Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

2.1.6

Bugfixes

• Fixed op.translate (and every op.resource call for non-human organisms) failing with HTTP 404 after HGNC dropped the genenames/ subtree from the EBI FTP mirror; the HCOP fifteen-column files are now fetched from the new HGNC Google Cloud Storage bucket (#303)

• Removed a redundant pd.read_csv call in op.translate that re-downloaded the HCOP file a second time and bypassed the retry logic in _download

• Fixed bm.benchmark ground-truth matrix row order not matching score matrix after pivot, which silently misaligned perturbation labels when obs names sort differently from their original order (#302)

2.1.5

Added

• pl.volcano now accepts a gene name (str) or list of gene names (list[str]) for the top parameter to annotate specific features on volcano plots

Changes

• Refactored ds.ensmbl_to_symbol to reuse _download and fixed mirror fallback to actually switch between Ensembl mirrors

• Made tqdm progress bar compact in _download

Bugfixes

• Fixed pl.dotplot ignoring vcenter=0 due to falsy check (#293)

• Fixed _log setting the root logging level to INFO for all packages (#296)

2.1.4

Changes

• Added alternative argument to decoupler.mt.query_set. By default 'greater', before it was 'two-sided'

• Unpinned scipy version limit

Bugfixes

• Fixed missing progressbar for decoupler._download._download

• Added missing decoupler.mt.query_set documentation

2.1.3

Changes

• pp.adjmat now returns the same features as used as input instead of the subset of net

• pp.pseudobulk now returns the same order features as used as input instead of shuffling them

• Added a dedicated header and 5 attemps to _download to mitigate 429 Client Error from Zenodo downloads

2.1.2

Added

• pp.query_set to test overlap between a given feature set against a database of sets

Changes

• tl.rankby_obsm now accepts AnnData.obs column names specified in the obs_keys argument

• Most plotting functions now accept extra arguments through kw_ arguments.

• p-values now are corrected using a custom numba-optimized version of scipy.stats.false_discovery_control called _fdr_bh_axis1_numba

2.1.1

Added

• mypy checks in CI

• notebook checks in CI

Bugfixes

• Fixed error in decoupler.pp.pseudobulk when adata.obs_names were not unique, now throws verbose error

• Fixed corner case in decoupler.mt.gsea when p-values were infitite and could not be corrected

2.1.0

Changes

• Updated logo

• decoupler._download._download now returns bytes instead of a dataframe. To transform to pandas.DataFrame use decoupler._download._bytes_to_pandas

• Enrichment methods and pseudobulking now work with backed AnnData objects, useful when working with big datasets and memory is limited

Bugfixes

• Fixed error in pl.obsm where default value of cmap_obs was not properly set.

2.0.7

Changes

• Added pre-commit functionality to the repository

2.0.6

Changes

• Modified links and paths to follow scverse’s repository

2.0.5

Bugfixes

• Fixed error message when extra dependencies where not installed

2.0.4

Bugfixes

• Fixed dcor import bug as an external dependency

2.0.3

Bugfixes

• Fixed error in pp.pseudobulk when obs columns were not categorical

2.0.2

Changes

• Allowed ordering functions (pp.bin_order, pl.order and pl.order_targets) not to be bound between 0 and 1

• Added ipywidgets as dependency

• Silenced xgboost warnings

Bugfixes

• Handled corner case in bm.metric.auc when scores are all 0

• Fixed error in bm.metric.hmean when metrics were str instead of list

• Fixed error when obs column is a list in pp.pseudobulk

2.0.1

Bugfixes

• Fixed an error in pp.pseudobulk when handling empty samples or features

2.0.0

Major update to accomodate the scverse template [VBH+23].

All functions have been rewritten to follow the new API, errors when running previous versions (1.X.X) are expected if decoupler >= 2.0.0 is installed.

Changes

• Methods are now in the mt module and are built from shared class Method

  • Use decoupler.mt.<method_name> to call a method

  • min_n argument has been renamed tmin

  • New argument bsize allows to run a method with batches in case excessive memory usage is an issue

  • $p_{values}$ of the enrichment scores are now corrected by Benjamini-Hochberg

  • mdt and udt are now based on xgboost instead of sklearn for better scalability. udt statistic is now the coefficient of determination $R^2$ instead of the importance of a single decision tree.

  • mlm and ulm now include a tval parameter, which allows returning either the t-value of the slope or the slope itself as the enrichment statistic

  • ora now returns the odds ratio of the contingency table as a statistic, and computes a two-sided Fisher exact test instead of a one-sided one

  • viper now correctly estimates shadow regulons when network weights are values other than -1 or +1

  • wsum and wmean are deprecated, instead now the method waggr allows to run both methods and any custom function. This makes it easier to quickly test new enrichment methods without having to deal with decoupler’s implementation

• Databases from Omnipath can now be accessed through the new op module

  • Use decoupler.op.<resource_name> to access a database

  • Removed the omnipath package as a dependancy

  • Fixed collectri to the publication version instead of the OmniPath one

  • Made progeny only return significant genes by default instead of the top N genes per pathway

• Plots are now in a new pl module

  • Use decoupler.pl.<plot_name> to call a plot

  • They use a common class Plotter to make it easier to share arguments between them

  • plot_violins has been deprecated

  • Names that have changed

    • plot_psbulk_samples to filter_samples

    • plot_running_score to leading_edge

    • plot_associations to obsm

    • plot_targets to source_targets

• Preprocessing functions are now in the new pp module

  • Renamed check_corr to net_corr, now also returns adjusted $p_{values}$

  • Renamed get_acts to get_obsm

  • Renamed get_pseudobulk to pseudobulk. Now it does not automatically remove low quality samples, this is now done with the function filter_samples

  • Deprecated get_contrast, get_top_targets and format_contrast_results. PyDESeq2 should be used instead

  • Moved rank_sources_groups to the new tl module as rankby_group

  • Moved get_metadata_associations to the new tl module as rankby_obsm

• Moved the benchmarking pipeline inside a new bm module, with its metrics and plotting functions in further submodules (bm.metric and bm.pl)

Added

• ds module with functions to download several datasets at different resolutions

  • Bulk: hsctgfb and knocktf

  • Single-cell: pbmc3k, covid5k and erygast1k

  • Spatial: msvisium

  • Toy data: toy and toy_bench

  • Utils: ensmbl_to_symbol

• New database functions in the op module

  • Added hallmark as a custom resource

• New plotting funcitons in the pl module

  • Added obsbar to plot size of metadata columns in anndata.AnnData.obs

  • Added order to plot sources or features along a continous process such as a trajectory

  • Added order_targets to plot the targets of a given source along a continous process

• New preprocessing functions in the pp module

  • Added two functions to format networks, adjmat to return an adjacency matrix, and idxmax to return a list of sets

  • Added filter_samples to filter pseudobulk profiles after running pseudobulk

  • Added knn to calculate K-Nearest Neighbors similarities based on spatial distances

  • Added bin_order to bin features across a continous process

• tl module with functions to perform statistical tests

  • Added rankby_order to test for non-linear associations of features with a continous process

• New benchmarking metrics and plotting related functions in the bm module

  • Added two more metrics, F-score and qrank

  • Added shared plots for metrics, bm.pl.auc, bm.pl.fscore and bm.pl.qrank

  • Added a summary plot across metrics bm.pl.summary