CCI (Consensus Clustering Imputation) is a novel imputation method inspired by consensus clustering techniques. It aims to address the issue of dropouts in single-cell RNA sequencing (scRNA-seq) and other high-dimensional data. CCI borrows the information from similar cells by ensemble clustering to identify robust neighbors, and employs the cell-to-cell similarities to impute gene expression levels.
You can install the development version of ccImpute from GitHub with:
#install.packages("devtools")
#devtools::install_github("wanlinjuan/ccImpute")library(ccImpute)
library(Seurat)
library(splatter)
library(scuttle)
set.seed(1)
params <- newSplatParams(batchCells = 200, nGenes = 100)
sim <- splatSimulate(params, verbose = FALSE)
sim <- logNormCounts(sim)
seurat_obj <- as.Seurat(sim)
data <- GetAssayData(seurat_obj, slot = "data")
hv_genes <- rownames(data)[1:50]
# Impute dropout values
imputed <- cc_impute(data, num_sampling = 5, select_genes = hv_genes, num_clusters = 2)If you use ccImpute in your research, please cite:
Juan, W.; Ahn, K.W.; Chen, Y.-G.; Lin, C.-W. CCI: A Consensus Clustering-Based Imputation Method for Addressing Dropout Events in scRNA-Seq Data. Bioengineering 2025, 12, 31. https://doi.org/10.3390/bioengineering12010031
This package is licensed under GPL (>= 3).