im_syng_tcga: Calculation and statistical assessment of synergistic...

In korkutlab/imogimap: What the Package Does (One Line, Title Case)

Calculation and statistical assessment of synergistic associations using TCGA data.

Description

Takes a list of Tumor-intrinsic pathway(TIP) genes and returns their combinatorial association with immune checkpoint(ICP) genes by evaluating their synergistic impact on immune-associated phenotypes(IAP) using RNASeq2GeneNorm expressions as provided by curatedTCGAData.

Usage

im_syng_tcga(
  onco_gene,
  icp_gene,
  cohort,
  select_iap,
  method,
  ndatamin = 8,
  specificity,
  N_iteration_specificity = 1000,
  sensitivity,
  N_iteration_sensitivity = 100,
  sample_list,
  add_receptor_ligand = FALSE
)

Arguments

onco_gene	A character vector of gene Hugo symbols.
icp_gene	An optional character vector of immune checkpoint gene Hugo symbols.
cohort	A list of TCGA diseases
select_iap	An optional character vector or numeric matrix or data.frame.
method	A character string indicating which synergy score to be used. one of "max" or "independence". Default is "max".
ndatamin	minimum number of samples. Synergy score calculation will be skipped for matrices with number of rows less than ndatamin
specificity	An optional logical indicating if specificity analysis should be done. Default is FALSE.
N_iteration_specificity	Number of iterations for random sampling for specificity p.value calculation. Default is 1000.
sensitivity	An optional logical indicating if a sensitivity analysis should be done. Default is FALSE.
N_iteration_sensitivity	Number of iterations for random sampling for sensitivity analysis. Default is 100.
sample_list	An optional character vector of TCGA samples barcodes indicating a subset of samples within a cohort.
add_receptor_ligand	An optional logical indicating whether receptor_ligands pair should be added. Default is FALSE.

Details

If no icp_gene is specified, the default icp_gene_list will be used. If add_receptor_ligand is TRUE, CellphoneDB database is used to add genes with known receptor-ligand interactions with either of icp_genes or onco_genes. If select_iap is a character vector, it must be any sub-list of IAP names as listed in TCGA_immune_features_list. If a numeric matrix or data.frame, each column represents a user-defined IAP and must have a range between [0,1]. If select_iap is missing all IAPs listed in TCGA_immune_features_list will be considered for analysis.

For synergy score calculations all features are normalized to be on [0,1] range. For details of synergy score and significance pvalue calculations see find_a_synergy function.

A specificity p.value is computed using random sampling with replacement from two null models, generated from one of the two genes against a set of genes randomly selected from the genome. Two P-values are calculated for the synergistic interaction of the pair against the two null models. The highest of the two P-values is used to assess the specificity of the interaction against the whole genome. The number of randomly selected genes in each null model is determined by N_iteration_specificity.

Sensitivity (Robustness) score defined as normalized root mean square deviation of scores calculated over 70% of samples, selected via random sampling. The number of sub-sample iterations is determined by N_iteration_sensitivity.

All barcodes in sample_list must be 15 character long and belong to the same cohort. When sample_list is provided, cohort should be the disease cohort that they belong to, otherwise only the first element of the cohort list will be used.

Value

A data.frame of synergy scores and bootstrapping p.values.

Examples

im_syng_tcga(onco_gene=c("TGFB1","SERPINB9"), cohort=c("ucec"))

korkutlab/imogimap documentation built on March 17, 2023, 8:22 a.m.