markersAUCPR: Test for markers of a population using a precision-recall...

In farrellja/URD: URD

Description

Test for markers of a population using a precision-recall curve

Usage

markersAUCPR(
  object,
  clust.1 = NULL,
  clust.2 = NULL,
  cells.1 = NULL,
  cells.2 = NULL,
  clustering = NULL,
  effect.size = 0.25,
  frac.must.express = 0.1,
  exp.thresh = 0,
  frac.min.diff = 0,
  auc.factor = 1,
  max.auc.threshold = 1,
  genes.use = NULL
)

Arguments

object	An URD object
clust.1	(Character)
clust.2	(Character)
cells.1	(Character vector)
cells.2	(Character vector)
clustering	(Character) Name of a clustering (i.e. a column in @group.ids)
effect.size	(Numeric) Minimum log fold-change for two genes to be considered differentially expressed
frac.must.express	(Numeric) Gene must be expressed in at least this fraction of cells in one of the two clusters to be considered.
exp.thresh	(Numeric) Minimum expression value to consider 'expressed.'
frac.min.diff	(Numeric) Fraction of cells expressing the gene must be at least this different between two populations to be considered. @param auc.factor (Numeric) The precision-recall AUC is determined for a random classifier is determined based on the size of populations. To be considered differential, genes must have an AUC that is auc.factor times the expected AUC for a random classifier.
max.auc.threshold	(Numeric) This acts as an upper bound for how high the AUC must be for a gene to be considered differential (i.e. you can choose that no matter the expectation for a random classifier, any gene with an AUC > 0.9 is considered differential).
genes.use	(Character vector) Genes to compare, default is NULL (all genes)

Value

(data.frame)

farrellja/URD documentation built on June 17, 2020, 4:48 a.m.