rpt.IBD: Replacing CPTs for selected mixture contributors with familial relationships

Description

Used after a call to DNAmixture with compile=FALSE,triangulate=FALSE, this function replaces the CPTs for the genotype allele count arrays for specified mixture contributors by those representing the specified relationship with each other and typed relatives

Usage

rpt.IBD(mixture, IBD="parent-child", typed.gts = NULL, inds = 1, 
jtyped = ncol(IBD$patt)/2 - length(typed.gts) + seq_along(typed.gts), 
jcontr = seq_along(inds), 
targets=attr(IBD,'targets'), contribs=NULL, quiet=FALSE, all.freq = NULL, compile = TRUE) 
rpt.typed.relatives(mixture, IBD="parent-child", typed.gts = NULL, inds = 1, 
jtyped = ncol(IBD$patt)/2 - length(typed.gts) + seq_along(typed.gts), 
jcontr = seq_along(inds), 
targets=attr(IBD,'targets'), contribs=NULL, quiet=FALSE, all.freq = NULL, compile = TRUE)
rpt.typed.child(mixture, aca, ind=1)
replace.Ui.tables(mixture, aca, ind=1)
rpt.typed.parents(mixture, Mgt, Fgt, ind=1, compile=TRUE)
rpt.typed.relative(mixture, Rgt, IBD=c(0.25,0.5,0.25), ind=1, compile=TRUE)

Value

Vector of marker-specific probabilities of the typed genotypes.

Arguments

mixture: DNAmixtures object created by previous call to DNAmixture with
triangulate=FALSE,compile=FALSE
IBD: relationships between the specified individuals, as multi-person condensed coefficients of identity, in one of several representation; see Details.
typed.gts: list of 0 or more genotypes of relatives; the components of this list must be named (with the id's of the relevant individuals) if targets and contribs are used to code the correspondences (see Details).
inds: vector of 1 or more integers: which `unknown' contributors are we modelling by amending their CPTs? The elements should be listed in the same order as the corresponding pairs of columns of the IBD patterns in IBD
jtyped: indices of pairs of columns of IBD$patt that correspond to the typed relatives (if any); default the last length(typed.gts) pairs of columns
jcontr: indices of pairs of columns of IBD$patt that correspond to the relevant mixture contributors; default the first length(inds) pairs of columns
targets: Character vector of the tags of the individuals referred to in IBD
contribs: Character vector of the tags of the individuals included in the mixture, in order
quiet: should calculated values of inds, jtyped and jcontr be reported?
all.freq: alternative allele frequency database(s), see Details.
compile: logical flag: should mixture object be compiled on exit?
ind: as inds, used above when only one allowed
aca, Mgt, Fgt, Rgt: individual genotypes, as allele count arrays

Author

Peter Green (P.J.Green@bristol.ac.uk)

Details

In using rpt.IBD or rpt.typed.relatives (which is identical), the correspondence between mixture contributors, specified relationships, and typed genotype profiles should be specified either (preferably) using targets, contribs and through the names of the components in typed.gts, or (to be deprecated) with inds, jcontr and jtyped: the two representations should not be mixed up. If either targets or contribs specified, the former representation is assumed.

Special cases are treated slightly more efficiently: rpt.typed.child: single contributor, single typed relative, parent or child; rpt.typed.parents: single contributor, both parents typed; rpt.typed.relative: single contributor, single typed relative.

Note that IBD$patt always has an even number of columns, two for each individual in the joint relationship specified; jtyped and jcontr are vectors of indices of these individuals, i.e. to pairs of adjacent columns of IBD$patt.

Multiple functions in this group can validly be called sequentially (with all but the last having compile=FALSE) providing they reference different sets of contributors among the targets, and that these sets are conditionally independent given the typed genotypes specified.

There are multiple valid representations for relationships in the argument IBD -- as an IBD pattern distribution, via a pedigree, or. in the case of just two individuals. via either a vector of 3 kappas or 9 Deltas (Jacquard's condensed coefficients of identity). For full details, see convertIBD. If IBD is missing, the default value represents parent-child.

In the interests of upward compatibility, in rpt.typed.child and replace.Ui.tables (which are identical), the argument Cgt can be given as either a genotype profile data frame, or an allele count array.

By default, the allele frequency database used for the founding genes is that used when the mixture object is created, in an earlier call to DNAmixture. A non-null value for the all.freq argument allows the user to specify alternative database (s) for the founding genes. If its value is an allele frequency database (in the format specified in formats) then that database is used for all founding genes; if the value of the argument is a list of such databases, then component k of the list is used for allele frequencies for the founding gene labelled k in the IBD argument. Note that this option allows modelling of mixtures where different contributors are drawn from different populations, whether or not there are relationships among individuals.

Examples

Run this code

data(test2data)
data(NGMDyes)

## Fit 2-person mixture - baseline model

mixD<-DNAmixture(list(epg),k=2,C=list(0.001),database=db)
pars<-mixpar(rho=list(2),eta=list(100),xi=list(0.1),phi=list(c(U1=0.7,U2=0.3)))
baseline<-logL(mixD)(pars)

## Fit 2-person mixture model in which contributor 1 is parent of a typed individual Cgt

mixD<-DNAmixture(list(epg),k=2,C=list(0.001),database=db,triangulate=FALSE,compile=FALSE)
rpt.IBD(mixD,,list(Cgt)) 
log10LR<-(logL(mixD)(pars)-baseline)/log(10)
cat('log10 LR',log10LR,'\n')

## Fit 2-person mixture model in which contributor 1 is father of a typed individual Cgt 
## with mother Mgt

mixD<-DNAmixture(list(epg),k=2,C=list(0.001),database=db,triangulate=FALSE,compile=FALSE)
rpt.IBD(mixD,,list(Mgt,Cgt)) 
log10LR<-(logL(mixD)(pars)-baseline)/log(10)
cat('log10 LR',log10LR,'\n')

## Fit 2-person mixture model in which contributors are two parents of a child with 
## genotype Cgt, and a parent of one of them has genotype Rgt. Note the encoding of allele 
## labels to reduce the complexity of the IBD pattern distribution IBD.

IBD<-list(patt=rbind(c(1,3,2,4,1,2,1,5),c(1,3,2,4,1,2,3,5)))

mixD<-DNAmixture(list(epg),k=2,C=list(0.001),database=db,triangulate=FALSE,compile=FALSE)
rpt.IBD(mixD,IBD,list(Cgt,Rgt),1:2) 
log10LR<-(logL(mixD)(pars)-baseline)/log(10)
cat('log10 LR',log10LR,'\n')

## the same, with individuals and relationships denoted by character tags

mixD<-DNAmixture(list(epg),k=2,C=list(0.001),database=db,triangulate=FALSE,compile=FALSE)
rpt.IBD(mixD,IBD,list(c=Cgt,gf=Rgt),targets=c('f','m','c','gf'),contribs=c('f','m')) 
log10LR<-(logL(mixD)(pars)-baseline)/log(10)
cat('log10 LR',log10LR,'\n')

Run the code above in your browser using DataLab