ZeroData: Zero replacement

In CoDaLoMic: Compositional Models to Longitudinal Microbiome Data

Zero replacement

Description

In this function the zeros are removed or replaced using functions of "zCompositions" package that can be used with longitudinal data (because they do not use the information of other rows to make the replacement).

Usage

ZeroData(DaTa, method = "multKM", seed = NULL)

Arguments

DaTa	data.frame. The first column contains the time point information (natural numbers 1,2,3...). The rest of the columns contain the relative abundance of each bacteria at the different time points. The values of each column must sum 1.
method	Character. If method="multKM" - The replacement is carried out with the "multiplicative Kaplan-Meier smoothing spline replacement" (Palarea-Albaladejo and Martín-Fernandez, 2015). Default method. The zeros must be written with a 0. If method="multRepl" - The replacement is carried out with the "multiplicative simple replacement" (Palarea-Albaladejo and Martín-Fernandez, 2015). The zeros must be written with a 0. If method="nozeros" - The bacteria that contains zeros are removed. One column is added to the dataset called "Other".
seed	Number. Set a seed. Default seed=NULL.

Value

The dataset without zeros.

References

Palarea-Albaladejo J. and Martín-Fernandez JA. zCompositions – R package for multivariate imputation of left-censored data under a compositional approach. Chemometrics and Intelligent Laboratory Systems 2015; 143: 85-96.

Examples

set.seed(2)
dat=gtools::rdirichlet(6,c(1,2,3,1,2,3))
dat2=dat
dat2[2,1]=0
dat2[2,2]=dat[2,1]+dat[2,2]
dat2[4,3]=0
dat2[4,4]=dat[4,3]+dat[4,4]

X <- cbind( c(1:6) ,dat2)

Final=ZeroData(X,"multKM",1)
Final2=ZeroData(X,"multRepl",1)

CoDaLoMic documentation built on April 12, 2025, 2:18 a.m.