penmodel: Estimate the penetrance model and penetrance curves

Description

Fits a penetrance model for family data based on a prospective likelihood with ascertainment correction and provides parameter estimates as well as the gender- and mutation-specific penetrance estimates.

Usage

penmodel(parms, vbeta, data, design="pop", base.dist="Weibull", robust=FALSE)

Arguments

parms

Vector of initial values for baseline parameters. parms=c(lambda, rho), where lambda and rho are the initial values for the scale and shape parameters, respectively. For the "lognormal" baseline distribution, rho > 0; for the other baselinse distributions, lambda > 0 and rho > 0.

vbeta

Vector of initial values for regression coefficients for gender and majorgene; vbeta=c(beta.s, beta.g).

data

Data frame generated from simfam or data frame containing specific variables: famID, indID, generation, gender, currentage, mgene, time, status and weight with attr(data,"agemin") specified.

design

Study design of the family data. Possible choices are: "pop", "pop+", "cli", "cli+" or "twostage", where "pop" is for the population-based design with affected probands whose mutation status can be either carrier or non-carrier, "pop+" is similar to "pop" but with mutation carrier probands, "cli" is for the clinic-based design that includes affected probands with at least one parent and one sib affected, "cli+" is similar to "cli" but with mutation carrier probands, and "twostage" is for the two-stage design with oversampling of high risks families. Default is "pop".

base.dist

Choice of baseline hazard distribution to fit. Possible choices are: "Weibull", "loglogistic", "Gompertz", "lognormal", or "gamma". Default is "Weibull".

robust

Logical; if TRUE, use robust `sandwich' standard errors and variance covariance matrix, otherwise use conventional standard errors and variance covariance matrix.

Value

Returns an object of class 'penmodel', including the following elements:

coefficients

Parameter estimates of transformed baseline parameters ($\lambda, \rho$) and regression coefficients for gender and mutation status ($\beta_s, \beta_g$).

varcov

Variance covariance matrix of parameter estimates. If robust=TRUE, robust `sandwich' variance covariance matrix is returned.

Standard errors of parameter estimates. If robust=TRUE, robust 'sandwich' standard errors are returned.

pen70.est

Penetrance estimates by age 70 specific to gender and mutation-status subgroups.

pen70.se

Standard errors of penetrance estimates by age 70 specific to gender and mutation-status subgroups.

pen70.ci

95% confidence interval for penetrance estimates by age 70 specific to gender and mutation-status subgroups.

ageonset

Vector of ages of onset ranging from agemin to 90 years.

pen.maleCarr

Vector of penetrance estimates for male carriers from agemin to 90 years.

pen.femaleCarr

Vector of penetrance estimates for female carriers from agemin to 90 years.

pen.maleNonCarr

Vector of penetrance estimates for male non-carriers from agemin to 90 years.

pen.femaleNonCarr

Vector of penetrance estimates for female non-carriers from agemin to 90 years.

logLik

Loglikelihood value for the fitted penetrance model.

Details

The penetrance model is fitted to family data with a specified baseline hazard distribution, $$ h(t|x_s, x_g) = h_0(t) \exp(\beta_s x_s+\beta_g x_g) $$ where $h_0(t)$ is the baseline hazards function specified by base.dist, which depends on the shape and scale parameters, $\lambda$ and $\rho$; $x_s$ indicates male (1) and female (0) and $x_g$ indicates carrier (1) or non-carrier (0) of a gene of interest (major gene).

For family data arising from population- or clinic-based study designs (design="pop", "pop+", "cli", or "cli+"), the parameters of the penetrance model are estimated from the ascertainment-corrected prospective likelihood approach (Choi, Kopciuk and Briollais, 2008).

For family data arising from a two-stage study design (design="twostage"), model parameters are estimated based on the composite likelihood approach (Choi and Briollais, 2011)

Transformed baseline parameters ($\lambda, \rho$) were used for estimation; log tranformation was applied to both scale and shape parameters for "Weibull", "loglogistic", "Gompertz" and "gamma" baseline distributions. For "lognormal" baseline distribution, the log transformation was applied only to shape parameter $\rho$, not to $\lambda$ which represents the location parameter in log-normal distribution.

Calculations of standard errors and 95% confidence intervals for penetrance estimates by age 70 were based on the penetrances obtained from 1000 Monte-Carlo simulations of the estimated penetrance model; for more details, see penci.

References

Choi, Y.-H., Kopciuk, K. and Briollais, L. (2008) Estimating Disease Risk Associated Mutated Genes in Family-Based Designs, Human Heredity 66, 238-251

Choi, Y.-H. and Briollais (2011) An EM Composite Likelihood Approach for Multistage Sampling of Family Data with Missing Genetic Covariates, Statistica Sinica 21, 231-253

Examples

Run this code

# NOT RUN {
# Family data simulated from population-based design using a Weibull baseline hazard 

fam <- simfam(N.fam=300, design="pop+", variation="none", base.dist="Weibull", 
       base.parms=c(0.01,3), vbeta=c(-1.13, 2.35), agemin=20, allelefreq=0.02)
 
# Penetrance model fit for simulated family data

fit <- penmodel(parms=c(0.01, 3), vbeta=c(-1.13, 2.35), data=fam, 
       design="pop+", base.dist="Weibull")

# Summary of the model parameter and penetrance estimates from model fit

summary(fit)

# Generate the lifetime penetrance curves from model fit for specific gender and 
# mutation status groups along with their non-parametric penetrance curves 
     
plot(fit)

# }

Run the code above in your browser using DataLab