KINSIMU (version 0.1.2-2)

LRgpgcam: LR in grandparent-child identification with reference

Description

LR when a child (C) is alleged to be a grand-child of another individual (GP), and an offspring(A) of the alleged grandparent participated, with or without the assistant of B's other parent (M). Hp is that, C is an offspring of A's full-sibling; and Hd that, C is unrelated to GP and A. Inbreeding is not considered.

Usage

LRgpgcam(A, C, GP, af, rare = NULL, allelename = FALSE, M = NULL)

Value

a data frame with one column and ss rows, containing log10 value of the CLR of each case

Arguments

A

Genotype data of the alleged uncle/aunt, should be data.frame with 2 columns and ss rows, where ss stand for sample size;

C

Genotype data of the child, with the same form with A

GP

Genotype data of the alleged grandparent of C, with the same form with A

af

name of allele frequency matrix, a data.frame of 1 column containing frequencies with allele names being row names

rare

frequency of rare allele on the locus

allelename

if TRUE, the input genotype data would be regarded as allelenames, otherwise, the position in the af matrix

M

Genotype data of the other parent of C, which can be NULL or as that of A

Details

There might be no allele sharing between GP and A, or between M and B, if so, the related part in the LR calculation would be treated as 0, which can be further optimized in future version.

Examples

# Construct pedi data.frames for two types of pedigrees
pedi1 <- data.frame(Person=c("GF","GM","F","A","M","C"),
Father=c("RI","RI","GF","GF","RI","F"),
Mother=c("RI","RI","GM","GM","RI","M"))
pedi2 <- data.frame(Person=c("GF","GM","F","A","M","C"),
Father=c("RI","RI","RI","GF","RI","F"),
Mother=c("RI","RI","RI","GM","RI","M"))
LR_1=LR_2=data.frame(Log10CLR=rep(0,10000))
for (i in 1:42) { 
# Simulate 10000 group of pedigrees where the Hp is true
  Genotype <- pedisimu(af = FortytwoSTR$afmatrix[[i]],ss = 10000,pedi = pedi1)
  LR_1 <-LR_1+ LRgpgcam(A=Genotype[,7:8],C=Genotype[,11:12],GP=Genotype[,1:2],M=Genotype[,9:10],
                        af=FortytwoSTR$afmatrix[[i]],rare=FortytwoSTR$rare[i][1,1])
#Simulate 10000 group of false pedigrees, i.e., P and C is unrelated to GP and A
Genotype <- pedisimu(af = FortytwoSTR$afmatrix[[i]],ss = 10000,pedi = pedi2)
  LR_2 <-LR_2+ LRgpgcam(A=Genotype[,7:8],C=Genotype[,11:12],GP=Genotype[,1:2],M=Genotype[,9:10],
                        af=FortytwoSTR$afmatrix[[i]],rare=FortytwoSTR$rare[i][1,1])
}
xmin<-floor(min(min(LR_1$Log10CLR),min(LR_2$Log10CLR)))
xmax<-ceiling(max(max(LR_1$Log10CLR),max(LR_2$Log10CLR)))
par(mfrow = c(1, 2))
hist(LR_2$Log10CLR,xlab = expression(log[10]~CLR),main = "False pedigree",
     xlim = c(xmin,xmax), col = "red")
hist(LR_1$Log10CLR,xlab = expression(log[10]~CLR),main = "True cases",
     xlim = c(xmin,xmax), col = "blue")