15)

Mash v1.0

First official release