Latch Bio - Automated Single Cell Filtering

LatchBio Implementation of Dropkick for automated single cell filtering

This workflow automates single cell filtering using Dropkick - source.

You can find the deployed version of this workflow on the Latch Console. This was built using the tutorial from the Dropkick repo and the docs from the Latch SDK.

Basic Usage:

Input

A directory containing 10x Single Cell data, which should contain:

1. barcodes.tsv
2. genes.tsv
3. matrix.mtx

Please also include a name for the run - e.g Human_PBMCs, Mouse_TH17

Output

A HTML file in the root directory named RUN_NAME__SCDK_Report.html, which is a converted Jupyter notebook containing:

1. Quality Control plots
2. Score plots
3. Gene coefficients list and plots
4. Prinicipal Component Analyses Plots
5. UMAP comparisons to other methods

Usage Notes:

• The cell cycle genes used are as defined in Tirosh et al, Science, 2016, these are not currently changable.
• The software packages that power this workflow (Scanpy & Dropkick) were initially designed to be used as interactive tools. To overcome this, this workflow programatically constructs notebooks using nbformat and nbconvert.

TO DO:

1. Integration of the outputs of this workflow with the inputs of downstream analysis.
2. Cleaner output HTML file.

Contributions

Please feel free to contribute to this repo and improve it!