rel_lynch: This estimates the Lynch & Ritland (1999)

In dyerlab/gstudio: Tools Related to the Spatial Analysis of Genetic Marker Data

This estimates the Lynch & Ritland (1999)

Description

This function estimates the allele-wise relatedness statistic from Lynch & Ritland (1999). It is very sensitive to the presence of rare alleles.

This is an symmetric estimator using the formula:

Usage

rel_lynch(x)

Arguments

x	A data.frame or vector that has locus objects in it. If you pass it a data.frame it will return the multilocus relatedness.

Details

\hat{r}_{XY} = \frac{r_{XY} + r_{YX}}{2}

where

r_{XY} = \frac{1}{\sum_{i=1}^L w_i} \sum_{i=1}^L \frac{p_a(\delta_{bc}+\delta_{bd}) + p_b(\delta_{ac} + \delta_{ad}) - 4p_a p_b}{2p_a p_b}

and

w_i = \frac{((1+\delta_{ab})(p_a+p_b) - 4*p_a p_b)}{2p_a p_b}

For missing genotypes, relatedness is not estimated for any pair-wise comparison. This does not constitute an error, specifically, but you should be careful of noting which comparisons are made from no or at least fewer than the full set of loci.

Value

A matrix of pairwise values

Note

For missing data and for data comparing heterozygotes at a 2-allele locus this estimator is undefined.

Author(s)

Rodney J. dyer <[e-mail]>

dyerlab/gstudio documentation built on Feb. 2, 2024, 8:24 p.m.