excellxgene 2.0.4

Web application for exploration of large scale scRNA-seq datasets, upgraded to enable end-to-end interactive analysis.

Exploratory CellxGene (ExCellxGene)

This fork implements some of the key features that have been highly requested by the data science team at CZBiohub.

Features include:

• Hotkeys (SHIFT+? to see a tooltip describing all available hotkeys)
• End-to-end interactive analysis and reembedding, with new embeddings hierarchically organized.
• LIDAR graph interaction mode (the airplane) - Show an interactive tooltip describing the cells underneath your cursor. Very helpful for the color impaired or for large datasets with hundreds of labels.
• Sankey plots
• Leiden clustering
• Label fusion and deletion
• Interactive selection of data layer for expression visualization
• Many other quality-of-life improvements.

Patch notes (v1.2.5)

• When displaying continuous metadata, cells with value zero are drawn as if they are unselected to send them to the background.
• Category and geneset menus now have a new menu item to include/exclude zeros from the histograms. This is useful when the distributions are super zero-inflated.

Patch notes (v1.2.3)

• Gene sets are now grouped based on their descriptions under collapsible headers.
• Gene sets are now more compact, displaying 10 genes at a time with buttons to flip through pages.
• Differential expression now calculates the top 100 genes.
• A new button in the menubar allows you to calculate marker genes for all labels in a selected category.
• Embeddings are now indented according to their hierarchical organization, and nested embeddings are collapsible.
• Categorical labels are now sortable based on the currently displayed continuous medatada.
• All preprocessing and reembedding parameters now have a tooltip.
• Added a button to display hotkey menu to the menubar.
• Various bugfixes.

Installation

1. Install miniconda if conda not available already:

wget https://repo.anaconda.com/miniconda/Miniconda3-latest-MacOSX-x86_64.sh -O ~/miniconda.sh
bash ~/miniconda.sh -b -p $HOME/miniconda

2. Create and activate a new environment:

conda create -n cxg python=3.7
conda activate cxg

3. Install excellxgene with pip:

pip install excellxgene

If your operating system is CentOS, then you may run into issues installing dependencies that require up-to-date gcc or g++ compilers. Please install with the following and try reinstalling excellxgene with pip:

conda install -c conda-forge gcc cxx-compiler

4. Download the git repository to get the example datasets (assumes git is available, if not install it with conda install -c anaconda git)

git clone https://github.com/czbiohub/cellxgene
cd cellxgene

Datasets are stored in example-dataset

5. Launch cellxgene with:

cellxgene launch example-dataset

This should launch a cellxgene session with all the datasets in example-datasets/ loaded in.

If you're running excellxgene remotely, please launch with:

cellxgene launch example-datasets --host 0.0.0.0

Ping me on the Biohub slack (@Alec) if you have any questions!