LLCT: Longitudinal Linear Combination Test for Gene Set Analysis
In its-likeli-jeff/LLCT: Longitudinal Linear Combination Test for Gene set Analysis

Description Usage Arguments Value Examples

View source: R/LLCT.R

LLCT is a two-step self-contained gene-set analysis method which is developed to handle multiple longitudinal outcomes. Analysis of within-subject variation in the first step is followed by examining the between-subject variation utilizing Linear Combination Test (LCT) in the second step. This method is also applicable in analysis of time-course microarray data.

LLCT(EXPR, GS, LongData, ID = "ID", time = "time", covariate = NULL, phenotype = "phenotype", familybased = FALSE, pedigree = NULL, FIX.formula = "~time+covariate", RANDOM.formula = "~1|ID", nbPermutations = 1000, family = "gaussian(link=identity)")

`EXPR`	Gene Expressions — Row names:Subjects (single row per subject); Col names: Genes names. Note: The rownames of GS must be matched with colnames of EXPR and the unique values of the rownames of LongData must be matched with the rownames of EXPR.
`GS`	Gene Set Matrix — Row names: list of genes; Col names: list of selected gene sets; Cells: 1 (if gene of the row belongs to the gene set of the column) 0 (otherwise). Note: The rownames of GS must be matched with colnames of EXPR
`LongData`	Phenotypes and Covariates Data in Longitudinal format
`ID`	Name of ID variable in LongData
`time`	Vector of the names of time variables in LongData
`covariate`	name of covariate(s) in LongData
`phenotype`	name of phenotype in LongData
`familybased`	TRUE/FALSE indicating if the subjects are sharing families/pedigress Defaults to FALSE
`pedigree`	name of pedigree variable in LongData
`FIX.formula`	Formula of Phenotype~time model Defaults to "phenotype~time+covariate"
`RANDOM.formula`	Formula of random effects as lme function requires - required in case of related subjects Defaults to NULL
`nbPermutations`	Number of Permutations Defaults to 1000
`family`	family of link function appropriate for the type of phenotype

`LLCT_Results`	geneset-specific p-values and q-values
`Step1_Coefs`	coefficients calculated in the first step using FIX_formula

data(data_for_LLCT)

#LLCT for unrelated subjects with two more complex time trend
LLCT(EXPR=EXPR, GS=GS, LongData=LongData, ID="ID",time=c("time","time2"), covariate=NULL,phenotype="phenotype",familybased=FALSE,pedigree=NULL,FIX.formula="~time+time2", RANDOM.formula=NULL,nbPermutations=1000)

#LLCT for unrelated subjects with a linear time trend
#LLCT(EXPR=EXPR, GS=GS, LongData=LongData, ID="ID",time=c("time"), covariate=NULL,phenotype="phenotype",familybased=FALSE,pedigree=NULL,FIX.formula="~time", RANDOM.formula=NULL,nbPermutations=1000)

#LLCT for unrelated subjects with adjustment for time dependent covariate
#LLCT(EXPR=EXPR, GS=GS, LongData=LongData, ID="ID",time=c("time"), covariate="covariate",phenotype="phenotype",familybased=FALSE,pedigree=NULL,FIX.formula="~time+covariate", RANDOM.formula=NULL,nbPermutations=1000)

#LLCT for unrelated subjects with binary phenotype
#LLCT(EXPR=EXPR, GS=GS, LongData=LongData, ID="ID",time=c("time"), covariate=NULL,phenotype="binary.phenotype",familybased=FALSE,pedigree=NULL,FIX.formula="~time", RANDOM.formula=NULL,nbPermutations=1000,family="binomial(link=logit)")

#LLCT for related subjects with adjustment for time dependent covariate
#LLCT(EXPR=EXPR, GS=GS, LongData=LongData, ID="ID",time=c("time"), covariate="covariate",phenotype="phenotype",familybased=FALSE,pedigree="pedigree",FIX.formula="~time+covariate", RANDOM.formula="~1|ID",nbPermutations=1000)

#LLCT for related subjects with adjustment for more complex time trend
#LLCT(EXPR=EXPR, GS=GS, LongData=LongData, ID="ID",time=c("time","time2"), covariate="covariate",phenotype="phenotype",familybased=FALSE,pedigree="pedigree",FIX.formula="~time+time2", RANDOM.formula="~1|ID",nbPermutations=1000)

#LLCT for unrelated subjects with two phenotypes
#LLCT(EXPR=EXPR, GS=GS, LongData=LongData, ID="ID",time=c("time"), covariate=NULL,phenotype=c("phenotype","phenotype2"),familybased=FALSE,pedigree=NULL,FIX.formula="~time", RANDOM.formula=NULL,nbPermutations=1000)