Pyscenic

Latest version: v0.12.1

Safety actively analyzes 685838 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 8 of 8

0.8.3

- Faster implementation of the sole remaining speed bottleneck i.e. module_from_adjacencies: from 23min to less than 3 minutes on benchmark.

0.8.2

BugFix: AUCell - When there is a complete mismatch between a gene signature/regulon and the genes in the expression matrix, AUCell does not abort anymore with an assertion error but warns the end-user and continues with calculations for the other supplied regulons.

0.8.1

- Several optimisations for computing on clusters using dask.distributed.
- Installation: version of pandas should be at least 0.20.1 (df2regulons uses groupby with an index column) - this dependency is enforced.

0.8.0

- __Easier and more robust Jupyter notebook API__:
- Removed nomenclature attribute from all functions.
- Changed name of parameter num_cores to num_workers for aucell function to make it more consistent with pruning for cis-regulatory footprints (prune function).
- In modules_from_adjacencies: the expression matrix is always converted to floating point numbers. This requirement might be violated when dealing with raw counts as input.
- In modules_from_adjacencies: removing duplicate genes in the expression matrix to avoid errors when looking up correlations between genes.
- __Better default values__:
- Adjusted default setting for threshold based modules: now percentile based instead of based on an absolute threshold. 75th and 90th percentiles are the new defaults.
- Masking of dropouts for calculation of Pearson correlation between a TF and its target genes based on expression levels across cells is the new default.
- __BugFixes__:
- Incorrect validation of IP-address when using dask distributed scheduler.
- AUC calculation based on weighted recovery without weighted recovery being used for target gene selection.

0.7.0

- Support for Drosophila melanogaster.
- Experimental - Support for region-based databases: instead of ranking genes based on the score of a motif we rank candidate regulatory regions (i.e. enhancers) and map genes to their putative regulatory regions. Regulons hereby gain enhancer-resolution.
- Experimental - Support for loom file format export.

Page 8 of 8

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.