Version 1.0

This is an important release with the following changes:

Results

• Compared with v0.2, results have changed in some cases.
• Various bugs have been fixed that slightly affected results, improving the results.
• Alignment is now performed with a band by default; in rare cases this might degrade results. This may be changed/deactivated via the --band option.

Performance

• Significant performance improvements; speed-ups of 3x to more than 10x (depends on data and options).
• To achieve full speed, you need to build with machine-specific optimisations (see build instructions).

Features

• Dynamically adapt the window size to the length of the current read (via --dyn-w-size).
• Automatically cache all input files in a local folder to reduce load on remote resources (via --cache-data-in-tmp).

Command-line interface

• The option --bam2 has been removed; instead just give a list of the combined files as single argument.
• The options that influence scoring in alignment have been deprecated. They are no longer visible on the command-line and will be removed in a future update.
• Please change all scripts to use the long form of the options (those that begin with --), because a future update will remove/change the short-forms.
• The option --dyn-w-size was added (see above).
• The option --cache-data-in-tmp was added (see above).

Under-the-hood

• Extensive refactoring, modernisation and clean-up of code.
• Continuous Integration via Github Actions.
• Code-formatting via clang-format.

Version 0.2

v0.2

v0.2

Version 0.1.2

v0.1.2

Disallowed hard clipped reads

Version 0.1.1

v0.1.1

Fixed valgrind memory error

Version 0.1

Initial release.