fit_curve: Output the fit curve for a given mixture of Gaussians

In fosterlab/PrInCE: Predicting Interactomes from Co-Elution

Description

For a Gaussian mixture model fit to a curve by fit_gaussians, output the fit curve using the coefficients rather than the nls object. This allows individual Gaussians to be removed from the fit model: for example, if their height is below a certain threshold, or their centres are outside the bounds of the chromatogram.

Usage

fit_curve(coef, indices)

Arguments

coef	numeric vector of coefficients for a Gaussian mixture model fit by fit_gaussians. This function assumes that the heights of the Gaussians are specified by coefficients beginning with "A" ("A1", "A2", "A3", etc.), centres are specified by coefficients beginning with "mu", and standard deviations are specified by coefficients beginning with "sigma".
indices	the indices, or x-values, to predict a fitted curve for (for example, the fractions in a given chromatogram)

Value

the fitted curve

Examples

data(scott)
chrom <- clean_profile(scott[1, ])
fit <- fit_gaussians(chrom, n_gaussians = 1)
curve <- fit_curve(fit$coefs, seq_along(chrom))

fosterlab/PrInCE documentation built on Dec. 13, 2020, 5:50 a.m.