gower: Gower

In ecodive: Parallel and Memory-Efficient Ecological Diversity Metrics

Gower

Description

Gower beta diversity metric.

Usage

gower(counts, weighted = TRUE, pairs = NULL, cpus = n_cpus())

Arguments

counts	An OTU abundance matrix where each column is a sample, and each row is an OTU. Any object coercible with as.matrix() can be given here, as well as phyloseq, rbiom, SummarizedExperiment, and TreeSummarizedExperiment objects.
weighted	If TRUE, the algorithm takes relative abundances into account. If FALSE, only presence/absence is considered.
pairs	Which combinations of samples should distances be calculated for? The default value (NULL) calculates all-vs-all. Provide a numeric or logical vector specifying positions in the distance matrix to calculate. See examples.
cpus	How many parallel processing threads should be used. The default, n_cpus(), will use all logical CPU cores.

Value

A dist object.

Calculation

Each row (OTU) of counts is rescaled to the range 0-1. In cases where a row is all the same value, those values are replaced with 0.

counts                 scaled recounts
     A B C  D                 A   B   C D  
OTU1 0 0 0  0    ->    OTU1 0.0 0.0 0.0 0
OTU2 0 8 9 10    ->    OTU2 0.0 0.8 0.9 1
OTU3 5 5 5  5    ->    OTU3 0.0 0.0 0.0 0
OTU4 2 0 0  0    ->    OTU4 1.0 0.0 0.0 0
OTU5 4 6 4  1    ->    OTU5 0.6 1.0 0.6 0

In the formulas below, x and y are two columns (samples) from the scaled counts. n is the number of rows (OTUs) in counts.

D = \displaystyle \frac{1}{n}\sum_{i = 1}^{n} |x_i - y_i|

  x <- c(0, 0, 0, 1, 0.6)
  y <- c(0, 0.8, 0, 0, 1)
  sum(abs(x-y)) / length(x)  
  #>  0.44

References

Gower JC 1971. A general coefficient of similarity and some of its properties. Biometrics. 27(4). \Sexpr[results=rd]{tools:::Rd_expr_doi("10.2307/2528823")}

Gower JC, Legendre P 1986. Metric and Euclidean Properties of Dissimilarity Coefficients. Journal of Classification. 3. \Sexpr[results=rd]{tools:::Rd_expr_doi("10.1007/BF01896809")}

See Also

Other beta_diversity: bray_curtis(), canberra(), euclidean(), generalized_unifrac(), jaccard(), kulczynski(), manhattan(), unweighted_unifrac(), variance_adjusted_unifrac(), weighted_normalized_unifrac(), weighted_unifrac()

Examples

    # Example counts matrix
    ex_counts
    
    # Gower weighted distance matrix
    gower(ex_counts)
    
    # Gower unweighted distance matrix
    gower(ex_counts, weighted = FALSE)
    
    # Only calculate distances for A vs all.
    gower(ex_counts, pairs = 1:3)
    

ecodive documentation built on Aug. 23, 2025, 1:13 a.m.