Data

Tools

Indexes

Applications

Experiments

Open Source Science

Recent Releases of unsupervised_analysis

unsupervised_analysis - v3.0.3 - Minor fixes

What's Changed

  • Update environment to avoid bug #72 by @bednarsky in https://github.com/epigen/unsupervised_analysis/pull/73

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v3.0.2...v3.0.3

- Python
Published by sreichl 8 months ago

unsupervised_analysis - v3.0.2 - Minor fixes & improvements

What's Changed

  • Fix stringi version by @bednarsky in https://github.com/epigen/unsupervised_analysis/pull/62
  • Avoid heatmap config default causing error by @bednarsky in https://github.com/epigen/unsupervised_analysis/pull/63
  • Give plotheatmap always enough colors by @bednarsky in https://github.com/epigen/unsupervisedanalysis/pull/64
  • improve-umap-colors-for-only-positive-values by @bednarsky in https://github.com/epigen/unsupervised_analysis/pull/71

New Contributors

  • @bednarsky made their first contribution in https://github.com/epigen/unsupervised_analysis/pull/62

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v3.0.1...v3.0.2

- Python
Published by sreichl 8 months ago

unsupervised_analysis - v3.0.1 - Enable module usage using `github()` directive

  • to enable module usage using github() directive
    • comment global.yaml (now requires full snakemake installation, not minimal)
  • add nodefaults to all env YAML

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v3.0.0...v3.0.1

- Python
Published by sreichl about 1 year ago

unsupervised_analysis - v3.0.0 - Snakemake 8 compatible

Breaking change: Requires Snakemake >= v8.

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v2.0.0...v3.0.0

- Python
Published by sreichl over 1 year ago

unsupervised_analysis - v2.0.0 - Performance improvements

Enhancements and new features

  • PCA: To improve performance n_components and svd_solver can be configured.
  • Heatmap: performance improvements
    • distance matrix calculation done by pdist from scipy and parallelized for observations and features
    • hierarchical clustering using fastcluster
    • observations can be downsampled using configuration n_observations
    • top features can be selected by variability using configuration n_features

The documentation was updated accordingly.

Bug fixes and other performance improvements are not mentioned.

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v1.1.0...v2.0.0

- Python
Published by sreichl over 1 year ago

unsupervised_analysis - v1.1.0 - small enhancements and bug fixes

Enhancements and new features

  • Additional PCA diagnostics: Visualization of the top 10 loadings per principal component using lollipop plots.
  • Internal cluster index calculation optional (very compute intensive).
  • Enable plotting of all features using the keyword "ALL".
  • Enhance Snakemake report using labels.
  • Switch from panels to solo plots.
  • Switch to data.table usage for accelerated read/write in R.

The documentation was updated accordingly.

Bug fixes and performance improvements are not mentioned.

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v1.0.1...v1.1.0

- Python
Published by sreichl over 1 year ago

unsupervised_analysis - v1.0.1 - update author ORCID

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v1.0.0...v1.0.1

- Python
Published by sreichl over 2 years ago

unsupervised_analysis - v1.0.0 - unsupervised analysis now includes cluster analysis methods

enhancements - added a config flag for 2D plot coord_fixed() option

new features - Clustering - Leiden algorithm - Clustification: an ML-based clustering approach that iteratively merges clusters based on misclassification - Clustree analysis and visualization - Cluster Validation - External cluster indices are determined by comparing all clustering results with all categorical metadata - Internal cluster indices are determined for each clustering and [metadataofinterest] - Multiple-criteria decision-making (MCDM) using TOPSIS for ranking clustering results by internal indices - Visualization - all clustering results as 2D and interactive 2D & 3D plots for all available embedings/projections. - external cluster indices as hierarchically clustered heatmaps, aggregated in one panel. - internal cluster indices as one heatmap with clusterings and selected metadata sorted by TOPSIS ranking from top to bottom and split cluster indices split by type (cost/benefit functions to be minimized/maximized).

documentation - add scRNA-seq analysis section to the documentation - update the documentation accordingly (Software, Methods, Features, Examples) - update report to include all new feature outputs - update rulegraph

Bug fixes and performance improvements are not mentioned.

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v0.2.0...v1.0.0

- Python
Published by sreichl over 2 years ago

unsupervised_analysis - v0.2.0 - enhancements, new features and a full example added

enhancements - 2D metadata plots: up to 10 columns per row, coordinates are fixed on both axes, numeric color scheme blue to red with midpoint 0 in grey

new features - 2D feature plots: specify features of interest, which values from the data, will be highlighted in the 2D plots (motivated by bioinformatics highlighting expression levels of marker genes) - densMAP support: local density preserving regularization as an additional dimensionality reduction method - additional PCA diagnostics: - pairs: sequential pair-wise PCs for up to 10 PCs using scatter- and density-plots colored by metadataofinterest - loadings: showing the magnitude and direction of the 10 most influential features for each PC combination - interactive 2D and 3D visualizations (self-contained HTML files) of all projections and embeddings including widgets to color by categorical and numerical metadata, respectively - hierarchically clustered heatmaps of scaled data (z-score) with configured distance metrics and clustering methods (all combinations are computed), and annotated with metadataofinterest

documentation - add a minimal example, using the digits dataset from sklearn, to show configuration, results, and report (.test/ folder) - update the documentation accordingly (Software, Methods, Features, Examples) - update report to include all new feature outputs (apart from interactive plots) - update rulegraph

Bug fixes and performance improvements are not mentioned.

Full Changelog: https://github.com/epigen/unsupervised_analysis/compare/v0.1.0...v0.2.0

- Python
Published by sreichl over 3 years ago