In ayurovsky/rnaGinesis: Simulate gene expression data

knitr::opts_chunk$set(echo = TRUE)

R Markdown

Creates mixture data from scRNA-seq data and saves expression matrix, trueH, trueW, unmixed

Load parsed scRNA-seq data

load("~/rnaGinesis/data/muraro.RData")

Creates mu (avg W) with cell-type gene expression profiles

mu   = makeCentroid(data)

# Subsets mu to contain gene expression profiles for 3 most abundant cell types
mu   = subsetCentroid(data, mu, 3)

Creates H with cell-type proportions in each sample, dimensions cell types by samples

# Creates H with 3 cell types, Samplesize = 50
Samplesize = 50
d.params   = c("alpha"  = 0.5,
               "beta"   = 0.2,
               "duct" = 0.3)
trueH      = sim_H(S = Samplesize,
           d.params = d.params)

Obtains mixture expression matrix data and list of individual Ws for each sample

total  = 500 # Number of cells want to sample out of population of all cells in data
result = make_W(data, 
                 trueH, 
                 total)
ex     = result[[1]]
trueW  = result[[2]]

Saves expression matrix, trueW, trueH, unmixed signatures (mu)

Dataset_ResampledSingleCell         = list()
Dataset_ResampledSingleCell$ex      = ex
Dataset_ResampledSingleCell$trueW   = trueW
Dataset_ResampledSingleCell$trueH   = trueH
Dataset_ResampledSingleCell$unmixed = mu

Saves Dataset_ResampledSingleCell data to .RData file

save(Dataset_ResampledSingleCell, file = "~/rnaGinesis/data/Dataset_ResampledSingleCell.RData")

ayurovsky/rnaGinesis documentation built on Dec. 19, 2021, 6:36 a.m.