| Title: | Performing Single-Cell RNA-Seq Imputation by Using Mean/Median Imputation |
|---|---|
| Description: | Performing single-cell imputation in a way that preserves the biological variations in the data. The package clusters the input data to do imputation for each cluster, and do a distribution check using the Anderson-Darling normality test to impute dropouts using mean or median (Yazici, B., & Yolacan, S. (2007) <DOI:10.1080/10629360600678310>). |
| Authors: | Mohamed Soudy [aut, cre], Sascha Jung [aut], Antonio DEL SOL [aut] |
| Maintainer: | Mohamed Soudy <[email protected]> |
| License: | GPL-3 |
| Version: | 0.1 |
| Built: | 2024-11-03 04:05:46 UTC |
| Source: | https://github.com/mohmedsoudy/scrnaimm |
Cluster cells based on scDHA methods to get cluster labels
cluster_cells(ScRNA_filtered, Normalize = TRUE, k=NULL, n=5000)cluster_cells(ScRNA_filtered, Normalize = TRUE, k=NULL, n=5000)
ScRNA_filtered |
ScRNA-seq data set generated by filter_ScRNA function |
Normalize |
Boolean parameter whether to apply log10 normalization for the data or not |
k |
Number of clusters if there is a prior knowledge about that |
n |
Number of genes to keep after feature selection step |
a vector that contains the cell labels
Mohamed Soudy [email protected]
Evaluate the clustering if you have the original labels
evaluate_clustering(cluster_labels, original_labels)evaluate_clustering(cluster_labels, original_labels)
cluster_labels |
Cluster labels generated by cluster_cells functions or user-defined |
original_labels |
Original labels of the ScRNA-seq data |
ARI of clustering 'a value between 0 and 1' 1 indicates best clustering
Mohamed Soudy [email protected]
evaluate_clustering(c(1,1,1,1,2,2,3,3), c(1,1,1,1,3,3,3,2))evaluate_clustering(c(1,1,1,1,2,2,3,3), c(1,1,1,1,3,3,3,2))
Filter out the genes that are not expressed in at least one cell type
filter_ScRNA(ScRNA_mat)filter_ScRNA(ScRNA_mat)
ScRNA_mat |
ScRNA-Seq matrix where genes are in rows and cells are in columns |
Filtered ScRNA-seq matrix that contains genes that are at least expressed in one cell type
Mohamed Soudy [email protected]
This function aims to get the indices where all genes are zeros or genes that is expressed on only one sample per cell
prepare_dataset(filtered_data, cluster_labels)prepare_dataset(filtered_data, cluster_labels)
filtered_data |
ScRNA-seq data set generate by filter_ScRNA function |
cluster_labels |
cell labels |
a data frame that contains the processed ScRNA-seq data
Mohamed Soudy [email protected]
Run the main pipeline for ScRNAIMM
run_pipeline(ScRNA,label=NULL,k=NULL,cells=TRUE,genes=TRUE,outdir=NULL,dataset=NULL)run_pipeline(ScRNA,label=NULL,k=NULL,cells=TRUE,genes=TRUE,outdir=NULL,dataset=NULL)
ScRNA |
ScRNA-seq data set generated by prepare_dataset function |
label |
Prior knowledge about cluster labels if NULL, will use our clustering function |
k |
Prior knowledge about number of clusters if NULL, will use our clustering function |
cells |
Boolean whether to do the imputation based on cell clustering or not |
genes |
Boolean whether to do the imputation based on genes or not |
outdir |
Path to output directory to write the imputed data |
dataset |
Name of the data set to be the name of the output directory |
a data frame with the imputed values
Mohamed Soudy [email protected]
Perform ScRNA-seq imputation using mean/Median
ScRNA_imp_MM(ScRNA_filtered, cluster_labels = NULL, cells = TRUE, genes = FALSE)ScRNA_imp_MM(ScRNA_filtered, cluster_labels = NULL, cells = TRUE, genes = FALSE)
ScRNA_filtered |
ScRNA-seq data set generated by prepare_dataset function |
cluster_labels |
Cluster labels generated by cluster_cells function or user-defined |
cells |
Boolean whether to do the imputation based on cell clustering or not |
genes |
Boolean whether to do the imputation based on genes or not |
a data frame with the imputed values
Mohamed Soudy [email protected]
Performs a distribution check for the data
scRNA_MMI(scRNA_dataset)scRNA_MMI(scRNA_dataset)
scRNA_dataset |
ScRNA-seq data set |
ScRNA-Seq Imputed data set
Mohamed Soudy [email protected]