sqsun / momf

MOMF

Deconvolution analysis with the bulk RNA-seq data and single-cell RNA-seq data

Installation

### install devtools packages (devtools package)
install.packages("devtools")

### install MOMF package
devtools::install_github("sqsun/MOMF")

Example data

The example data (toy_example.Rdata) is a simulated data.

• sc_counts: scRNA-seq gene expression matrix (#cells x #genes);
• sc_cell_type: cell types for scRNA-seq data (#cells x 1);
• bulk_counts: bulk RNA-seq gene expression matrix (#individuals x #genes).
  

Example Code

Two main functions momf.fit and momf.computeRef are used to do deconvoluation analysis.

### load MOMF package
> library(MOMF)

### load example data
> load("toy_example.RData")

### compute the cell type specific expression level as reference
> priorU <- momf.computeRef(sc_counts, sc_cell_type)

### create the gene list for MOMF 
> GList <- list(X1 = t(sc_counts), X2 = t(bulk_counts))

### run MOMF
> momf_res <- momf.fit(DataX = GList, DataPriorU=priorU, method="KL", rho=2, num_iter=100)

### output the cell type proportions
> cell_prop <- momf_res$cell.prop
> heatmap(cell_prop)

Citation

Xifang Sun, Shiquan Sun, and Sheng Yang. An efficient and flexible method for deconvoluting bulk RNAseq data with single-cell RNAseq data, 2019, DIO: 10.5281/zenodo.3373980.

Supports

Please reach out Xifang Sun or Sheng Yang (email: [email protected] or [email protected]) if you have any questions.

momf's Issues

Release Version

We are currently working on putting MOMF on bioconda, but for that we would need you to create a release. Would that be possible?
Thank you very much in advance!

Error in momf.computeRef function

Hi,

momf.computeRef function returns error due to lack of names in weight vector. This implementation should solve it.

momf.computeRef <- function(sc_counts, cell_type){

   U <- sapply(unique(cell_type), function(ct){
          y = sc_count[ ,cell_type %in% ct, drop = FALSE]
          weight = sum(y)/sum(sc_count)
          rowSums(y)/sum(y)*weight
      })
	U <- 1000000 * U
	colnames(U) <- unique(cell_type)
	rownames(U) <- rownames(sc_count)
	# return estimated mean for each cell type
	return(U)
}

P.D: I would also be straight forward to implement this function in parelel.

Thanks