config.yaml

system: imac
## clustering on which level
## so on L1, means we perform L2 level clustering
clustering_level: L1
retries: 1
conda:
  imac: sa2
  encoder: sa2
  tscc: sa2
  tscc_test: sa2
  test_unite: sa2

project_dir:
  imac: /Users/szu/git-recipes/mouseBrainAtlas/CEMBA2
  encoder: /projects/ps-renlab/szu/projects/CEMBA2
  tscc: /projects/ps-renlab/szu/projects/CEMBA2
  tscc_test: /projects/ps-renlab/szu/projects/CEMBA2
  test_unite: /Users/szu/git-recipes/mouseBrainAtlas/CEMBA2
  
work_dir:
  imac: /Users/szu/git-recipes/mouseBrainAtlas/CEMBA2/17.snapatac2
  encoder: /projects/ps-renlab/szu/projects/CEMBA2/17.snapatac2
  #tscc: /oasis/tscc/scratch/szu/projects/CEMBA2/17.snapatac2
  # use condo since scratch has slow IO.
  tscc: /projects/ps-renlab/szu/projects/CEMBA2/17.snapatac2
  tscc_test: /oasis/tscc/scratch/szu/projects/CEMBA2/17.snapatac2
  test_unite: /users/szu/test/17.snapatac2
out_dir: result
blacklist_file: 17.snapatac2/meta/mm10.blacklist.bed
barcode2id_file: 17.snapatac2/resource/barcode2id.csv
sample2fragment_file: 17.snapatac2/resource/sample2fragment.csv
pre_clustering_meta: 17.snapatac2/resource/cluster2size.csv
cemba_anndata_file: 17.snapatac2/resource/merge_cemba_all.h5ad
max_united_size: 200000
embed_nfeat: 500000
embed_nsample: 3000000
embed_ncomp: 30
embed_name: nfeat-default_nsample-all_nc30
embed:
  nfeat: 500000
  ncomp: 30
  embed_nsample: 3000000
  name: allfeat_nc30
knn:
  n: 50
  method: exact
  name: knn50
leiden:
  n_iter: -1
  weight: True
  obj: modularity
  min_size: 50
  repeat: 1
  minr: 0.1
  maxr: 2.0
  byr: 0.1
  seed: 0
  name: igraph_leiden_modularity
  n_sample: 20000
umap:
  n_neigh: 15
  n_comp: 2
  metric: euclidean
  init: spectral
  min_dist: 0.01
  a: 1.8956
  b: 0.8006
  seed: 0