aucprTestByFactor: Test for differential gene expression along tree using...
In farrellja/URD: URD

Test for differential gene expression along tree using binomial test

aucprTestByFactor(
  object,
  cells.1,
  cells.2,
  label,
  groups,
  log.effect.size = 0.25,
  auc.factor = 1,
  min.auc.thresh = 0.1,
  max.auc.thresh = Inf,
  frac.must.express = 0.1,
  frac.min.diff = 0,
  genes.use = NULL,
  min.groups.to.mark = 1,
  report.debug = F
)

`object`	An URD object
`cells.1`	(Character vector) Cells in population of interest
`cells.2`	(List) Cells to compare against. If a list, each entry is a population to compare against, and markers must beat all populations.
`label`	(Character) Label to use to split cells. Must be a column of `group.ids`
`groups`	(List) List of label values to include in each group.
`log.effect.size`	(Numeric) Minimum fold-change difference in mean expression to be considered differential (default, `log(2)` is 2-fold change)
`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 this factor multiplied by the expected AUC of a random classifier.
`frac.must.express`	(Numeric) Fraction of cells of interest that must express the gene to consider it differential.
`frac.min.diff`	(Numeric) Minimum difference in fraction of cells expressing a gene to consider it differential.
`genes.use`	(Character vector) Which genes to test for differential expression (default `NULL` is all genes)
`min.groups.to.mark`	(Numeric) For how many of the groups must a gene be a marker in order to be considered differential?
`report.debug`	(Logical) If `TRUE`, this function returns a list instead of a data.frame, with `$stats` containing information about all of the comparisons? (nGene, nTrans, pseudotime, n.cells) and `$marker.chain` containing the markers from each branchpoint.
`min.auc.threshold`	(Numeric) This acts as a lower bound for the AUC threshold, no matter how unbalanced the populations are.
`max.auc.threshold`	(Numeric) This acts as an upper bound for how high the AUC must be for a gene to be considered differential.