| Title: | Single Cell Entropy Analysis of Gene Heterogeneity in Cell Populations |
|---|---|
| Description: | Analyse single cell RNA sequencing data using entropy to calculate heterogeneity and homogeneity of genes amongst the cell population. From the work of Michael J. Casey, Ruben J. Sanchez-Garcia and Ben D. MacArthur (2020) <doi:10.1101/2020.10.01.322255>. |
| Authors: | Hugh Warden [aut, cre] |
| Maintainer: | Hugh Warden <[email protected]> |
| License: | GPL (>= 3) |
| Version: | 0.1.0 |
| Built: | 2024-06-25 05:31:56 UTC |
| Source: | https://github.com/hwarden162/scent |
Find the Number of Times Each Gene Has Been Expressed
gene_counts(expr, transpose = FALSE)gene_counts(expr, transpose = FALSE)
expr |
A matrix of gene expressions with cells as rows and genes as columns |
transpose |
A logical value indicating whether the matrix should be transposed before operations are carried out |
A vector of counts of expression for each gene
# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) genes <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(genes) <- paste0("cell", 1:7) colnames(genes) <- paste0("gene", 1:5) #Calculating Gene Counts gene_counts(genes)# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) genes <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(genes) <- paste0("cell", 1:7) colnames(genes) <- paste0("gene", 1:5) #Calculating Gene Counts gene_counts(genes)
Find the Heterogeneity of a Gene Within a Population
gene_het(expr, unit = "log2", normalise = TRUE, transpose = FALSE)gene_het(expr, unit = "log2", normalise = TRUE, transpose = FALSE)
expr |
A vector or matrix of gene expressions. For the matrix, genes should be represented as rows and cells as columns. |
unit |
The units to be parsed to the entropy function. |
normalise |
A logical value representing whether the gene frequencies should be normalised into a distribution. |
transpose |
A logical value representing whether the matrix should be transposed before any calculations are performed. |
A vector of the information gained from the gene distribution compared to the uniform distribution. The higher the value more heterogeneous the cell is within the population.
# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) # Calculating Heterogeneity For Each Gene gene_het(gene1) gene_het(gene2) gene_het(gene3) gene_het(gene4) gene_het(gene5) # Calculating Heterogeneity For a Matrix gene_het(gene_counts)# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) # Calculating Heterogeneity For Each Gene gene_het(gene1) gene_het(gene2) gene_het(gene3) gene_het(gene4) gene_het(gene5) # Calculating Heterogeneity For a Matrix gene_het(gene_counts)
Find the Homogeneity of a Gene Within a Population
gene_hom(expr, unit = "log2", normalise = TRUE, transpose = FALSE)gene_hom(expr, unit = "log2", normalise = TRUE, transpose = FALSE)
expr |
A vector or matrix of gene expressions. For the matrix, genes should be represented as rows and cells as columns. |
unit |
The units to be parsed to the entropy function. |
normalise |
A logical value representing whether the gene frequencies should be normalised into a distribution. |
transpose |
A legical value representing whether the matrix should be transposed before any calculations are performed. |
A vector of the information contained in the distribution of each gene. The higher this is, the more homogeneous the gene is within the cell population.
# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) # Calculating Homogeneity For Each Gene gene_hom(gene1) gene_hom(gene2) gene_hom(gene3) gene_hom(gene4) gene_hom(gene5) # Calculating Homogeneity For a Matrix gene_hom(gene_counts)# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) # Calculating Homogeneity For Each Gene gene_hom(gene1) gene_hom(gene2) gene_hom(gene3) gene_hom(gene4) gene_hom(gene5) # Calculating Homogeneity For a Matrix gene_hom(gene_counts)
A function that takes frequency count data and normalises it into a probability distribution. Only available internally within SCEnt.
normalise(dist)normalise(dist)
dist |
A vector of a frequency distribution. |
A vector of a probability distribution relative to the frequencies.
Remove Lowly Expressed Genes From Expression Data
rm_low_counts( expr, count_threshold = NULL, perc_threshold = NULL, transpose = FALSE )rm_low_counts( expr, count_threshold = NULL, perc_threshold = NULL, transpose = FALSE )
expr |
A matrix of gene expression with cells as rows and genes as columns |
count_threshold |
A threshold for the number of counts a gene must have to be included. Only one threshold may be used at a time. |
perc_threshold |
A threshold for what percentile the gene counts should be cut off at. Only one threshold may be used at a time. |
transpose |
A logical value indicating whether the expression matrix should be transposed before any operations are carried out. |
A matrix of gene expressions with the low count genes, as specified by the user, removed.
# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) # Removing Low Count Genes rm_low_counts(gene_counts, count_threshold = 7) rm_low_counts(gene_counts, perc_threshold = 0.1)# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) # Removing Low Count Genes rm_low_counts(gene_counts, count_threshold = 7) rm_low_counts(gene_counts, perc_threshold = 0.1)
Tidy Wrapper To Remove Lowly Expressed Genes From Expression Data
rm_low_counts_tidy( expr, count_threshold = NULL, perc_threshold = NULL, transpose = FALSE )rm_low_counts_tidy( expr, count_threshold = NULL, perc_threshold = NULL, transpose = FALSE )
expr |
A tibble of gene expression with cells as rows and genes as columns |
count_threshold |
A threshold for the number of counts a gene must have to be included. Only one threshold may be used at a time. |
perc_threshold |
A threshold for what percentile the gene counts should be cut off at. Only one threshold may be used at a time. |
transpose |
A logical value indicating whether the expression matrix should be transposed before any operations are carried out. |
A tibble of gene expressions with the low count genes, as specified by the user, removed.
# Creating Data library(tibble) gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) gene_counts <- as_tibble(gene_counts) # Removing Low Count Genes rm_low_counts_tidy(gene_counts, count_threshold = 7) rm_low_counts_tidy(gene_counts, perc_threshold = 0.1)# Creating Data library(tibble) gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) gene_counts <- as_tibble(gene_counts) # Removing Low Count Genes rm_low_counts_tidy(gene_counts, count_threshold = 7) rm_low_counts_tidy(gene_counts, perc_threshold = 0.1)
Feature Selection by Gene Heterogeneity
scent_select( expr, bit_threshold = NULL, count_threshold = NULL, perc_threshold = NULL, unit = "log2", normalise = TRUE, transpose = FALSE )scent_select( expr, bit_threshold = NULL, count_threshold = NULL, perc_threshold = NULL, unit = "log2", normalise = TRUE, transpose = FALSE )
expr |
A matrix of gene expression data. Cells should be represented as rows and genes should be represented as columns. |
bit_threshold |
The threshold for the amount of bits of information a gene must add to be selected as a feature. Only one threshold can be used at a time. |
count_threshold |
A number represented how many of the most heterogeneous cells should be selected. Only one threshold can be used at a time. |
perc_threshold |
The percentile of the hetergeneity distribution above which a gene should be to be selected as a feature. |
unit |
The units to be used when calculating entropy. |
normalise |
A logical value representing whether the gene counts should be normalised into a probability distribution. |
transpose |
A logical value representing whether the matrix should be transposed before having any operations computed on it. |
A matrix of gene expression values where genes with low heterogeneity have been removed.
# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) # Performing Feature Selection scent_select(gene_counts, bit_threshold = 0.85) scent_select(gene_counts, count_threshold = 2) scent_select(gene_counts, perc_threshold = 0.25)# Creating Data gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) # Performing Feature Selection scent_select(gene_counts, bit_threshold = 0.85) scent_select(gene_counts, count_threshold = 2) scent_select(gene_counts, perc_threshold = 0.25)
A Tidy Wrapper for Feature Selection by Heterogeneity
scent_select_tidy( expr, bit_threshold = NULL, count_threshold = NULL, perc_threshold = NULL, unit = "log2", normalise = TRUE, transpose = FALSE )scent_select_tidy( expr, bit_threshold = NULL, count_threshold = NULL, perc_threshold = NULL, unit = "log2", normalise = TRUE, transpose = FALSE )
expr |
A tibble of gene expression data. Cells should be represented as rows and genes should be represented as columns. |
bit_threshold |
The threshold for the amount of bits of information a gene must add to be selected as a feature. Only one threshold can be used at a time. |
count_threshold |
A number represented how many of the most heterogeneous cells should be selected. Only one threshold can be used at a time. |
perc_threshold |
The percentile of the hetergeneity distribution above which a gene should be to be selected as a feature. |
unit |
The units to be used when calculating entropy. |
normalise |
A logical value representing whether the gene counts should be normalised into a probability distribution. |
transpose |
A logical value representing whether the matrix should be transposed before having any operations computed on it. |
A tibble of gene expression values where genes with low heterogeneity have been removed.
# Creating Data library(tibble) gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) gene_counts <- as_tibble(gene_counts) # Performing Feature Selection scent_select_tidy(gene_counts, bit_threshold = 0.85) scent_select_tidy(gene_counts, count_threshold = 2) scent_select_tidy(gene_counts, perc_threshold = 0.25)# Creating Data library(tibble) gene1 <- c(0, 0, 0, 0, 1, 2, 3) gene2 <- c(5, 5, 3, 2, 0, 0, 0) gene3 <- c(2, 0, 2, 1, 3, 0, 1) gene4 <- c(3, 3, 3, 3, 3, 3, 3) gene5 <- c(0, 0, 0, 0, 5, 0, 0) gene_counts <- matrix(c(gene1, gene2, gene3, gene4, gene5), ncol = 5) rownames(gene_counts) <- paste0("cell", 1:7) colnames(gene_counts) <- paste0("gene", 1:5) gene_counts <- as_tibble(gene_counts) # Performing Feature Selection scent_select_tidy(gene_counts, bit_threshold = 0.85) scent_select_tidy(gene_counts, count_threshold = 2) scent_select_tidy(gene_counts, perc_threshold = 0.25)