Options for normalize_counts(). More...

#include <normalize_counts.hpp>

Public Attributes
double	pseudo_count = 1

bool	preserve_sparsity = false

bool	log = true

double	log_base = 2

Detailed Description

Options for normalize_counts().

Member Data Documentation

◆ pseudo_count

double scran_norm::NormalizeCountsOptions::pseudo_count = 1

Pseudo-count to add to each value prior to log-transformation. All values should be positive to ensure that log-transformed values are finite. The default value of 1 preserves sparsity in the log-count matrix. Larger values shrink the differences between cells towards zero, reducing spurious differences (but also signal) at low counts - see choose_pseudo_count_raw() for comments. Ignored if NormalizeCountsOptions::log = false.

◆ preserve_sparsity

bool scran_norm::NormalizeCountsOptions::preserve_sparsity = false

Whether to preserve sparsity for non-unity pseudo-counts. If true, we multiply the size factors by the NormalizeCountsOptions::pseudo_count and add 1 before log-transformation. This does not change the differences between entries of the resulting matrix, and adding log(pseudo_count/log(log_base) will recover the expected log-count values. Ignored if NormalizeCountsOptions::log = false or NormalizeCountsOptions::pseudo_count = 1.

◆ log

bool scran_norm::NormalizeCountsOptions::log = true

Whether to log-transform the normalized counts in the output matrix.

◆ log_base

double scran_norm::NormalizeCountsOptions::log_base = 2

Base for the log-transformation. Only used if NormalizeCountsOptions::log = true.

The documentation for this struct was generated from the following file: