RADnormalization: Normalizes an abundance vector to the desired number of...

In RADanalysis: Normalization and Study of Rank Abundance Distributions

Description

Normalizes an abundance vector to the desired number of ranks.

Usage

RADnormalization(input, max_rank, average_over = 1, min_rank = 1,
  labels = FALSE, count_data = TRUE, method = "upperlimit")

Arguments

input	A vector which contains the abundance values (an abundance vector).
max_rank	The desired rank to which this method normalizes the input.
average_over	Number of times, a normalized RAD is created and averaged to produce the result.
min_rank	The minimum rank to which this method normalizes the input.
labels	A logical. If TRUE the label of each rank (ids in the input vector) will be returned.
count_data	A logical. TRUE means that the input vector contains counts (integer values) otherwise contains the relative abundances. In the current version only counts are accepted.
method	Sets the stop criterion for normalization. This should be one of "lowerlimit", "middle" or "upperlimit". Method affects the final result. lowerlimit: Samples from species pool one by one, until reaches max_rank. middle: Samples from species pool with random size until the sampled vector has desired ranks (max_rank). upperlimit: Removes from species pool one by one, until reaches max_rank.

Value

A list of following items:

$norm_rad: Normalized RAD sum up to 1. If labels = TRUE, it would also contain the labels.

$norm_rad_count: A matrix of average_over rows and max_rank columns. Each row contains one normalized RAD. These normalized RADs are averaged and sum up to 1 in order to make norm_rad

$norm_rad_mean_sd: Standard deviation of the mean for all the ranks in norm_rad. This vector is created using the values in norm_rad_count

$inputs: A list which contains inputs used for creating normalized RADs.

See Also

RADnormalization_matrix for normalize an entire otutable, representative_point for study the representative of groups of samples in a multi-dimensional scaling plot, representative_RAD for study the representative of group of norm rads.

Examples

data("gut_otu_table")
rads <- gut_otu_table
original_rad <- sort(rads[1,],decreasing = TRUE)
#removing zeros
original_rad <- original_rad[original_rad > 0]
plot(original_rad,ylim = c(1,max(original_rad)),log = "xy", xlab = "Rank",ylab = "Abundance",
     main = "RAD of first sample",pch = 19,type = "b",cex = 0.5)
print(paste("number of ranks present in the original rad is:",length(original_rad)))
norm_rad <- RADnormalization(input = rads[1,],max_rank = 500,average_over = 50)
points(x = norm_rad$norm_rad * sum(norm_rad$norm_rad_count[1,]) ,pch = 19,cex = 1, type = "l",
       col = "blue",lwd = 4)
points(x = norm_rad$norm_rad_count[1,],pch = 19,cex = 1, type = "l",col = "red",lwd = 3)
points(x = norm_rad$norm_rad_count[2,],pch = 19,cex = 1, type = "l",col = "green",lwd = 3)
legend("bottomleft",legend = c("Original RAD","possible norm rad","possible norm rad",
                               paste("nrad averaged over 50 realizations, times",
                               sum(norm_rad$norm_rad_count[1,]))),
       col = c("black","red","green","blue"),lwd = 2,bty = "n")

Example output

[1] "number of ranks present in the original rad is: 1116"

RADanalysis documentation built on May 2, 2019, 6:13 a.m.