simulate_decreasing: Simulate decreasing trends

In ssokolen/metcourse: Tools for Metabolic Time-course Data

Description

Simulates decreasing trends by mixing two sigmoid curve families. Sigmoid parameters are sampled from uniform distributions with boundaries provided as input to this function. Function: y = (1 + exp((x-a)/b))**(-1).

Usage

simulate_decreasing(n, n.samples, par1, par2 = NULL, p = 0.5)

Arguments

n	Number of trends to generate.
n.samples	Number of timepoints within each trend.
par1	Parameter vector (a_min, a_max, b_min, b_max) for first set of trends
par2	Optional parameter vector (a_min, a_max, b_min, b_max) for second set of trends
p	Proportion of trends to generate using par1 vs. par2. Note, p is ignored if par2 is not specified.

Value

A dataframe with each row representing a metabolite concentration time-course scaled between 0 and 1. The corresponding x variables are assumed to be equally spaced between 0 and 1 i.e. x <- seq(0, 1, length.out=n.samples).

Examples

par1 <- c(0.2, 0.6, 0.10, 0.18)
par2 <- c(0.6, 0.9, 0.10, 0.18)
trends <- simulate_decreasing(1000, 100, par1, par2, 0.05)

# Conversion for plotting
y_mat <- t(as.matrix(trends))
x <- seq(0, 1, length.out=100)
matplot(x, y_mat, type = 'l', lty = 1, lwd = 4, col = grey(0, 0.05),
        xlab = 'Relative culturing time', ylab = 'Relative concentration')

ssokolen/metcourse documentation built on May 30, 2019, 8:43 a.m.