hd_pca() runs a PCA 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 PCs
to be calculated is defined by the user. The function returns a tibble with the PCA results.
Arguments
- dat
An HDAnalyzeR object or a dataset in wide format and sample ID as its first column.
- components
The number of PCs to be calculated. Default is 10.
- by_sample
If TRUE, points represent samples. If FALSE, points represent features. Default is TRUE.
- seed
The seed to be used in the PCA analysis. Default is 123.
Examples
# Create the HDAnalyzeR object providing the data and metadata
hd_object <- hd_initialize(example_data, example_metadata)
# Run the PCA analysis
hd_pca(hd_object, components = 5, by_sample = TRUE, seed = 123)
#> $pca_res
#> # A tibble: 586 × 6
#> DAid PC1 PC2 PC3 PC4 PC5
#> <fct> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 DA00001 -3.67 -4.28 -2.34 -3.10 -2.65
#> 2 DA00002 4.11 -2.64 2.04 -0.441 -4.43
#> 3 DA00003 -3.34 4.72 1.41 0.881 -0.561
#> 4 DA00004 -4.78 0.443 1.41 0.107 -1.10
#> 5 DA00005 -4.98 -3.67 0.711 -5.70 -0.807
#> 6 DA00006 0.395 0.0572 -1.90 -7.75 0.707
#> 7 DA00007 -10.5 -2.91 -0.382 -0.841 -1.61
#> 8 DA00008 2.64 -2.01 2.75 -0.128 -0.852
#> 9 DA00009 -1.79 -0.461 2.79 -2.61 -2.71
#> 10 DA00010 3.57 0.821 -0.00756 1.50 -2.66
#> # ℹ 576 more rows
#>
#> $pca_loadings
#> # A tibble: 10,000 × 3
#> terms value component
#> <chr> <dbl> <chr>
#> 1 AARSD1 -0.133 PC1
#> 2 ABL1 -0.198 PC1
#> 3 ACAA1 -0.163 PC1
#> 4 ACAN 0.0123 PC1
#> 5 ACE2 -0.0583 PC1
#> 6 ACOX1 -0.135 PC1
#> 7 ACP5 -0.0626 PC1
#> 8 ACP6 -0.0933 PC1
#> 9 ACTA2 -0.0751 PC1
#> 10 ACTN4 -0.0424 PC1
#> # ℹ 9,990 more rows
#>
#> $pca_variance
#> # A tibble: 5 × 3
#> component percent_variance cumulative_percent_variance
#> <int> <dbl> <dbl>
#> 1 1 15.7 15.7
#> 2 2 6.72 22.4
#> 3 3 3.73 26.2
#> 4 4 3.31 29.5
#> 5 5 2.90 32.4
#>
#> $by_sample
#> [1] TRUE
#>
#> attr(,"class")
#> [1] "hd_pca"
# Run the PCA analysis by feature
hd_pca(hd_object, components = 5, by_sample = FALSE, seed = 123)
#> $pca_res
#> # A tibble: 100 × 6
#> Assay PC1 PC2 PC3 PC4 PC5
#> <fct> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 AARSD1 -39.5 6.28 0.0147 -2.84 6.05
#> 2 ABL1 -14.0 -7.20 -7.91 -5.03 -0.464
#> 3 ACAA1 -0.858 -7.96 -3.86 11.6 -1.02
#> 4 ACAN 6.40 4.48 1.87 -0.667 -1.10
#> 5 ACE2 1.22 3.49 -5.89 9.29 -1.86
#> 6 ACOX1 8.36 -7.34 1.39 1.88 0.806
#> 7 ACP5 0.214 2.36 0.0138 2.68 -0.182
#> 8 ACP6 -3.40 1.38 -0.348 -0.757 -1.98
#> 9 ACTA2 -11.3 5.26 -4.83 -1.78 -5.80
#> 10 ACTN4 9.90 1.23 0.842 -0.685 -0.0497
#> # ℹ 90 more rows
#>
#> $pca_loadings
#> # A tibble: 58,600 × 3
#> terms value component
#> <chr> <dbl> <chr>
#> 1 DA00001 -0.0373 PC1
#> 2 DA00002 -0.0225 PC1
#> 3 DA00003 -0.0429 PC1
#> 4 DA00004 -0.0451 PC1
#> 5 DA00005 -0.0414 PC1
#> 6 DA00006 -0.0393 PC1
#> 7 DA00007 -0.0390 PC1
#> 8 DA00008 -0.0342 PC1
#> 9 DA00009 -0.0437 PC1
#> 10 DA00010 -0.0420 PC1
#> # ℹ 58,590 more rows
#>
#> $pca_variance
#> # A tibble: 5 × 3
#> component percent_variance cumulative_percent_variance
#> <int> <dbl> <dbl>
#> 1 1 46.9 46.9
#> 2 2 6.18 53.1
#> 3 3 3.01 56.1
#> 4 4 2.37 58.5
#> 5 5 2.29 60.8
#>
#> $by_sample
#> [1] FALSE
#>
#> attr(,"class")
#> [1] "hd_pca"