SupervisedTransformBulk {BisqueRNA}

Transforms bulk expression of a gene given overlapping data

Description

For a specific gene, this function uses linear regression to learn a transformation of the bulk expression to match the values produced by the single-cell based reference and observed single-cell based cell proportions.

Usage

SupervisedTransformBulk(gene, Y.train, X.train, X.pred)

Arguments

gene

Character string. Gene name that corresponds to row in Y.train

Y.train

Numeric Matrix. Number of gene rows by number of overlapping individuals columns. Contains weighted sum of reference profile by single-cell based cell proportion estimates for each individual

X.train

Numeric Matrix. Number of gene rows by number of overlapping individuals columns. Contains observed bulk expression for each individual

X.pred

Numeric Matrix. Number of gene rows by number of remaining individuals columns. Contains observed bulk expression for each individual to be transformed.

Details

If a linear transformation cannot be learned for a gene (zero variance in observed bulk or single-cell based weighted sums), a vector of NaNs will be returned of the expected length (length of X.pred)

Value

Y.pred Numeric Matrix. One row for given gene by number of remaining individuals columns. Contains transformed bulk expression for each individual.