ConstExpRef: Construct mRNA expression reference matrix

In aet21/EpiSCORE: Epigenetic cell-type deconvolution with Single-Cell Omic References

Construct mRNA expression reference matrix

Description

This function takes as input a tissue-specific scRNA-Seq atlas and produces an expression reference matrix for the cell-types specified by the user

Usage

ConstExpRef(exp.m, celltype.idx, namesCellT.v, markspecTH.v = NULL)

Arguments

exp.m	The normalized scRNA-Seq data matrix with rows labeling unique genes (either human gene symbol or Entrez gene ID) and columns labeling cells. Missing values are not allowed and values should be positive or zero. Importantly, the value corresponding zero read counts should also be zero. Typically, this matrix will be obtained by scaling the readcounts, adding a pseudocount of +1 and then taking the log2 transform.
celltype.idx	An integer index vector with as many elements as there are columns in 'exp.m', specifying the cell-type of each cell. Each unique integer in the vector defines one cell-type and integers must be sequential, i.e. no gaps are allowed. For instance, for 4 cell-types, the integers would be 1 to 4.
namesCellT.v	A vector of unique names for the cell-types, of length equal to the number of unique elements in celltype.idx. The order of the names must match the numerical integers in celltype.idx.
markspecTH.v	This is the marker specificity score (MSS) threshold for each cell-type, i.e. for a given marker gene and cell-type, the minimum number of other cell-types where we require this gene not to be expressed. For n cell-types, the maximum MSS value is n-1 and minimum value should be 1. The parameter must be a vector with one threshold value for each cell-type. If 'NULL' (the default option), each entry will be set to n-1.

Value

A list with two elements.

ref	a list itself also with 2 elements, giving the reference expression matrices defined using averages (av) or medians (med).
markers	a list of matrices, one for each cell-type listing for each marker gene in the reference, their Wilcoxon rank sum statistics, median value in each cell-type and their marker specificity score.

References

Teschendorff AE, Zhu T, Breeze CE, Beck S. Cell-type deconvolution of bulk tissue DNA methylomes from single-cell RNA-Seq data Genome Biol.2020

Examples

data(lungSS2mca1)
out.l <- ConstExpRef(lungSS2mca1.m,celltypeSS2.idx,celltypeSS2.v,markspecTH.v=rep(3,4));
print(head(out.l$ref$med));

aet21/EpiSCORE documentation built on June 26, 2022, 5:50 p.m.