Get binary (aka beta) score

Returns a beta score which indicates the binaryness of a gene across clusters. High scores (near 1) indicate that a gene is either on or off in nearly all cells of every cluster. Scores near 0 indicate a cells is non-binary (e.g., not expressed, ubiquitous, or randomly expressed). This value is used for gene filtering prior to defining clustering.

getBetaScore(propExpr, returnScore = TRUE, spec.exp = 2)

Arguments

propExpr

a matrix of proportions of cells (rows) in a given cluster (columns) with CPM/FPKM > 1 (or 0, HCT uses 1)

returnScore

if TRUE returns the score, if FALSE returns the ranks

spec.exp

scaling factor (recommended to leave as default)

Value

returns a numeric vector of beta score (or ranks)