get_mean_ratio: Get Mean Ratio for Each Gene x Cell Type
In lahuuki/DeconvoBuddies: Helper Functions for LIBD Deconvolution

get_mean_ratio

R Documentation

Get Mean Ratio for Each Gene x Cell Type

Description

Calculate the Mean Ratio value and rank for each gene for each cell type in the sce object, to identify effective marker genes for deconvolution.

Usage

get_mean_ratio(
  sce,
  cellType_col,
  assay_name = "logcounts",
  gene_ensembl = NULL,
  gene_name = NULL
)

Arguments

`sce`	SummarizedExperiment-class (or any derivative class) object containing single cell/nucleus gene expression data.
`cellType_col`	A `character(1)` name of the column in the colData() of `sce` that denotes the cell type or group of interest.
`assay_name`	A `character(1)` specifying the name of the assay() in the `sce` object to use to rank expression values. Defaults to `logcounts` since it typically contains the normalized expression values.
`gene_ensembl`	A `character(1)` specifying the `rowData(sce_pseudo)` column with the ENSEMBL gene IDs. This will be used by `layer_stat_cor()`.
`gene_name`	A `character(1)` specifying the `rowData(sce_pseudo)` column with the gene names (symbols).

Details

Note if a cell type has < 10 cells the MeanRatio results may be unstable. See rational in OSCA: https://bioconductor.org/books/3.19/OSCA.multisample/multi-sample-comparisons.html#performing-the-de-analysis.

Value

A tibble::tibble() with the MeanRatio values for each gene x cell type.

gene is the name of the gene (from rownames(sce)).
cellType.target is the cell type we're finding marker genes for.
mean.target is the mean expression of gene for cellType.target.
cellType.2nd is the second highest non-target cell type.
mean.2nd is the mean expression of gene for cellType.2nd.
MeanRatio is the ratio of mean.target/mean.2nd.
MeanRatio.rank is the rank of MeanRatio for the cell type.
MeanRatio.anno is an annotation of the MeanRatio calculation helpful for plotting.
gene_ensembl & gene_name optional columns from rowData(sce) specified by the user to add gene information.

Examples

## load example SingleCellExperiment
if (!exists("sce_DLPFC_example")) sce_DLPFC_example <- fetch_deconvo_data("sce_DLPFC_example")
## Explore properties of the sce object
sce_DLPFC_example

## this data contains logcounts of gene expression
SummarizedExperiment::assays(sce_DLPFC_example)$logcounts[1:5, 1:5]

## nuclei are classified in to cell types
table(sce_DLPFC_example$cellType_broad_hc)

## Get the mean ratio for each gene for each cell type defined in
## `cellType_broad_hc`
get_mean_ratio(sce_DLPFC_example, cellType_col = "cellType_broad_hc")

# Option to specify gene_name as the "Symbol" column from rowData
# this will be added to the marker stats output
SummarizedExperiment::rowData(sce_DLPFC_example)

## specify rowData col names for gene_name and gene_ensembl
get_mean_ratio(sce_DLPFC_example,
    cellType_col = "cellType_broad_hc",
    gene_name = "gene_name",
    gene_ensembl = "gene_id"
)

lahuuki/DeconvoBuddies documentation built on Dec. 18, 2024, 6:24 a.m.