Solutions

Snakefile

@@ -19,6 +19,11 @@ rule download_data:
     script:
         "scripts/download_data.py"
 
+# This rule should preprocess downloaded data by calling the `preprocess.py`
+# Python script.
+# The rule should read in the raw dataset, and filter cells + genes based off
+# of provided parameters as explained in the github issue. The rule should write
+# the newly processed data to a new `.h5ad` file. 
 rule preprocess_data:
     input:
         adata="data/raw/{dataset}.h5ad"
@@ -32,6 +37,10 @@ rule preprocess_data:
     script:
         "scripts/preprocess.py"
 
+# This rule should cluster cells using the `cluster_cells.py` script.
+# rule should read in preprocessed data, and clsuter cells according to user-
+# provided parameters, k and resolution. The rule should produce 3 output files
+# as csvs: a count matrix, a cell metadata table, and a gene metadata table.
 rule cluster_cells:
     input:
         adata="data/processed/{dataset}.h5ad"
@@ -45,6 +54,10 @@ rule cluster_cells:
     script:
         "scripts/cluster_cells.py"
 
+# This rule should plot clusters on UMAP projections using ggplot and the 
+# plot_cells.R script. 
+# the rule should read in the previously generated csvs, color cells according
+# to values in a user-specified column, and create a .png file containing the plot
 rule plot_clusters:
     input:
         X="data/clustered/{dataset}/X.csv",

install.sh

@@ -1,5 +1,5 @@
 module load miniconda
-conda create --name workshop seaborn scikit-learn statsmodels numba pytables
+conda create --name workshop seaborn scikit-learn statsmodels numba pytables pip
 conda activate workshop
-conda install -c conda-forge igraph python-igraph leidenalg imagemagick 
-pip install scanpy snakemake scikit-misc louvain jinja2 pygraphviz pygments
+conda install -c conda-forge igraph python-igraph leidenalg imagemagick louvain
+pip install scanpy==1.8.1 snakemake scikit-misc jinja2 pygraphviz pygments openpyxl