Cuttlefish 2.2.0

New Features

• Introduces option --poly-N-stretch to cuttlefish 1 usage, specifying which would include information of poly-N stretches in the tiling output of the GFA-reduced format.

Enhancements

• Can now build and run on M1 and M2 machines (thanks to @theJasonFan).

Full Changelog: v2.1.1...v2.2.0

Cuttlefish 2.1.1

Fixes

• A bug in outputting the compacted graph in the ''reduced'' GFA format has been fixed. The bug would freeze the process at completion of the GFA-reduced output, when the input contains sequences with length >= k but with no unitig.

Full Changelog: v2.1.0...v2.1.1

Cuttlefish 2.1.0

New features

• Introduces option --track-short-seqs to cuttlefish 1 usage, specifying which would make it track the existence of sequences shorter than k bases, in the output .json file.

Enhancements

• Upgraded the bit-vector library to v0.1.1.
• Using std::filesystem moving forward.

Full Changelog: v2.0.0...v2.1.0

Cuttlefish 2.0.0

New features

• Cuttlefish is now able to construct compacted de Bruijn graphs from short read sets, along with reference sequences. This is made possible through the design of a new algorithm, Cuttlefish 2, that works on both forms of input. An associated pre-print describing this new algorithm is available on bioRxiv.

Enhancements

The earlier Cuttlefish algorithm (also referred to as Cuttlefish 1) implementation has been enhanced in a number of ways:

• An explicitly built KMC-database is not required as an input to Cuttlefish anymore. The database construction has been incorporated into the Cuttlefish execution.
• The plain-text output has been replaced with the FASTA format.
• A meta-information file is also output along with the compacted graph, containing summary statistics of the graph.
• The memory-usage can now be traded-off for faster execution time, through providing a (soft) memory-bound, or lifting off any strict-memory requirements.

All these features are also available under execution of the Cuttlefish 2 algorithm as well.

Fixes

• A bug in outputting the compacted graph in the GFA 1 format has been fixed (reported in #8, #9). The bug would produce an extra pre-pending overlap value in the ''Overlaps'' field of a ''Path'' line.

Other changes

• The default k value is bumped to 27 from 25.
• The default thread-count t is set to a quarter of the number of concurrent threads supported, instead of just 1.

Full Changelog: v1.0.0...v2.0.0

Cuttlefish 1.0.0

First release of the associated software implementation for the algorithm proposed in Cuttlefish: Fast, parallel, and low-memory compaction of de Bruijn graphs from large-scale genome collections.

Publication: https://doi.org/10.1093/bioinformatics/btab309