RADanalysis (version 1.0.1)

representative_point: Computes representative point based on the coordinates of points which are in the same group.

Description

Computes representative point based on the coordinates of points which are in the same group.

Usage

representative_point(
  input,
  ids = NULL,
  coord_names = c(1, 2),
  standard_error_mean = TRUE,
  plot = FALSE,
  ...
)

Value

A list of following parameters:

$mean: Contains the average of points. A vector with the length of coordinates used for computing the average. These coordinates are preset in coord_names.

$sd: A vector with a length similar to mean which contains the standard deviation for each coordinate.

$mean_standard_error: A vector with a length similar to mean which contain the standard deviation of the mean for each coordinate. This vector is the result of sd / sqrt(n), when n is the number of members of the group (length of sample_ids).

If plot = TRUE, representative points would be added to the previous plot.

If standard_error_mean = TRUE, the standard error of the mean would be added to the representative points.

Arguments

input

A matrix which contains the coordinates of samples. Usually this is the result of ordination of normalized RADs using multi-dimensional scaling (cmdscale). In the input matrix each row contains vector of coordinates of one sample.

ids

Vector of row numbers of the desired group, from which a representative point is going to be represented

coord_names

A vector which contains the coordintes number that should be used to create representative point. Default is c(1,2).

standard_error_mean

A logical. If TRUE, uses the standard error of the mean and plot it with representative points. It works only if plot = TRUE.

plot

A logical. If TRUE, shows the representative points on the previous plot.

...

other graphical parameters to use for plotting. This function uses internally the functions points and arrows to plot.

See Also

RADnormalization for normalize an abundance vector. This function return more details compared to RADnormalization_matrix, RADnormalization_matrix for normalize an entire otutable, representative_RAD for study the representative of group of norm rads.

Examples

line_cols <- c("green","red","blue")
sample_classes <- c(1,1,1,1,2,2,3,3,1,1,2,3,3,1,1,2,3,3)
maxrank <- 400
data("gut_nrads")
nrads <- gut_nrads
nrads <- nrads$norm_matrix

#distance matrix using manhattan distance
d <- dist(x = nrads,method = "manhattan")
#ordination using classical multi-dimensional scaling
mds <- cmdscale(d = d,k = 5,eig = TRUE)

#plot the points
plot(mds$points,xlab = "First coordinate",ylab = "Second coordinate",pch = 19,cex =1,
    col = line_cols[sample_classes],
    main = "MDS plot with representative points \n of each group and error bars")

#add the representative points wit erorr bar to the previous plot
a <- representative_point(input = mds$points,ids = which(sample_classes == 1),
    col = scales::alpha(line_cols[1],0.5),
    plot = TRUE,standard_error_mean = TRUE,pch = 19, cex = 4)
a <- representative_point(input = mds$points,ids = which(sample_classes == 2),
    col = scales::alpha(line_cols[2],0.5),
    plot = TRUE,standard_error_mean = TRUE,pch = 19, cex = 4)
a <- representative_point(input = mds$points,ids = which(sample_classes == 3),
    col = scales::alpha(line_cols[3],0.5),
    plot = TRUE,standard_error_mean = TRUE,pch = 19, cex = 4)

legend("bottomleft",bty = "n",legend = c("pre Cp","under Cp","post Cp"),
    col = line_cols,pch = 19)