Leiden usage workflow

example data:

# a raw scRNA-seq data, goolam.rds, has been stored in the ./scRNAsqData.
# https://hemberg-lab.github.io/scRNA.seq.datasets/mouse/edev/#goolam
da <- readRDS("./scRNAseqData/goolam.rds")
da

output:

class: SingleCellExperiment
dim: 41428 124
metadata(0):
assays(2): counts logcounts
rownames(41428): ENSMUSG00000000001 ENSMUSG00000000003 ... ERCC-00170 ERCC-00171
rowData names(19): is_feature_control is_feature_control_ERCC ... feature_symbol feature_id
colnames(124): X2cell_1_A X2cell_1_B ... ME_4cell_6_C ME_4cell_6_D
colData names(31): cell_type1 total_features ... pct_counts_top_50_features_ERCC is_cell_control
reducedDimNames(0):
mainExpName: NULL
altExpNames(0)

Step 1: Normalizing and mapping the raw scRNA-seq data to multiple low-dimensional latent spaces. A 9_latent_data folder is produced and saved in the *./OutputData*.

# activate the environment
conda activate DEPF
cd DEPF/HierarchicalAutoencoder/
Rscript runHA.R

output:

[1] "goolam.rds"
output 9 of  1  latent:  goolam .
output 9 of  2  latent:  goolam .
output 9 of  3  latent:  goolam .
output 9 of  4  latent:  goolam .
output 9 of  5  latent:  goolam .
output 9 of  6  latent:  goolam .
output 9 of  7  latent:  goolam .
output 9 of  8  latent:  goolam .
output 9 of  9  latent:  goolam .

Step 2: Selecting Leiden to generate a clustering ensemble.

R
source("runLeiden.R")
runLeiden(res=1, ensemble_num = 10)

output:

********************************************** .
output 9 of  1  latent:  goolam .
********************************************** .
---------------------------------------------- .
resolution:  1 .
---------------------------------------------- .
Computing nearest neighbor graph
Computing SNN
[1] "============  Leiden   ============"
...
...
...
********************************************** .
output 9 of  9  latent:  goolam .
********************************************** .
---------------------------------------------- .
resolution:  1 .
---------------------------------------------- .
Computing nearest neighbor graph
Computing SNN
[1] "============  Leiden   ============"

The Leiden_resolution_1.csv is produced and saved in the ./OutputData.

Step 3: Performing dynamic ensemble pruning. The final_clustering.csv is produced and saved in the ./OutputData.

runBioFOA("Louvain", 5, 1)

output:

==========run Ensemble Pruning.==================
goolam
============load Leiden ensemble=================
=========  NMI  ==========
    0.8400

=========  ARI  ==========
    0.6700