I was wondering if anybody knows of a package or some method to simulate RNA-seq sequencing reads from samples where we have difference in the variance of gene expression or more importantly, splicing as opposed to the classical difference in mean gene expression or mean splicing values.
I was thinking about getting read counts for transcripts using some of the distributions from the extraDistr package, and feeding these read counts to polyester.
library("extraDistr") library("polyester") # different read counts simulated somehow count <- c(10, 50, 100) # different alpha values depending on uniform random splicing or a dominant transcript alphas <- c(1, 2, 10) # simulate read counts count_random <- rdirmnom(10, count, alphas) # run polyester with the simulated counts simulate_experiment_countmat(fasta = NULL, gtf = NULL, seqpath = NULL, count_random, outdir = ".", paired = TRUE, seed = NULL, ...)
Anybody ever tried something similar? Are there tools for simulating reads with differential variance?