fitXYCurve: Fitting a smooth curve through paired (x,y) data

In aroma.light: Light-Weight Methods for Normalization and Visualization of Microarray Data using Only Basic R Data Types

Description

Fitting a smooth curve through paired (x,y) data.

Usage

## S3 method for class 'matrix'
fitXYCurve(X, weights=NULL, typeOfWeights=c("datapoint"), method=c("loess", "lowess",
  "spline", "robustSpline"), bandwidth=NULL, satSignal=2^16 - 1, ...)

Arguments

X	An Nx2 matrix where the columns represent the two channels to be normalized.
weights	If NULL, non-weighted normalization is done. If data-point weights are used, this should be a vector of length N of data point weights used when estimating the normalization function.
typeOfWeights	A character string specifying the type of weights given in argument weights.
method	character string specifying which method to use when fitting the intensity-dependent function. Supported methods: "loess" (better than lowess), "lowess" (classic; supports only zero-one weights), "spline" (more robust than lowess at lower and upper intensities; supports only zero-one weights), "robustSpline" (better than spline).
bandwidth	A double value specifying the bandwidth of the estimator used.
satSignal	Signals equal to or above this threshold will not be used in the fitting.
...	Not used.

Value

A named list structure of class XYCurve.

Missing values

The estimation of the function will only be made based on complete non-saturated observations, i.e. observations that contains no NA values nor saturated values as defined by satSignal.

Weighted normalization

Each data point, that is, each row in X, which is a vector of length 2, can be assigned a weight in [0,1] specifying how much it should affect the fitting of the normalization function. Weights are given by argument weights, which should be a numeric vector of length N.

Note that the lowess and the spline method only support zero-one {0,1} weights. For such methods, all weights that are less than a half are set to zero.

Details on loess

For loess, the arguments family="symmetric", degree=1, span=3/4, control=loess.control(trace.hat="approximate", iterations=5, surface="direct") are used.

Author(s)

Henrik Bengtsson

Examples

 # Simulate data from the model y <- a + bx + x^c + eps(bx)
x <- rexp(1000)
a <- c(2,15)
b <- c(2,1)
c <- c(1,2)
bx <- outer(b,x)
xc <- t(sapply(c, FUN=function(c) x^c))
eps <- apply(bx, MARGIN=2, FUN=function(x) rnorm(length(x), mean=0, sd=0.1*x))
Y <- a + bx + xc + eps
Y <- t(Y)

lim <- c(0,70)
plot(Y, xlim=lim, ylim=lim)

# Fit principal curve through a subset of (y_1, y_2)
subset <- sample(nrow(Y), size=0.3*nrow(Y))
fit <- fitXYCurve(Y[subset,], bandwidth=0.2)

lines(fit, col="red", lwd=2)

# Backtransform (y_1, y_2) keeping y_1 unchanged
YN <- backtransformXYCurve(Y, fit=fit)
points(YN, col="blue")
abline(a=0, b=1, col="red", lwd=2)

aroma.light documentation built on Nov. 8, 2020, 4:56 p.m.