SimulateTrueCounts: Generate both evf and gene effect and simulate true...

In YosefLab/SymSim: Simulate Single Cell RNA Sequencing Data

Generate both evf and gene effect and simulate true transcript counts

Description

Generate both evf and gene effect and simulate true transcript counts

Usage

SimulateTrueCounts(
  ncells_total,
  min_popsize,
  i_minpop = 1,
  ngenes,
  evf_center = 1,
  evf_type = "one.population",
  nevf = 10,
  phyla,
  randseed,
  n_de_evf = 0,
  vary = "s",
  Sigma = 0.4,
  geffect_mean = 0,
  gene_effects_sd = 1,
  gene_effect_prob = 0.3,
  bimod = 0,
  param_realdata = "zeisel.imputed",
  scale_s = 1,
  impulse = F,
  gene_module_prop = 0,
  prop_hge = 0.015,
  mean_hge = 5
)

Arguments

ncells_total	number of cells
min_popsize	the number of cells in the smallest population
i_minpop	specifies which population has the smallest size
ngenes	number of genes
evf_center	the value which evf mean is generated from
evf_type	string that is one of the following: 'one.population','discrete','continuous'
nevf	number of evfs
phyla	the cell developmental tree if chosing 'discrete' or 'continuous' evf type. Can either be generated randomly (using pbtree(nclusters) function from phytools package) or read from newick format file using the ape package
randseed	random seed
n_de_evf	number of differential evfs between populations
vary	which kinetic parameters should the differential evfs affect. Default is 's'. Can be "kon", "koff", "s", "all", "except_kon", "except_koff", "except_s". Suggestions are "all" or "s".
Sigma	parameter of the std of evf values within the same population
geffect_mean	the mean of the normal distribution where the non-zero gene effect sizes are sampled from
gene_effect_prob	the probability that the effect size is not 0
bimod	the amount of increased bimodality in the transcript distribution, 0 being not changed from the results calculated using evf and gene effects, and 1 being all genes are bimodal
param_realdata	pick from zeisel.imputed or NULL; zeisel.imputed means using the distribution of kinetic parameters learned from the Zeisel 2015 dataset. This option is recommended.
scale_s	a factor to scale the s parameter, which is used to tune the size of the actual cell (small cells have less number of transcripts in total)
impulse	use the impulse function when generating continuous population or not. Default is F.
gene_module_prop	proportion of genes which are in co-expressed gene module
prop_hge	the proportion of very highly expressed genes
mean_hge	the parameter to amplify the gene-expression levels of the very highly expressed genes
gene_effect_sd	the standard deviation of the normal distribution where the non-zero gene effect sizes are sampled from

Value

a list of 4 elements, the first element is true counts, second is the gene level meta information, the third is cell level meta information, including a matrix of evf and a vector of cell identity, and the fourth is the parameters kon, koff and s used to simulation the true counts

YosefLab/SymSim documentation built on Sept. 30, 2024, 2:22 p.m.