fitLmm: Fit linear mixed-effects models (LMMs)

In kwanjeeraw/lmm2met: Linear mixed-effects modeling to improve clinical metabolomics information

Description

Fit a linear mixed-effects model (LMM) to each measured variable (e.g. metabolite) via lmer function of lme4.

Usage

fitLmm(fix, random, data, start, end=NULL, auto=FALSE, major=NULL, pval=0.05, ...)

Arguments

fix	list or vector; list or vector of fixed effects
random	list or vector; list or vector of random effects
data	dataframe; a dataframe contains metadata and measured variables
start	integer; index of the first measured variable
end	integer; index of the last measured variable, If not given, all variables are fitted
auto	logical; automatically include only significant fixed effects based on chi-square test, defaut = FALSE
major	list or vector; list or vector of fixed effects always be included, if provided, auto is = TRUE
pval	numeric; value of significance level for chi-square test, defaut = 0.05
...	other arguments of lmer function

Details

The function processes a given data matrix (e.g. metabolomics data) by fitting a LMM to each measured variable via lmer function of lme4. Significance level of the fixed-effect term is assessed by chi-square test as implemented in drop1. If auto = FALSE, all fixed effects are included in models. If auto = TRUE, only significant fixed effects with Pr(Chi) < pval, are included in models. If major is provided, major-fixed effects and significant fixed effects with Pr(Chi) < pval are included in models. Random effect must be provided to form LMMs. The random-effect term can be in several forms as given in Table 2 of Bates et al. (2015).

Value

fitLmm returns a LMM for each measured variable. Model information is returned as an object of class lmm2met, with the following components: completeMod = list of LMMs containing all given fixed and random effects

updateMod = list of LMMs containing fixed and random effects, if auto = TRUE or major is provided

testRes = list of outputs from chi-square tests

fittedDat = a dataframe of processed data

rawDat = a dataframe of original data

dataIndex = vector of indices of the first and last measured variable

Author(s)

Kwanjeera W [e-mail]

References

https://cran.r-project.org/web/packages/lme4/index.html

Bates, D., et al., Fitting Linear Mixed-Effects Models Using lme4. Journal of Statistical Software, 2015. 67(1): p. 1-48.

See Also

lmer, drop1, update, summary

Examples

##Ex 1. LMMs containing all given fixed and random effects
#fitMet = fitLmm(fix=c('Sex','Age','BMI','Stage','Location','Tissue'), 
#random='(1|Id)', data=adipose, start=10, end=14)
##Ex 2. LMMs containing only significant fixed effects and random effects
#fitMet = fitLmm(fix=c('Sex','Age','BMI','Stage','Location','Tissue'), 
#random='(1|Id)', data=adipose, start=10, end=14, auto=TRUE)
##Ex 3. LMMs containing only significant fixed effects, Tissue effect and random effects
#fitMet = fitLmm(fix=c('Sex','Age','BMI','Stage','Location','Tissue'), 
#random='(1|Id)', data=adipose, start=10, end=14, major='Tissue')
#structure(fitMet) #show lmm2met object
#summary(fitMet$updateMod[[1]]) #summarize outputs of fitting a LMM to variable X1

kwanjeeraw/lmm2met documentation built on May 14, 2019, 3:07 a.m.