msca_p: Multilevel Simultaneous Component Analysis (MSCA-P)

In ricoderks/Rcpm: General functions for the Center of Proteomics and Metabolomics (LUMC)

Multilevel Simultaneous Component Analysis (MSCA-P)

Description

Multilevel Simultaneous Component Analysis (MSCA-P).

Usage

msca_p(X, persID)

Arguments

X	a data table with structure ((I x K) x J), see details for more information
persID	vector length (IxK) person ids (numeric e.g. for the situation with 4 samples per person 1 1 1 1 2 2 2 2 3 3 3 3, or factor)

Details

Unlike MSCA, MSCA-P does not make any assumptions about the similarities between the time-dynamic variations of different individuals, so scores and loadings will be computed separately for each individual. X needs to have the following data table structure ((I x K) x J). Where I are the individuals, J are the variables, and K are samples of an individual.

Note: msca_p can NOT handle NA's!

Based on the algorithm Jansen et al. ANALYTICA CHIMICA ACTA (2005) 530(2):173-183 Based on MatLab code available here http://www.bdagroup.nl/content/Downloads/software/software.php

Value

A data frame with the following structure :

$between$data	table used for the between-individual level
$between$scores	between-individual scores
$between$loadings	between-individual loadings
$between$percexp	percentage of variation explained in the between-individual model
$within[[i]]$data	table used for the within-individual level for individual i
$within[[i]]$scores	within-individual scores for individual i
$within[[i]]$loadings	within-individual loadings for individual i
$within[[i]]$percexp	percentage of variation explained in the within-individual model for individual i

Author(s)

E. Nevedomskaya

References

Jansen et al. ANALYTICA CHIMICA ACTA (2005) 530(2):173-183

ricoderks/Rcpm documentation built on May 18, 2022, 7:49 a.m.