Setup

This pipeline requires Snakemake to run.

Configuration

All configuration for generating the files required by DrEdGE is placed in a file named config.yaml. Create that file in this directory.

Loading data

There are two methods for loading your dataset which you can configure in config.yaml. The first takes precedence over the second.

1. Using a pre-configured Seurat object

If you know how to configure your dataset with Seurat and you want to manually tune clustering, UMAP generation, and cluster labeling yourself, you can load in a pre-configured Seurat object via an .rds file. To do so, add the following line to your configuration file:

seurat_object: /path/to/your/object.rds

This Seurat object should already have been normalized, scaled, and had FindNeighbors, FindClusters, and RunUMAP performed on it.

2. Using an expression count matrix

If you do not have experience with Seurat, or you want to let this pipeline run the typical Seurat workflow for you, then you need to provide an expression count matrix to be read. To do so, add the following line to your configuration file:

count_matrix: /path/to/your/counts/file.matrx

Parsing transcript names

If the transcript names in your dataset have a prefix, you can add an option to the configuration to separate it out of the name with the option transcript_id_prefix. For example, if you have transcripts with bardcodes, named something like 98439713|wnt-1, you could add the option:

transcript_id_prefix: (\d+)\|(.*)

(Where (\d+)\|(.*) is a regular expression that separates the name of your transcript from the barcode). This will tell DrEdGE that your transcript can be named one of three things: 98439713|wnt-1, 98439713, or wnt-1. The last captured group will be the canonical name of your transcript.

Running

Once you have configured your dataset, run snakemake -j 1 in this directory. Generated files will be placed in the out directory.