1.12

Minor bug fixes.

• Improved UI for smaller screens (tablets, phones)
• Fixed a bug in the gmt marker genes file that required two tabs to separate annotation from genes instead of one. This caused the first marker gene of each annotation to get lost (gseapy source code: https://github.com/zqfang/GSEApy/blob/8229edcbe629c6309bee540829e96b3da820a515/gseapy/enrichr.py#L89).

1.11

• Added support for .h5 files.
• Allowing spatial information for CODEX to be stored in anndata via x, y coordinates or spatial_idx key.
• Added a missing dependency for seurat_v3.
• Switched the order of selecting highly variable genes and log-transforming the data for flavors other than seurat_v3.

1.10

• Migrated Cellar to dash 2.x and dash-bootstrap-components 1.x. These new versions introduced several breaking changes and the UI now uses a bigger font.
• Allowed for arbitrary directory names when uploading cellranger tar.gz files.
• Switched to conda based bioconductor; now using conda's version of SingleR.
• Updated installation instructions.

1.00

This marks the first release of Cellar. Cellar is an interactive tool for analyzing single-cell omics data and it supports several data types including, but not limited to scRNA-seq, scATAC-seq, CODEX, SNARE-seq, sciRNA-seq, Visium. It supports preprocessing, dimensionality reduction, clustering, DE gene testing, enrichment analysis, cluster and gene visualization modules, projection to spatial tiles, label transfer, and semi-supervised clustering among others.

Cellar is entirely built in Python using the Dash framework. It relies on several open-source packages such as: scanpy, scikit-learn, anndata, pandas, numpy, leidenalg, and more.

Cellar can be used as a standalone app by pulling the docker image at euxhen/cellar or by using our demo web server (for a link see the main page). Full documentation can be found at https://euxhenh.github.io/cellar/