A function that takes PMR observations, and (given a prior distribution for degrees of relatedness) returns the posterior probabilities of all pairs of individuals being (a) the same individual/twins, (b) first-degree related, (c) second-degree related or (d) "unrelated" (third-degree or higher). The highest posterior probability degree of relatedness is also returned as a hard classification. Options include setting the background relatedness (or using the sample median), a minimum number of overlapping SNPs if one uses the sample median for background relatedness, and a minimum number of overlapping SNPs for including pairs in the analysis.
callRelatedness(
pmr_tibble,
class_prior = rep(0.25, 4),
average_relatedness = NULL,
median_co = 500,
filter_n = 1
)results_tibble: A tibble containing 13 columns:
row: The row number
pair: the pair of individuals that are compared.
relationship: the highest posterior probability estimate of the degree of relatedness.
pmr: the pairwise mismatch rate (mismatch/nsnps).
sd: the estimated standard deviation of the pmr.
mismatch: the number of sites which did not match for each pair.
nsnps: the number of overlapping snps that were compared for each pair.
ave_re;: the value for the background relatedness used for normalisation.
Same_Twins: the posterior probability associated with a same individual/twins classification.
First_Degree: the posterior probability associated with a first-degree classification.
Second_Degree: the posterior probability associated with a second-degree classification.
Unrelated: the posterior probability associated with an unrelated classification.
BF: A strength of confidence in the Bayes Factor associated with the highest posterior probability classification compared to the 2nd highest. (No longer included)
a tibble that is the output of the processEigenstrat function.
the prior probabilities for same/twin, 1st-degree, 2nd-degree, unrelated, respectively.
a single numeric value, or a vector of numeric values, to use as the average background relatedness. If NULL, the sample median is used.
if average_relatedness is left NULL, then the minimum cutoff for the number of overlapping snps to be included in the median calculation is 500.
the minimum number of overlapping SNPs for which pairs are removed from the entire analysis. If NULL, default is 1.
callRelatedness(counts_example,
class_prior=rep(0.25,4),
average_relatedness=NULL,
median_co=5e2,filter_n=1
)
Run the code above in your browser using DataLab