Learn R Programming
umx (version 1.9.1)

umxCF_SexLim: umxCF_SexLim

Description

Build a multivariate twin analysis with sex limitation based on a correlated factors model. This allows Quantitative & Qualitative Sex-Limitation. The correlation approach ensures that variable order does NOT affect ability of model to account for DZOS data. Restrictions: Assumes means and variances can be equated across birth order within zygosity groups

Usage

umxCF_SexLim(name = "ACE_sexlim", selDVs, mzmData, dzmData, mzfData, dzfData,
  dzoData, C_or_A = "A", suffix = NA)

Arguments

name

The name of the model (Default = "CF_sexlim")

selDVs

BASE NAMES of the variables in the analysis. You MUST provide suffixes.

mzmData

Dataframe containing the MZ male data

dzmData

Dataframe containing the DZ male data

mzfData

Dataframe containing the MZ female data

dzfData

Dataframe containing the DZ female data

dzoData

Dataframe containing the DZ opposite-sex data (be sure and get in right order)

C_or_A

Whether to model sex-limitation on C or on A. (Defaults to "A")

suffix

Suffix used for twin variable naming. Allows using just the base names in selVars

Value

- CF SexLim model

References

- Neale et al. (2006). Multivariate genetic analysis of sex-lim and GxE interaction. Twin Research & Human Genetics, 9, pp. 481--489.

See Also

Other Twin Modeling Functions: plot.MxModel, umxACESexLim, umxACE_cov_fixed, umxACEcov, umxACEv, umxACE, umxCP, umxGxE_window, umxGxE, umxIP, umxPlotACEcov, umxPlotCP, umxPlotGxE, umxPlotIP, umxReduceACE, umxSummaryACEcov, umxSummaryACEv, umxSummaryACE, umxSummaryCP, umxSummaryGxE, umxSummaryIP, umx_long2wide, umx_wide2long, umx, xmu_twin_check

Examples

Run this code
# NOT RUN {
# Load Libraries
require(umx)
# Create Functions to Assign Labels
laLower <- function(la,nVar) {
	paste(la,rev(nVar+1-sequence(1:nVar)),rep(1:nVar,nVar:1),sep="_")
}
laSdiag <- function(la,nVar) {
	paste(la,rev(nVar+1-sequence(1:(nVar-1))),rep(1:(nVar-1),(nVar-1):1),sep="_") 
}
laFull  <- function(la,nVar) {
	paste(la,1:nVar,rep(1:nVar,each=nVar),sep="_") 
}
laDiag  <- function(la,nVar) {
	paste(la,1:nVar,1:nVar,sep="_") 
}
laSymm  <- function(la,nVar) {
	paste(la,rev(nVar+1-sequence(1:nVar)),rep(1:nVar,nVar:1),sep="_") 
}
# =========================
# = Load and Process Data =
# =========================
data("us_skinfold_data")
# rescale vars
us_skinfold_data[,c('bic_T1', 'bic_T2')] <- us_skinfold_data[,c('bic_T1', 'bic_T2')]/3.4
us_skinfold_data[,c('tri_T1', 'tri_T2')] <- us_skinfold_data[,c('tri_T1', 'tri_T2')]/3
us_skinfold_data[,c('caf_T1', 'caf_T2')] <- us_skinfold_data[,c('caf_T1', 'caf_T2')]/3
us_skinfold_data[,c('ssc_T1', 'ssc_T2')] <- us_skinfold_data[,c('ssc_T1', 'ssc_T2')]/5
us_skinfold_data[,c('sil_T1', 'sil_T2')] <- us_skinfold_data[,c('sil_T1', 'sil_T2')]/5
# describe(us_skinfold_data, skew = FALSE)

# Select Variables for Analysis
varList = c('ssc','sil','caf','tri','bic')
selVars = umx_paste_names(varList, "_T", 1:2)
nVar = length(selVars)

# Data objects for Multiple Groups
mzmData = subset(us_skinfold_data, zyg == 1, selVars)
dzmData = subset(us_skinfold_data, zyg == 3, selVars)
mzfData = subset(us_skinfold_data, zyg == 2, selVars)
dzfData = subset(us_skinfold_data, zyg == 4, selVars)
dzoData = subset(us_skinfold_data, zyg == 5, selVars)

m1 = umxACESexLim(selDVs = varList, suffix = "_T",
       mzmData = mzmData, dzmData = dzmData, 
       mzfData = mzfData, dzfData = dzfData, 
       dzoData = dzoData)
m1 = mxRun(m1)
summary(m1)

