impute_zeros: Impute Rounded Zeros

In warrenmcg/sleuth-ALR: Compositional Approach to Sleuth Data Analysis

Description

This uses various strategies to impute zeros that are considered "rounded zeros", that is, values that are zero because the true value is positive but below the detection limit (i.e. the sequencing depth). See the references for more information.

Usage

impute_zeros(mat, method = "multiplicative", delta = NULL,
  impute_proportion = 0.65)

Arguments

mat	a numeric matrix to be imputed, with N targets and M samples. This method assumes that N > M. N x M and M x N matrices are both valid inputs.
method	the choice of how to impute the rounded zeros. only "multiplicative" and "additive" is supported at this time.
delta	the value to impute; if NULL, delta = impute_proportion x (the minimum value in a sample)
impute_proportion	the proportion of the minimum value that should be the delta value (default = 65%).

Value

a matrix with the same dimensions as input and all zero values imputed by sample.

warrenmcg/sleuth-ALR documentation built on Oct. 27, 2020, 4:30 a.m.