correct_counts: Correct data by setting all latent factors to their median...

In sctransform: Variance Stabilizing Transformations for Single Cell UMI Data

Correct data by setting all latent factors to their median values and reversing the regression model

Description

This version does not need a matrix of Pearson residuals. It takes the count matrix as input and calculates the residuals on the fly. The corrected UMI counts will be rounded to the nearest integer and negative values clipped to 0.

Usage

correct_counts(
  x,
  umi,
  cell_attr = x$cell_attr,
  scale_factor = NA,
  verbosity = 2
)

Arguments

x	A list that provides model parameters and optionally meta data; use output of vst function
umi	The count matrix
cell_attr	Provide cell meta data holding latent data info
scale_factor	Replace all values of UMI in the regression model by this value. Default is NA which uses median of total UMI as the latent factor.
verbosity	An integer specifying the verbosity level: 0 (silent, no messages), 1 (show messages only), or 2 (show messages and progress bars); default is 2

Value

Corrected data as UMI counts

Examples

vst_out <- vst(pbmc, return_cell_attr = TRUE)
umi_corrected <- correct_counts(vst_out, pbmc)

sctransform documentation built on Jan. 10, 2026, 9:17 a.m.