The Latent Class Twin Method

By Stuart G. Baker, 2015

Introduction

The latent class twin method assumes two types of latent classes: an environmental susceptibility (ES) latent class and a genetic susceptibility (GS) latent class. The GS latent class represents a genetic contribution (arising from an additive, autosomal dominant, or autosomal recessive model) associated with high probabilities of independent outcomes. The ES latent class involves background probabilities of possibly dependent outcomes. The latent class twin method estimates the genetic prevalence, which is the fraction of persons in the genetic susceptibility latent class, and the heritability fraction, which is the fraction of persons in the genetic susceptibility latent class with the trait or outcome.

Reference

Baker, SG. The latent class twin method.

Requirement

Mathematica Version 10 or later.

Set-Up

copy all files into some folder called "FOLDER" a new Mathematica session SetDirectory["FOLDER"] << twinfit.m

To run computations in paper for breast cancer twin survival data

type TwinFit[dataBem, parsym, MaxBoot->2000]

To run computations in paper for breast cancer twin data with sum of counts

type TwinFit[dataBobs, parsym]

To try on your own data,

type TwinFit[ dataset, parsym,options]

Options

Option Default Explanation
NewFitQ TRUE New fitting or use stored result of previous fit
CensoringStart "during" Censoring in relation to interval; other choice is "after" interval
NumCatFinal 1 Number of intervals in final category
AMin 0.7 Minimum outcome probability for GS latent class
ShowPlot FALSE Profile plot of deviance versus s for different values of a
ShowProgress FALSE Show progress in terms of profile of likelihood
MaxBoot 0 Number of bootstrap iterations

dataset={{m2,m1,m0,d2,d1,d0}},datatype,dataname,datafilename}

m2 L x L matrix of incidence, incidence for MZ twins L x L matrix of incidence, censoring for MZ twins L x L matrix of censoring, censoring for MZ twins L x L matrix of incidence, incidence for DZ twins L x L matrix of incidence, censoring for DZ twins L x L matrix of censoring, censoring for DZ twins "em" or "obs" . If datatype ="obs", m2,m1,m0, d2, d1,d0 are scalar name of dataset name used for files and plots

parsym={a,b,c,s}

File Contents

File name Description
twinfit.m Calls other packages
twinfitcore.m Core computations
twinfitprofille.m Fits profile likelihood as a function of allele frequency s
twinfitprofileplot.m Plot of profile likelihood
twinfitem.m Fits survival data with EM algorithm
twinfitprob.m Probability of latent class and outcome given latent class
twinfitfreq.m Probability of genotype
twinfitvar.m Variance components method (for comparison)
matrixlx.m Matrix functions
twinfitsummary.m Summary report with bootstap
twinfitcheck.m TwinFitCheck[dataset,parysm] checks convergence (extra)
twinfitsim.m TwinFitSim[parsym] compare estimates under different amin (extra)
twinfitheritability.m TwinFitPlotH[] plots heritability versus heritability fraction (extra)

Disclaimer

This code is provided "as is", without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and non-infringement. In no event shall the NCI or the individual developers be liable for any claim, damages or other liability of any kind. Use of this code by recipient is at recipient's own risk. NCI makes no representations that the use of the code will not infringe any patent or proprietary rights of third parties.

Last updated: November 16, 2015