FamEvent (version 3.2)

carrierprob: Compute mutation carrier probabilities for individuals with missing gentoypes

Description

Computes model- or data-based carrier probabilities for individuals with missing genotypes based on the observed mutation status of family members and the individual's phenotype.

Usage

carrierprob(condition = "geno", method = "data", fit = NULL, data, mode = "dominant", 
q = 0.02)

Value

Returns a data frame with a vector of carrier probabilities called carrp.geno when condition="geno" or carrp.pheno when condtion="geno+pheno" added after the last column of the family data.

Arguments

condition

Choice of conditional information to use for computing the carrier probability. Possible choices are "geno" for using observed genotypes and "geno+pheno" for using both observed genotype and phenotype information in the calculation of the carrier probability.

method

Choice of methods to calculate the carrier probability. Possible choices are "data" for empirical calculation of the carrier probabilities based on data, "mendelian" using Mendelian transmission probabilities based on observed carriers within families, or "model" using the parametric model fit; see details below. Default is "data". If method = "data", only data is required to be specified.

fit

An object of class penmodel, a fitted model by penmodelEM function for inferring missing mutation statuses in the family.

data

Family data that includes missing genotypes using the same data format generated by the function simfam.

mode

Choice of modes of inheritance when using method="model". Possible choices are "dominant" for dominant model or "recessive" for recessive model. Default is "dominant".

q

Frequency of the disease causing allele when using method="model". The value should be between 0 and 1. If NULL, the estimated allele frequency from data will be used. Default value is 0.02.

Author

Yun-Hee Choi

Details

When method="model" along with the choice of condition="geno+pheno", the carrier probability for individual \(i\) is calculated by conditioning on her/his observed phenotype and carrier statuses of family members P(X_i = 1 | Y_i , X^o ) = P(Y_i | X_i = 1) * P(X_i = 1 | X^o) / (P(Y_i | X_i = 1) * P(X_i = 1 | X^o) + P(Y_i | X_i = 0) * P(X_i = 0 | X^o)) where X_i indicates the unknown carrier status of individual \(i\) and X^o represents the observed carrier statuses in his or her family members; Y_i represents the observed phenotype t_i, δ_i of individual \(i\) in terms of age at onset t_i and disease status indicator δ_i with 1 used for affected individuals and 0 for unaffected individuals. When method="mendelian" along with the choice of condition="geno", the carrier probability is calculated based on Mendelian laws of genetic transmission with a fixed allele frequency.

See Also

simfam, penmodelEM, plot.simfam, summary.simfam

Examples

# Simulated family data with 30% of members missing their genetic information.

set.seed(4321)
fam <- simfam(N.fam = 100, design = "pop+", base.dist = "Weibull", mrate = 0.3,
       base.parms = c(0.01,3), vbeta = c(-1.13, 2.35), agemin = 20)
 
# EM algorithm for fitting family data with missing genotypes assuming a Weibull
# baseline hazard and dominant mode of Mendelian inheritance for a major gene.

fitEM <- penmodelEM(Surv(time, status) ~ gender + mgene, cluster = "famID", gvar = "mgene", 
        parms = c(0.01, 3, -1.13, 2.35), data = fam, design = "pop+", base.dist = "Weibull", 
        method = "mendelian", mode = "dominant")

# Carrier probability obtained by conditioning on the observed genotypes and phenotype, 
# assuming a dominant Mendelian mode of inheritance

fam.added <- carrierprob(condition = "geno+pheno", method = "model", fit = fitEM, 
             data = fam, mode = "dominant", q = 0.02)

# pedigree plot for family 1 displaying carrier probabilities

plot.simfam(fam.added, famid = 1)