config.yaml

---
project_name: clustering_test
# path to the Cell Ranger ouput; it can even be a CSV file like the one in a Cell Ranger aggregation
input_expression: "/home/ciro/ad_hoc/fungal_allergy/raw/cellranger/aggr/non_prolif"
# A table of the metadata per library, or the whole metadata ready.
# It can be multiple files: [metadata_lib.csv, demultiplexed_hashtags.rdata]
metadata:
  - "/home/ciro/fungal_allergy/info/metadata_library.csv"
  - "/home/ciro/ad_hoc/fungal_allergy/results/quality_control/non_prolif/metadata_prefilter.rdata"
# if you don't add "/" at the end, it will append 'project_name'
output_dir: "/home/ciro/ad_hoc/fungal_allergy/results/clustering/"
filtering:
  subset: {expr: "nFeature_RNA >= 200 & nFeature_RNA <= Inf & nCount_RNA >= -Inf & nCount_RNA <= Inf & percent.mt <= 40"}
  nSamples_expressed: 0.001 # % of cells in which that gene is present
regress_var: [nCount_RNA, percent.mt]
norm: LogNormalize
variable_features:
  file: ./data/prot_coding_genes.csv
  method: vst
  nfeatures: 2000
  percent: [10, 15, 20, 25]  # iterated
  mean.cutoff: [0.01, 8]
  dispersion.cutoff: [1, Inf]
resolution: [0.1, 0.2, 0.4, 0.6, 0.8] # iterated during markers
dim_reduction:
  base: {type: pca, n_comp: 40, chosen_comp: [10, 15, 25]} # iterated: chosen
  tsne: {perplexity: 'auto', reduction.use: "pca"}
  umap: {n.neighbors: 30, min.dist: 0.3}
markers:
  select: snn_res. # pattern in the metadata's columns
  test: MAST
  avg_logFC: 0.25
  p_val_adj: 0.05
tool: seurat
script: ./R/seurat.R
exec: Rscript
pipeline: ./
cluster_config: ./cluster.json
environment: clustering
...