RADanalysis (version 1.0.1)

representative_RAD: Computes representative normalized RAD of a group of normalized RADs.

Description

Computes representative normalized RAD of a group of normalized RADs.

Usage

representative_RAD(
  norm_rad,
  sample_ids = NULL,
  plot = FALSE,
  min_rank = 1,
  confidence = 0.95,
  with_conf = TRUE,
  ...
)

Value

A list of following parameters:

$average: Contains a vector of length equal to the columns of norm_rad. This in the representative normalized RAD which is the average of normalized RAD of the group.

$population_sd: A vector of length equal to the columns of norm_rad which contains the standard deviation for each rank.

$standard_error: A vector of length equal to the columns of norm_rad which contains the standard deviation of the mean for each rank. This vector is the result of population_sd / sqrt(n), when n is the number of members of the group (length of sample_ids).

If plot = TRUE, plot of the repRAD is produced and would be added to the previous plot.

If with_conf = TRUE, confidence interval would be added to the repRAD plot.

Arguments

norm_rad

A matrix which contains the normalized RADs (samples in rows).

sample_ids

Vector of row numbers of the desired group, from which a representative RAD is going to be produced.

plot

A logical. If TRUE, plots the repRAD. The plot would be added to the previous plot.

min_rank

The minimum rank to be considered for making repRADs.

confidence

Confidence interval of plotted repRAD. Default is 0.9.

with_conf

A logical. If TRUE, plots the confidence interval in addition to repRAD. Only works when plot is TRUE.

...

Other graphical parameters to use for plotting. This function uses internally the functions lines and polygon 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_point for study the representative of groups of samples in a multi-dimensional scaling plot,

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

#plot nrads
plot(1e10,xlim = c(1,maxrank),ylim = c(2e-5,1),log="xy",
     xlab = "rank",ylab = "abundance",cex.lab = 1.5,axes = FALSE)
sfsmisc::eaxis(side = 1,at = c(1,10,100,1000,10000))
sfsmisc::eaxis(side = 2,at = c(1e-4,1e-3,1e-2,1e-1,1),las = 0)
for(i in 1:nrow(nrads)){
    points(nrads[i,],type = 'l',col = line_cols[sample_classes[i]],lwd = 0.8)
}
#plot confidence intervals of representative nrads
a <- representative_RAD(norm_rad = nrads,sample_ids = which(sample_classes == 1),
                      plot = TRUE,confidence = 0.9,with_conf = TRUE,
                      col = scales::alpha(line_cols[1],0.5),border = NA)
a <- representative_RAD(norm_rad = nrads,sample_ids = which(sample_classes == 2),
                      plot = TRUE,confidence = 0.9,with_conf = TRUE,
                      col = scales::alpha(line_cols[2],0.5),border = NA)
a <- representative_RAD(norm_rad = nrads,sample_ids = which(sample_classes == 3),
                      plot = TRUE,confidence = 0.9,with_conf = TRUE,
                      col = scales::alpha(line_cols[3],0.5),border = NA)
#plot representative nrads
a <- representative_RAD(norm_rad = nrads,sample_ids = which(sample_classes == 1),
                      plot = TRUE,with_conf = FALSE,
                      col = scales::alpha(line_cols[1],0.8),lwd = 4)
a <- representative_RAD(norm_rad = nrads,sample_ids = which(sample_classes == 2),
                      plot = TRUE,with_conf = FALSE,
                      col = scales::alpha(line_cols[2],0.8),lwd = 4)
a <- representative_RAD(norm_rad = nrads,sample_ids = which(sample_classes == 3),
                      plot = TRUE,with_conf = FALSE,
                      col = scales::alpha(line_cols[3],0.8),lwd = 4)
legend("bottomleft",bty = "n",legend = c("pre Cp","under Cp","post Cp"),
col = line_cols,lwd = 3)