Release 0.14.4

[0.14.4] - 2020-04-02

Fixed

• Fixed "TypeError: an integer is required (got type str)" when creating GC index.
• Fixed combine_peaks with Ensembl chromosome names (thanks @JGASmits).
• Fixed bug with pandas>=1.0.

Release 0.14.3

Fixed

• Fixed 'AttributeError: can't delete attribute' in gimme maelstrom and gimme motifs (#108, #109).

Release 0.14.2

[0.14.2] - 2020-01-31

Bugfix release

Added

• The combine_peaks script now supports .narrowPeak files.

Removed

• Removed seqlogo dependency.
• Removed obsolete code.

Changed

• Refactored the tools section.

Fixed

• Configuration issue with size instead of width (#103).
• Updated tqdm requirement (#98).

Version 0.14.1

Release 0.14.1

Version 0.14.0

Changelog

Added

• The gimme motifs command supports new de novo motif prediction tools: DREME, ProSampler, YAMDA, DiNAMO and RPMCMC.
• A set of non-redundant motifs is selected using recursive feature elimination for gimme motifs.
• Motif scan results for non-redundant motifs are now included in the output.
• Plot motif logos using different styles of visualization (information content, frequency, energy or Ensembl).
• New scoring scheme that uses a z-score based on genomic sequences with a similar GC%.
• CIS-BP motif database version 2.0 (Lambert et al. 2019).
• JASPAR 2020 motif database.

Changed

• Only three motif prediction tools are selected by default for gimme motifs: BioProspector, Homer and MEME.
• The xl motif setting (motif width 6-20) is selected by default for gimme motifs.
• Output names for files and reports of gimme motifs are now more consistent.
• Command line tools gimme roc has been removed. This functionality has now been merged with gimme motifs. The gimme motifs command now scans for both known and de novo motifs.
• Replaced weblogo/seqlogo with logomaker.
• GC% background now uses a genomic index of GC% frequencies.
• GC% z-score is now the default score for all gimme tools.

Fixed

• FASTA files with > symbols in the header are now correctly parsed.
• Fixed GC% background. The GC% background is now correct for all input formats.
• factorial import for scipy >= 1.3.0.
• All output columns in HTML reports can now be sorted correctly.

Version 0.13.1

Added

• Improved docstrings of several modules.
• Added new API examples.

Fixed

• The MEME motif tools should now be recognized after install.
• Output of MEME 5.0.2 is now parsed correctly.
• If the inputfile of gimme motifs is not recognized, a clear error message is printed.
• Duplicate factors are removed from the motif factors list.

Changed

• MEME is no longer included with GimmeMotifs. When installing via conda meme will be included. If GimmeMotifs is installed via pip, then MEME needs to be installed separately.
• Changed "user" background to "custom" background.
• Updated Posmo to run with a wider variety of settings.

Version 0.13.0

Release 0.13.0 of GimmeMotifs.

[0.13.0] - 2018-11-19

Added

• Multiple other motif databases (JASPAR, HOMER, HOCOMOCO, CIS-BP, ENCODE,
  Factorbook, SwissRegulon, IMAGE).
• Helper script to combine peaks (summit files from MACS2)
• Helper script to create coverage table (similar to bedtools multicov)
• Option to report z-score normalized motif scores.
• Added precision-recall AUC to stats and gimme roc.
• gimme motifs now supports narrowPeak input.
• Updated documentation with an explanation of the score that gimme maelstrom reports.

Changed

• The maelstrom tools now use z-score normalized motif scores.
• Improved efficiency of motif scanning (>10X speed improvement).
• Removed dependency on R for rank aggregation.
• Dropped support for Python 2.
• Use versioneer for versioning.
• Removed the default genome in config file.
• Config file is now independent from GimmeMotifs version and will be created by
  default on first use.
• Simplified setup.py script.
• Updated parameters for ChIPMunk motif finder.

Fixed

• Fixed the seqcor similarity metric to use a non-random sequence and to also take
  the reverse complement of motif 2 into account.
• Improved the speed of gimme roc.
• Fixed memory leak of gimme roc.
• Fixed scale for newer pandas/sklearn combo
• FIxed bug related to backgroundgradient with new pandas

Version 0.12.0

I am happy to announce GimmeMotifs version 0.12.0.

Notable changes

• You can specify the number of threads to use on all command line tools
• GimmeMotifs now uses the faidx index, which means that gimme index is no longer necessary. This also means that you might have to reinstall your genomes.

See further changes here: CHANGELOG.md.

Version 0.11.1

Bugfix release

• Fixed issue with gimme genome
• Fixed issue with gimme cluster
• Fixed bug in gimme roc where html report would crash when no significant motif were found

Version 0.11.0

Changelog:

• Support for Python 3
• Updated documentation
• API for a lot of the functionality
• New maelstrom command for differential motif analysis
• Updated stats and graphical output for gimme roc
• ... and many more small changes