Perform PCoA on a rewiring matrix.

Run classical multidimensional scaling (PCoA) on a Euclidean distance matrix derived from the ROR matrix.

Usage

perform_pcoa(mat, k = 2)

Arguments

mat

A numeric matrix from prepare_rewiring_matrix().

k

Number of dimensions. Default 2.

Value

A list with elements:

• coordinates: a tibble with isodecoder, PC1, PC2, ...

• variance_explained: numeric vector of percent variance explained for each dimension

• eigenvalues: full eigenvalue vector from stats::cmdscale()

Examples

mat <- matrix(
  c(1.5, -0.8, 0.3, 2.1, 0.5, -1.2),
  nrow = 3,
  dimnames = list(
    c("tRNA-Ala", "tRNA-Gly", "tRNA-Ser"),
    c("20_vs_34", "34_vs_58")
  )
)
perform_pcoa(mat)
#> $coordinates
#> # A tibble: 3 × 3
#>      PC1    PC2 isodecoder
#>    <dbl>  <dbl> <chr>     
#> 1  1.98   0.299 tRNA-Ala  
#> 2 -0.484 -1.03  tRNA-Gly  
#> 3 -1.50   0.726 tRNA-Ser  
#> 
#> $variance_explained
#> [1] 79.39345 20.60655
#> 
#> $eigenvalues
#> [1]  6.425576e+00  1.667757e+00 -2.664535e-15
#>