dot-getDistMat: Get Distances between 2D Points and a Loess Fit

In fmicompbio/swissknife: Handy code shared in the FMI CompBio group

Get Distances between 2D Points and a Loess Fit

Description

This function gets the euclidean (by default) distances between all points in a 2-D coordinate system (x,y) to a group of points that fall on a fitted loess curve. It uses the dist.matrix function from the wordspace package.It is used in the selVarGenes function.

Usage

.getDistMat(
  loessModel = NULL,
  x = NULL,
  y = NULL,
  x_curve = NULL,
  method = "euclidean",
  ...
)

Arguments

loessModel	the loess fit to all points.
x	the x values of all points.
y	the y values of all points.
x_curve	the x values of the points on the loess curve.
method	distance measure to be used.
...	additional parameters for the dist.matrix function.

Value

An object of class matrix and dist.matrix, containing the distances between all points to the specified points on the loess curve.

Author(s)

Dania Machlab

Examples

  library(stats)
  
  # create data
  x <- 1:1000
  y <- sort(rnorm(n = 1000, mean = 0, sd = 5)) + runif(n = 1000, min = 0, max = 10)
  lo <- loess(y ~ x)
  x_curve <- seq(range(x)[1], range(x)[2], length.out = 100)
  
  # plot
  plot(x, y, pch = 16, col = "#BEBEBE90")
  lines(x_curve, predict(lo, newdata = x_curve), col = "red")
  
  # get euclidean distance of all points to the curve (red)
  euclMat <- swissknife:::.getDistMat(loessModel = lo, x = x, y = y, x_curve = x_curve)
  euclMat[1:6, 1:6]

fmicompbio/swissknife documentation built on June 11, 2025, 4:17 p.m.