Estimate the optimized cut-off

Estimate the optimized cut-off that maximizes the coefficient of variation (CV) of each cell or sample.

estimate_optimized_cutoffs(
  data_exp_mat = NULL,
  interval = seq(from = floor(min(data_exp_mat)), to = ceiling(max(data_exp_mat)),
    length.out = 1000),
  gene_name_col = "GeneID",
  gene_type_col = "gene_type",
  anno_signature_genes = NULL,
  weight_by_gene_count = TRUE,
  prior_count = 2,
  do_parallel = TRUE,
  n_cores = NULL
)

Arguments

data_exp_mat

An expression matrix, e.g., raw count matrix or log2CPM matrix

interval

A sequence of cut-offs used for calculating the CVs, and the cut-off that maximize the CV is the optimized cut-off

gene_name_col

Colname name of row (gene) names used in the expression matrix

gene_type_col

Colname name of signature gene type annotation

anno_signature_genes

A data.frame containing signature gene annotation

weight_by_gene_count

Whether to divide the signature gene number by the total signature gene name, default is TRUE

prior_count

Add a prior count to avoid signature gene number to be 0, default is 2 but can be set to a different one

do_parallel

Whether do parallel computation or not, logical value, default is TRUE

n_cores

Number of cores used for parallel computation, half of the total cores will be used if not provided

Examples

# Set 'weight_by_gene_count' to TRUE is recommended for estimating the optimized cut-offs

start_time <- proc.time()

estimated_cutoffs <- estimate_optimized_cutoffs(
  data_exp_mat = edgeR::cpm(example_dge_data$counts,
                            log = TRUE),
  anno_signature_genes = anno_signature_genes_mouse,
  gene_name_col = "GeneID",
  gene_type_col = "gene_type",
  weight_by_gene_count = TRUE,
  prior_count = 2,
  do_parallel = TRUE,
  n_cores = 2
)

end_time <- proc.time() - start_time

end_time[3]
#> elapsed 
#>    3.05 

estimated_cutoffs
#> 10_6_5_11  9_6_5_11   purep53    JMS8-2    JMS8-3    JMS8-4    JMS8-5  JMS9-P7c 
#>  9.354354 11.496496  8.053053  9.774775 11.896897 10.075075  8.553554  8.253253 
#>  JMS9-P8c 
#> 10.675676 

data_for_ternary <- generate_data_for_ternary(
  data_exp_mat = edgeR::cpm(example_dge_data$counts,
                            log = TRUE),
  anno_signature_genes = anno_signature_genes_mouse,
  gene_name_col = "GeneID",
  gene_type_col = "gene_type",
  weight_by_gene_count = TRUE,
  cutoff_exp = estimated_cutoffs,
  prior_count = 2
)

vcdTernaryPlot(data = data_for_ternary,
  order_colnames = c(2,3,1),
  group = example_dge_data$samples$group,
  group_color = c("red","green","blue"),
  point_size = 1,
  legend_point_size = 0.6,
  legend_position = c(0.3,0.5),
  scale_legend = 1)