fit_zi_model: Fit a zero-inflated model for a single gene

In switchde: Switch-like differential expression across single-cell trajectories

Description

Fits a zero-inflated sigmoidal model for a single gene vector, returning MLE model parameters and p-value.

Usage

fit_zi_model(y, pst, maxiter = 10000, log_lik_tol = 0.001,
  verbose = FALSE)

Arguments

y	Vector of gene expression values
pst	Pseudotime vector, of same length as y
maxiter	Maximum number of iterations for EM algorithm if zero inflation enabled. Default 100
log_lik_tol	If the change in the log-likelihood falls below this for zero inflated EM the algorithm is assumed to have converged
verbose	Print convergence update for EM algorithm

Value

A vector with 6 entries: maximum likelihood estimates for μ_0, k t0, λ, σ^2 and a p-value

Examples

data(synth_gex)
data(ex_pseudotime)
y <- synth_gex[1, ]
fit <- fit_zi_model(y, ex_pseudotime)

switchde documentation built on Nov. 8, 2020, 5:19 p.m.