scNoiseSim: simulation to generate scRNA-seq data with varying level of...

Description Usage Arguments Value Examples

View source: R/simulation.R

Description

simulation to generate scRNA-seq data with varying level of gene detection noise versus gene count noise

Usage

1
scNoiseSim(zinb, celltype, disper, var_dropout = 1, var_count = 1, delta)

Arguments

zinb

a ZINB-WaVE object representing ZINB-WaVE fit to real data to get realistic simulation parameters

celltype

a factor to specify the ground-truth cell types in the original dataset that the parameter of zinb object is fit to. Since we filter out some simulated cells due to low amount of genes detected in that cell, we subset the ground truth cell types correspondingly

disper

numeric value, parameter to control the size factor r in NB(μ, r). r is varied in the set 0.5,1,5 in our simulation(as outlined in our paper)

var_dropout

numeric value, parameter to control the noise level added to a common embedding space for to generate gene detection matrix. This parameter is formulated as σ_π and in the paper is selected from the set 0.1, 0.5, 1, 2, 3

var_count

numeric value, parameter to control the noise level added to a common embedding space to generate gene count matrix. This parameter is formulated as σ_μ and and in the paper is selected from the set 0.1, 0.5, 1, 2, 3

delta

intercept to control the overall gene detection rate. and in the paper is selected from the set -2, -0.5, 1,2.5,4

Value

GeneExpr,a count matrix with rows number of genes and columns number of cells

celltype,a vector specify the corresponding celltype after QC measures.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
## raw counts matrix with rows are genes and columns are cells
data("zinb_toy",package = "scBFA", envir = environment())
## a vector specify the ground truth of cell types provided by conquer database
data("celltype_toy",package = "scBFA",envir = environment())

scData = scNoiseSim(zinb = zinb_toy,
         celltype = celltype_toy,
         disper = 1,
         var_dropout =1,
         var_count = 1,
         delta = 1)

quon-titative-biology/scBFA documentation built on March 24, 2020, 7:30 p.m.