v2022.12.0

Updating notebooks to use RAPIDS version 22.12. Broke atacworks into a separate conda file since scikit-learn, numpy, and scipy dependencies have changed since version 22.08 of RAPIDS.

v2022.08.0

This release updates the notebooks and conda environments to work with RAPIDS version 2022.08.

v2022.02.0

This release updates the notebooks and conda environments to work with RAPIDS version 2022.02.

Reverted back to single GPU for clustering and visualization steps in 1M cells multi-GPU notebook

v2021.12.0

This release updates the notebooks and conda environments to work with RAPIDS version 2021.12.

1. The 70k lung cell and 1.3M neuron example notebooks no longer need to batch the PCA.
2. 1.3M neuron multi-gpu example notebook now performs umap inference, kmeans, and PCA across multiple gpus.
3. 1.3M neuron notebooks now contain differential gene expression analysis.

v2021.06.0

Changes in this release includes:

• Upgrade to Rapids 21.06
• Introduce batch processing in 'Single-cell RNA-seq of 1 Million Mouse Brain Cells' to enable processing 1.3 million cells.

v2021.03.0

This release includes the following five example notebooks, covering single-cell RNA-seq and single-cell ATAC-seq analysis with RAPIDS.

Example 1: Single-cell RNA-seq of 70,000 Human Lung Cells
Example 2: Single-cell RNA-seq of 1 Million Mouse Brain Cells
Example 3: GPU-based Interactive Visualization of 70,000 Human Lung Cells
Example 4: Droplet Single-cell ATAC-seq of 60K Bone Marrow Cells
Example 5: Visualizing Chromatin Accessibility in 5,000 PBMCs with RAPIDS and AtacWorks

The release includes many functions, which are demonstrated in the notebooks, to enable easy adaptation of RAPIDS for single-cell analysis.
rapids_scanpy_funcs.py: preprocessing, clustering and differential analysis
coverage.py: functions to calculate coverage of a genomic region in single-cell ATAC-seq and integrate with AtacWorks for denoising and peak calling
visualize.py: functions to create interactive visualizations of single-cell data using RAPID and Plotly.