ds.exposome_pca: Principal components analysis of an Exposome Set

In isglobal-brge/dsExposomeClient: DataSHIELD client site Exposome functions

Principal components analysis of an Exposome Set

Description

Performs a non-disclosive PCA given an Exposome Set on the study server, the Exposome Set can be subsetted by families to perform the PCA

Usage

ds.exposome_pca(
  Set,
  fam = NULL,
  scale = TRUE,
  type = "meta",
  pca = TRUE,
  npc = 10L,
  datasources = NULL
)

Arguments

Set	character Name of the ExposomeSet on the study server
fam	character vector (default NULL) Families to subset the ExposomeSet
scale	bool (default TRUE) If TRUE the ExposomeSet wil be scaled, if FALSE it will not be scaled. It is always advisable to scale the data when performing a PCA.
type	character (default "meta") Type of analysis, if meta-analysis ("meta"), a PCA to each study server will be performed. If pooled-analysis ("pooled") a pooled methodology will be performed.
pca	bool (default TRUE) If TRUE perform PCA (only numerical variables), if FALSE FAMD (numerical and categorical). This argument only affects type = "meta", the pooled methodology does not have the option of performing a FAMD analysis.
npc	integer (default 10L) Number of principal components to compute.
datasources	a list of DSConnection-class objects obtained after login

Details

The pooled PCA uses the block method for SVD decomposition. More information can be found on https://arxiv.org/pdf/1907.07410.pdf

Value

This function does not have an output. It creates a object on the study server named ds.exposome_pca.Results, this object can be passed to the ds.exposome_pca_plot to visualize the results of the PCA.

Examples

## Not run: Refer to the package Vignette for examples.

isglobal-brge/dsExposomeClient documentation built on March 5, 2024, 12:26 p.m.