sc-flow

setup

This package is designed for use on Andes/Frontier and assumes you have a working miniconda/conda installation. To generate your own copy of our R env, folllow the following steps:

1. Run git clone https://github.com/jeanmerlet/sc-flow.git
2. In /env, run echo "prefix: /path/to/env/dir" >> frontier_seurat.yml
3. In /env, run conda env create -f frontier_seurat.yml
4. In the root directory, run ./project_init.sh
5. Place your fastqs in /data/raw
6. Create and fill a /data/metadata/sample_meta.tsv file (see metadata)

Metadata

Broad order of steps

In general, a typical scRNA-seq workflow will go through these steps (in order).

1. align
2. preprocess
3. cluster
4. plot
5. diff_exp

Each of these corresponds to an argument for the --workflow command in the sc-flow pipeline.

1. align: --workflow align
2. preprocess: --workflow preprocess
3. cluster: --workflow cluster
4. plot: --workflow plot
5. diff_exp: --workflow diff_exp

Example workflow

1. Rscript sc-flow.R --workflow align
2. Rscript sc-flow.R --workflow preprocess --species mouse --load_meta
3. Rscript sc-flow.R --workflow plot --plot_type qc --species mouse --color_by condition
4. Rscript sc-flow.R --workflow plot --plot_type umap --color_by sample_ids
5. Rscript sc-flow.R --workflow plot --plot_type umap --color_by condition
6. Rscript sc-flow.R --workflow cluster
7. Rscript sc-flow.R --workflow plot --plot_type umap --color_by clusters
8. Rscript sc-flow.R --workflow plot --plot_type umap --color_by clusters --split_by condition
9. Rscript sc-flow.R --workflow diff_exp --diff_type cluster
10. Rscript sc-flow.R --workflow diff_exp --diff_type condition --condition_group atopic_dermatitis

NOTE: talk about interactivity with --run\_r

NOTE: make sure --run\_r works with some weird species options stuff for preprocess / qc plot

NOTE: tell users where their data is located (especially for plots, count matrices, etc.)

NOTE: generalize jobscript generation into one function with arguments (including account to charge (e.g. syb111))