pkgdown/extra.css

Skip to contents

Calculate and plot read fraction correlation across wells (T-SHELL)

plot_tshell.Rd

[Experimental]

This function takes in an object with well-level read counts for a sample, loaded by load_well_counts_binary() along with a CDR3 sequence of interest for TCRalpha or TCRbeta. It calculates the correlation of the input sequence with all of its potential partner chains and returns the top N chains with the highest correlation/T-SHELL value.

Usage

plot_tshell(
  well_data,
  nuc_seq,
  chain,
  n_plot = 9,
  plot = TRUE,
  plot_log_scale = FALSE,
  interactive = FALSE,
  normalize_counts = TRUE,
  loss_frac_cutoff = 0.5,
  calc_fisher_pval_all = FALSE
)

Arguments

well_data

well-level read count data loaded by load_well_counts_binary()

nuc_seq

the CDR3 nucleotide sequence of the chain of interest

chain

the chain of the input nucleotide sequence, either "alpha" or "beta"

n_plot

the number of scatter plots to return (one for each of top N potential partners, default is 9)

plot

whether to plot the data or only return data frames (default is TRUE)

plot_log_scale

whether to plot log(readFraction) instead of readFraction (default is FALSE)

interactive

whether to return interactive plots (default is FALSE)

normalize_counts

whether to normalize the counts by the total count for each well (default is TRUE)

loss_frac_cutoff

the cutoff to use for "loss fraction" of potential partners for plots. Must be between 0 and 1. 1 or higher will not filter any clones, lower values will apply more strict filtering. (default is 0.5).

calc_fisher_pval_all

whether to calculate the Fisher exact test p-value of well overlap between the input chain and all of the filtered clones. Default is FALSE because this is computationally demanding.

Value

A list of output with the following elements:

  • data (list)

  • df_top_n - a data frame with results for the top N potential partner chains

  • df_all - a data frame with results for all potential partners passing the loss_frac_cutoff

  • input_meta - a one-row data frame with the metadata for the input chain

  • plots (list)

  • scatter - a scatter plot of read fractions for the input chain vs. the top N potential partner chains

  • r_vs_p - a point plot of correlation (r) vs. -log10(raw p-value)

  • r_vs_p_adj - a point plot of correlation (r) vs. -log10(FDR adjusted p-value)

  • rank_vs_p_adj - a Manhattan plot of chain rank vs. -log10(FDR adjusted p-value)

  • call - a list with all arguments given to the function

See also

Other well: choose_pair_manual(), get_well_subset(), get_wells_from_edges(), load_well_counts_binary(), write_well_data_to_binary()