v0.6.0

New features

• Simulate RNA-DNA data from first principles with bardic simulate.
• Specify uniform background (= no modeling chromatin heterogeneity) with bardic run -bt uniform.

Other

• More docstrings.

Full Changelog: v.0.5.0...v0.6.0

v.0.5.0

New features

• CLI arguments can be specified in a file. Check docs for an example.
• Added an option to not estimate scaling for all RNAs (all scaling factors will be equal 1).
• q-values are now estimated both globally and within each RNA individually. User can specify which q-value type to use to control the FDR.
• Due to changes in q-values, the version of the Rdc schema has been updated to 1.1.
• In selection.tsv, numbers of zero bins are added.

Bug fixes

• Fixes for scaling of RNAs with zero cis contacts.

Improvements

• Performance of multiprocessing is enhanced due to scheduling jobs in chunks (modification is possible from API side). The size of a chunk is internally adjusted to balance between the number of available cores and pending jobs.

Breaking changes

• Now only version 1.1 of Rdc will be created. Version 1 can be read, but there might be potential problems with rerunning some parts of the pipeline.

Other

• Added code for filtering of "bad" RNAs and contacts, the command will be introduced in the future.

Full Changelog: v0.4.1...v.0.5.0

v0.4.1

Bug fixes

• Packaging process is fixed to allow for clean installations.

Breaking changes

• Most methods and functions are renamed for better semantics and are split into private and public ones. Closer to public API!
• Removed legacy folder.
• Removed scripts folder.

Full Changelog: v0.4.0...v0.4.1

v0.4.0

In this release:

• CLI
• Better code quality
• Docs
• Ability to call peaks for one-to-all data

New features

• Finally, CLI! With pretty help messages due to custom formatter class. Run bardic -h to see for yourself.
• New util run that allows to run the whole pipeline with the single command. Also, run allows for supplying a custom background track (bedGraph with evenly sized bins).
• Versioning of dnah5 and rdc file schemas via version attribute. Current version is "1".
• Docs: for basic running of the pipeline and description of rdc and dnah5 schemas.

Bug fixes

Improvements

• Enforced checking of Rdc/rdc and DnaDataset/dnah5 status attributes in utils. E.g., there will be no creation of rdc file if are_binsizes_selected in dnah5 is False.
• Removed reduntant peak estimatation once their p-values and q-values are estimated (i.e. if Rdc.are_peaks_estimated is True).

Breaking changes

• Changed code and namings of status attributes for DnaDataset/dnah5 (binsizes_selected -> are_binsizes_selected) and Rdc/rdc (scaling_fitted -> is_scaling_fitted, peaks_estimated -> are_peaks_estimated).
• Minor API changes for better code quality.
• dnah5 and rdc files generated with previous versions of BaRDIC won't be open with the new one due to added version attribute.

Other

• Pipeline runs even with single RNA, so calling peaks for one-to-all data is possible now! Please ensure that the background track is a bedGraph with evenly sized bins. Parameters need to be tuned by the user, specifically for bin size selection and q-value threshold. The quality of peaks wasn't checked, though.

v0.3.0

Welcome to the first public release of BaRDIC!

New features

• Finally, nice API that covers all features from scripts.

Bug fixes

• Fixed a bug in spline refinement procedure.

Improvements

• Took HDF5 to store DNA parts and RDC (contact pixels) -> pixels are calculated only once -> increased performance.
• Parallel HDF5 reading allowed to ditch ray. Switched to tqdm.contrib.concurrent.process_map for multiprocessing -> no need for tweaking shared storage, API is simpler.
• Type hints are introduced, code is checked with mypy.

Breaking changes

• Scripts in scripts will fail. Will be removed in the future.
• Legacy in legacy will fail, too. Are kept here for the sake of interoperability. Ideally will be removed in the future.

Other