# ===============================
# = 1 Nonscalar Sex Limitation  =
# ===============================
# Quantitative Sex Differences & Qualitative Sex Differences for A
# Male and female paths, plus male and female Ra, Rc and Re between variables
# Male-Female correlations in DZO group between 
# A factors Rao FREE, Rc constrained across male/female and opp-sex
# ===================================================
# = Test switching specific a from Males to females =
# ===================================================

m2 = umxSetParameters(m1, labels = "asm_.*", free = FALSE, values = 0, regex = TRUE)
m2 = umxSetParameters(m1, labels = "asf_.*", free = TRUE , values = 0, regex = TRUE)
m2 = mxRun(m2)
summary(m2)
mxCompare(m2, m1)
# ===============================
# = 2 Nonscalar Sex Limitation  =
# ===============================
# Quantitative Sex Differences & Qualitative Sex Differences for C
# Male and female paths, plus male and female Ra, Rc and Re between variables
# Male-Female correlations in DZO group between C 
# factors Rco FREE, Ra constrained across male/female and oppsex

# -------|---------|---------|---------|---------|---------|---------|---------|---------|-----|
# 3 Scalar Sex Limitation 
# Quantitative Sex Differences but NO Qualitative Sex Differences
# Male and female paths, but one set of Ra, Rc and Re between variables (same for male & female)
# ---------------------------------------------------------------------------------------------|

# =================================
# =================================
frODiag   <- c(rep(c(FALSE,rep(TRUE,nVar)),nVar-1),FALSE)
svODiag   <- c(rep(c(1,rep(.4,nVar)),nVar-1),1)
m3 = umxSetParameters(m2, labels = "asm_.*", free = FALSE, values = 0, regex = TRUE)
pathRam = mxMatrix(name="Ram", "Stand", nrow= nVar, free = TRUE, values = .4, 
		label = laSdiag("ra", nVar), lbound = -1, ubound = 1)
pathRaf = mxMatrix(name = "Raf", "Stand", nrow = nVar, free = TRUE, values = .4, 
		label=laSdiag("ra", nVar), lbound = -1, ubound = 1)
pathRcm = mxMatrix(name="Rcm", "Stand", nrow = nVar, free = TRUE, values = .4, 
		label=laSdiag("rc", nVar), lbound = -1, ubound = 1)
pathRcf = mxMatrix(name="Rcf", "Stand", nrow = nVar, free = TRUE, values = .4, 
		label=laSdiag("rc", nVar), lbound = -1, ubound = 1)
pathRem = mxMatrix(name="Rem", "Stand", nrow= nVar, free=TRUE, values = .4, 
		label = laSdiag("re", nVar), lbound = -1, ubound = 1)
pathRef = mxMatrix(name="Ref", "Stand", nrow= nVar, free=TRUE, values = .4, 
		label = laSdiag("re", nVar), lbound = -1, ubound = 1)
corRao  = mxMatrix(name="Rao", "Symm" , nrow= nVar, free = frODiag, values = svODiag,
         label = laSymm("ra", nVar), lbound = -1, ubound = 1)
corRco  = mxMatrix(name="Rco", "Symm" , nrow= nVar, free = frODiag, values = svODiag, 
         label = laSymm("rc", nVar), lbound = -1, ubound = 1)

# m3 <- makeModel("HetCfAce")
# m3 <- mxRun(m3)
# summary(m3)
# round(m3$VarsZm$result,4); round(m3$CorsZm$result,4)
# round(m3$VarsZf$result,4); round(m3$CorsZf$result,4)
# mxCompare(HetCfAceRgFit, m3)

# ===================
# = 4 Homogeneity 
# = NO Quantitative Sex Differences AND NO Qualitative Sex Differences
# = Same paths for males and females
# ===================

# =====================================
# = Equate [ace]m and [ace]f matrices =
# =====================================

pathAm = mxMatrix(name="am", "Diag", nrow = nVar, free = TRUE, values = .5, 
label = laDiag("a", nVar))
pathCm = mxMatrix(name="cm", "Diag", nrow = nVar, free = TRUE, values = .5, 
label = laDiag("c", nVar))
pathEm = mxMatrix(name="em", "Diag", nrow = nVar, free = TRUE, values = .5, 
label = laDiag("e", nVar))
pathAf = mxMatrix(name="af", "Diag", nrow = nVar, free = TRUE, values = .5, 
label = laDiag("a", nVar))
pathCf = mxMatrix(name="cf", "Diag", nrow = nVar, free = TRUE, values = .5, 
label = laDiag("c", nVar))
pathEf = mxMatrix(name="ef", "Diag", nrow = nVar, free = TRUE, values = .5, 
label = laDiag("e", nVar))

# m4 <- makeModel("HomCfAce")
# m4 <- mxRun(m4)
# summary(m4)
# round(m4$VarsZm$result,4); round(m4$CorsZm$result,4)
# round(m4$VarsZf$result,4); round(m4$CorsZf$result,4)
# mxCompare(m3, m4)

# ==============================================
# = Generate Output Table of all Nested Models =
# ==============================================

# mxCompare(HetCfAceRgFit, c(HetCfAceRcFit, m3, m4))

# rbind(
# 		mxCompare(HetCfAceRgFit, HetCfAceRcFit),
#  	mxCompare(HetCfAceRcFit, m3)[2,],
#  	mxCompare(m3, m4)[2,]
# )
# }

Run the code above in your browser using DataLab