Provides estimates of admixture proportions in offspring and ungenotyped parents, using phased data
Usage
LEAPFrOG_EM(data,p,chr,alpha=1e-6)
Arguments
data
2 column matrix of allele counts, with each row as a SNP. Columns 1 and 2 refer to the 2 haplotypes. Each entry is either 1, 0 or NA.
p
Matrix of allele frequencies. Each row corresponds with a SNP. Number of rows must equal length of data. Each column is a population
chr
Vector of chromosome identifiers, one for each SNP. Each entry is an integer, 1-22 for the autosomes. If two X chromosomes for a female are being analysed, it should be identified by the number 23.
alpha
Convergence tolerance for the EM algorithm. The optimisation will stop when an interation fails to change the parental admixture proportions (total change across all parameters) by this amount
Details
LEAPFrOG_EM requires python to be installed. Only the parental admixture proportions are estimated directly (all except the last population), and therefore standard errors are only reported for these only.
Value
A list including elements
m
A vector: Admixture proportions (one per population) for the genotyped offspring.
P1
A vector: Admixture proportions (one per population) for parent 'A'.
P2
A vector: Admixture proportions (one per population) for parent 'B'.
P1se
A vector: Standard errors (All populations except the last) foradmixture proportions in parent 'A'.
P2se
A vector: Standard errors (All populations except the last) foradmixture proportions in parent 'A'.
iterations
Number of expectation steps performed.
value
Value of the likelihood function for the final maximisation step.
Author(s)
Daniel Crouch & Michael Weale, Department of Medical and Molecular Genetics, King's College London