data-scripts

Scripts to demultiplex or pre-process user data to get into Cellenics.

Utils.R

Contains utilities for the other functions.

filter_empty_drops.R

Usage:

1. Download samples using aws s3 cp s3://biomage-originals-production/PROJECTUUID input --recursive
2. Copy samples table (Pure json as opposed to dynamodb JSON) into samples.json
3. run "python3 rename_samples.py"
4. Open data-scripts.rproj and load renv dependencies renv::restore()
5. Use the filter_empty_drops function to filter all samples in the input dir

hto_demultiplex.R

Usage:

1. Follow usage instructions for filter_empty_drops.R to filter all samples in the input dir
2. Use the hto_demux function to demultiplex all samples in the out dir, which were previously filtered by the filter_empty_drops function

cellset extraction

To extract cellsets, you only need an experiment ID, and the index of the cellset in the cellsets file. The cellset index is composed from the cellset class, as listed below, and the cellset number inside each class (with 1-based indexing).

• 1 = louvain
• 2 = scratchpad
• 3 = samples
• 4 = metadata tracks

The easiest way to get this is to use Rstudio's list viewer: View(parsed_json_object). You can download the cellset file using the download_cellset_file function, import it with jsonlite::read_json and explore it with the list viewer.

After getting the positions, The function extract_cellset will do everything automagically, returning a subsetted seurat object.