ct.targetSetEnrichment: Run a (limited) Pathway Enrichment Analysis on the results of...

In OscarBrock/gCrisprTools: Suite of Functions for Pooled Crispr Screen QC and Analysis

Run a (limited) Pathway Enrichment Analysis on the results of a Crispr experiment.

Description

This function enables some limited geneset enrichment-type analysis of data derived from a pooled Crispr screen using the PANTHER pathway database. Specifically, it identifies the set of targets significantly enriched or depleted in a summaryDF object returned from ct.generateResults and compares that set to the remaining targets in the screening library using a hypergeometric test.

Note that many Crispr gRNA libraries specifically target biased sets of genes, often focusing on genes involved in a particular pathway or encoding proteins with a shared biological property. Consequently, the enrichment results returned by this function represent the pathways containing genes disproportionately targeted *within the context of the screen*, and may or may not be informative of the underlying biology in question. This means that pathways not targeted by a Crispr library will obviously never be enriched within the positive target set regardless of their biological relevance, and pathways enriched within a focused library screen are similarly expected to partially reflect the composition of the library and other confounding issues (e.g., number of targets within a pathway). Analysts should therefore use this function with care. For example, it might be unsurprising to detect pathways related to histone modification within a screen employing a crispr library targeting epigenetic regulators.

This is a function that invokes the PANTHER.db Bioconductor library to extract a list of pathway mappings to be used in gene set enrichment tests. Specifically, the function returns a named list of pathways, where each element contains Entrez IDs. Users should not generally call this function directly as it is invoked internally by the higher-level ct.PantherPathwayEnrichment() function.

This function takes in a resultsDF and a vector of targets (contained in the geneID column of resultsDF) and determines whether the specified targets are enriched within the set of all significantly altered targets. It does this by iteratively testing whether targets are more likely to be among the set of enriched or depleted targets at various significance thresholds using a hypergeometric test. Note that the returned Hypergeometric P-values are not corrected for multiple testing.

Returns a list detailing the targets used in the tests, and tables indicating the results of the hypergeometric test at various significance thresholds.

Usage

ct.targetSetEnrichment(
  summaryDF,
  targets,
  enrich = TRUE,
  ignore = "NoTarget",
  ...
)

Arguments

summaryDF	A dataframe summarizing the results of the screen, returned by the function ct.generateResults. Internally coerced via 'ct.simpleResult()'.
targets	A character vector containing the names of the targets to be tested; by default these are assumed to be 'geneID's, but specifying 'collapse=geneSymbol' enables setting on 'geneSymbol' by passing that value through to 'ct.simpleResult'.
enrich	Logical indicating whether to consider guides that are enriched (default) or depleted within the screen.
ignore	Optionally, a character vector containing elements of the summaryDF that should be ignored in the analysis (e.g., unassignable or nonfunctional targets, such as nontargeting controls). By default, this function omits targets with geneSymbol 'NoTarget'.
...	Additional parameters to pass to 'ct.simpleResult'.
pvalue.cutoff	A gene-level p-value cutoff defining targets of interest within the screen. Note that this is a nominal p-value cutoff to preserve end-user flexibility.
organism	The species of the cell line used in the screen; currently only 'human' or 'mouse' are supported.
db.cut	Minimum number of genes annotated to a given to a pathway within the screen in order to consider it in the enrichment test.
species	The species of the cells used in the screen. Currently only 'human' or 'mouse' are supported.

Value

A dataframe of enriched pathways.

A named list of pathways from PANTHER.db.

A named list containing the tested target set and tables detailing the hypergeometric test results using various P-value and Q-value thresholds.

Author(s)

Russell Bainer, Steve Lianoglou

Russell Bainer, Steve Lianoglou.

Russell Bainer

Examples

data(resultsDF)
tar <-  sample(unique(resultsDF$geneSymbol), 20)
res <- ct.targetSetEnrichment(resultsDF, tar)

OscarBrock/gCrisprTools documentation built on Oct. 25, 2022, 7:29 a.m.