dot-generateSubsets: Generate balanced subsets for divide and conquer BASiCS

In BASiCS: Bayesian Analysis of Single-Cell Sequencing data

Description

Partitions data based on either cells or genes. Attempts to find a partitioning scheme which is "balanced" for either total reads per cell across all genes (partitioning by gene) or total expression per gene across all cells (partitioning by gene). When partitioning by cell, at least 20 cells must be in each partition or BASiCS_MCMC will fail. If this partitioning fails, it will continue recursively up to a maximum number of iterations (20 by default).

Usage

.generateSubsets(
  Data,
  NSubsets,
  SubsetBy = c("cell", "gene"),
  Alpha = 0.05,
  WithSpikes = FALSE,
  MaxDepth = 20,
  .Depth = 1
)

Arguments

Data	a SingleCellExperiment object
NSubsets	Integer specifying the number of batches into which to divide Data for divide and conquer inference.
SubsetBy	Partition by "cell" or by "gene".
Alpha	p-value threshold for ANOVA testing of "balance"
WithSpikes	Similar to argument for BASiCS_MCMC - do the Data contain spikes?
MaxDepth	Maximum number of recursive
.Depth	Internal parameter. Do not set.

Value

A list of SingleCellExperiment objects

BASiCS documentation built on April 16, 2021, 6 p.m.