Run UMAP analysis

hd_umap() runs a UMAP analysis on the provided data. If data contain missing values, the function imputes them using the k-nearest neighbors algorithm (k = 5). The number of components to be calculated is defined by the user. The function returns a tibble with the UMAP results.

Usage

hd_umap(dat, by_sample = TRUE, components = 2, seed = 123)

Arguments

dat

An HDAnalyzeR object or a dataset in wide format and sample ID as its first column.

by_sample

If TRUE, points represent samples. If FALSE, points represent features. Default is TRUE.

components

The number of components to be calculated. Default is 10.

seed

The seed to be used in the UMAP analysis. Default is 123.

Value

A list with the UMAP results.

Examples

# Create the HDAnalyzeR object providing the data and metadata
hd_object <- hd_initialize(example_data, example_metadata)

# Run the UMAP analysis
hd_umap(hd_object, components = 2, by_sample = TRUE, seed = 123)
#> $umap_res
#> # A tibble: 586 × 3
#>    DAid     UMAP1   UMAP2
#>    <fct>    <dbl>   <dbl>
#>  1 DA00001 -2.39  -0.211 
#>  2 DA00002  1.42   2.15  
#>  3 DA00003 -1.07  -2.34  
#>  4 DA00004 -1.57  -2.31  
#>  5 DA00005 -2.71  -0.450 
#>  6 DA00006  1.88  -0.286 
#>  7 DA00007 -2.90  -1.75  
#>  8 DA00008 -0.300  1.73  
#>  9 DA00009 -0.803  0.0411
#> 10 DA00010  1.03   2.13  
#> # ℹ 576 more rows
#> 
#> $by_sample
#> [1] TRUE
#> 
#> attr(,"class")
#> [1] "hd_umap"

# Run the UMAP analysis by feature
hd_umap(hd_object, components = 2, by_sample = FALSE, seed = 123)
#> $umap_res
#> # A tibble: 100 × 3
#>    Assay   UMAP1  UMAP2
#>    <fct>   <dbl>  <dbl>
#>  1 AARSD1  3.31  -2.38 
#>  2 ABL1    1.52  -1.85 
#>  3 ACAA1   0.238 -1.50 
#>  4 ACAN   -0.289  1.69 
#>  5 ACE2   -0.307 -0.731
#>  6 ACOX1  -2.58   1.23 
#>  7 ACP5    1.00   0.229
#>  8 ACP6    1.35  -0.122
#>  9 ACTA2   1.71  -0.962
#> 10 ACTN4  -2.06   0.455
#> # ℹ 90 more rows
#> 
#> $by_sample
#> [1] FALSE
#> 
#> attr(,"class")
#> [1] "hd_umap"