met_smooth_gam: GAM smoothing

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

Description

Wraps GAM smoothing function for use in bias correction algorithm.

Usage

met_smooth_gam(time, concentration, new.time = NULL, warn = FALSE, ...)

Arguments

time	Time or sample number for metabolite time-course.
concentration	Metabolite concentration.
new.time	Optional vector for new independent variables.
warn	The gam function uses internal warning suppression that doesn't appear to work when used in conjunction with dplyr. All warning messages are suppressed by default here. Set to TRUE to unsuppress.
...	Arguments to be passed into the s() smoothing function, such as k – the dimension of the basis used to represent the smooth term. If no additional arguments are provided, a reasonable default for k (5) is assumed.

Value

A vector of smoothed concentrations.

Examples

# Simulating concave curve with 10 points
par <- c(3.5, 4.5, 2.5, 3.5, 0.0, 0.2, 0.8, 0.9)
concentration <- as.numeric(simulate_concave(1, 10, par))

# Adding noise and changing scale
concentration <- concentration * 18 + 2
abs.sd <- mean(concentration) * 0.1
concentration <- concentration + rnorm(10, 0, abs.sd)
concentration[concentration < 0] <- 0

# Original sampling time
time <- seq(0, 9, length.out=10) * 24

# New time for smoothing
new.time <- seq(0, 9, length.out=100) * 24

# Smoothing
corrected <- met_smooth_gam(time, concentration, new.time = new.time, k = 5)

# Plotting
plot(time, concentration, type='p',
     xlab='Time post inoculation (hours)', ylab='Concentration (mM)')
lines(new.time, corrected, type='l')

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