cor_mutations: Correlate cell line response with gene mutations

In dwassarman/cellpanelr: Analyze Data from Cell Line Panels

Correlate cell line response with gene mutations

Description

For each of the 19.537 genes present in the data_mutations data set, perform a Mann-Whitney test wilcox.test to determine whether cell lines containing a mutant version of the gene respond differently than cell lines with the wild-type version of the gene. The Mann-Whitney test is chosen, rather than the Student's t-test, because it does not assume that the response values in the two groups are normally distributed.

Usage

cor_mutations(data, response, ids = "depmap_id", fdr = 0.05)

Arguments

data	A tibble.
response	Column containing response data
ids	Column containing DepMap IDs of cell lines
fdr	False discovery rate. Number between 0 and 1 representing the likelihood that a gene predicted to be significant is actually a false-positive.

Value

A tibble with 19,537 rows and 4 columns. Each row contains the correlation values for a single gene.

gene

Hugo gene symbol

effect

The relative increase or decrease in cell line response associated with mutation of the gene. Calculated as log2( mutant response / wildtype response )

p.value

Probability that the null hypothesis is true (there is no relationship between gene mutation and cell line response)

significant

Whether the correlation is deemed significant after multiple hypothesis correction with the given false discovery rate

Examples

# Setup example data set
df <- tibble::tibble(
  CellLine = c("LS513", "253-J", "NIH:OVCAR-3"),
  DepMapID = c("ACH-000007", "ACH-000011", "ACH-000001"),
  logIC50 = c(-2.8, -4.04, -6.23)
)

cor_mutations(
  data = df,
  response = "logIC50",
  ids = "DepMapID",
  fdr = 0.01
)

dwassarman/cellpanelr documentation built on Jan. 3, 2023, 8:27 a.m.