Principal Components Analysis Calculation

Source: R/pca.R

[Experimental]

The calc_pca() function performs principal components analysis of the gene count vectors across all samples.

A corresponding autoplot() method then can visualize the results.

calc_pca(object, assay_name = "counts", n_top = NULL)

Arguments

object

(AnyHermesData)
input.

assay_name

(string)
name of the assay to use.

n_top

(count or NULL)
filter criteria based on number of genes with maximum variance.

Value

A HermesDataPca object which is an extension of the stats::prcomp class.

Details

  • PCA should be performed after filtering out low quality genes and samples, as well as normalization of counts.

  • In addition, genes with constant counts across all samples are excluded from the analysis internally in calc_pca(). Centering and scaling is also applied internally.

  • Plots can be obtained with the ggplot2::autoplot() function with the corresponding method from the ggfortify package to plot the results of a principal components analysis saved in a HermesDataPca object. See ggfortify::autoplot.prcomp() for details.

See also

Afterwards correlations between principal components and sample variables can be calculated, see pca_cor_samplevar.

Examples

object <- hermes_data %>%
  add_quality_flags() %>%
  filter() %>%
  normalize()

result <- calc_pca(object, assay_name = "tpm")
summary(result)
#> Importance of first k=18 (out of 19) components:
#>                            PC1     PC2     PC3      PC4      PC5     PC6
#> Standard deviation     22.9971 18.7315 16.3042 13.47009 13.05843 11.4881
#> Proportion of Variance  0.2212  0.1467  0.1112  0.07589  0.07132  0.0552
#> Cumulative Proportion   0.2212  0.3679  0.4791  0.55500  0.62632  0.6815
#>                             PC7     PC8     PC9    PC10    PC11    PC12    PC13
#> Standard deviation     10.60653 9.67291 9.29607 8.97324 8.54474 8.11786 7.70847
#> Proportion of Variance  0.04705 0.03913 0.03614 0.03368 0.03054 0.02756 0.02485
#> Cumulative Proportion   0.72857 0.76770 0.80384 0.83752 0.86805 0.89562 0.92047
#>                           PC14    PC15    PC16    PC17    PC18
#> Standard deviation     7.20798 6.91976 6.11309 5.77360 4.42960
#> Proportion of Variance 0.02173 0.02003 0.01563 0.01394 0.00821
#> Cumulative Proportion  0.94220 0.96222 0.97785 0.99179 1.00000

result1 <- calc_pca(object, assay_name = "tpm", n_top = 500)
summary(result1)
#> Importance of first k=18 (out of 19) components:
#>                            PC1    PC2     PC3     PC4     PC5     PC6     PC7
#> Standard deviation     11.2652 9.6518 6.89353 6.34548 5.29254 5.20812 4.70857
#> Proportion of Variance  0.2538 0.1863 0.09504 0.08053 0.05602 0.05425 0.04434
#> Cumulative Proportion   0.2538 0.4401 0.53517 0.61570 0.67172 0.72597 0.77031
#>                           PC8     PC9   PC10    PC11   PC12    PC13    PC14
#> Standard deviation     4.2012 4.06409 3.7884 3.51647 3.2405 3.07753 3.04570
#> Proportion of Variance 0.0353 0.03303 0.0287 0.02473 0.0210 0.01894 0.01855
#> Cumulative Proportion  0.8056 0.83864 0.8673 0.89208 0.9131 0.93202 0.95058
#>                           PC15    PC16    PC17    PC18
#> Standard deviation     2.87077 2.65226 2.46289 1.83598
#> Proportion of Variance 0.01648 0.01407 0.01213 0.00674
#> Cumulative Proportion  0.96706 0.98113 0.99326 1.00000

# Plot the results.
autoplot(result)

autoplot(result, x = 2, y = 3)

autoplot(result, variance_percentage = FALSE)

autoplot(result, label = TRUE, label.repel = TRUE)