Round Robin Allele Frequencies

This function utilizes rrmlg to calculate multilocus genotypes and then subsets each locus by the resulting MLGs to calculate the round-robin allele frequencies used for pgen and psex.

Usage

rraf(gid, pop = NULL, res = "list", by_pop = FALSE, correction = TRUE, ...)

Arguments

gid: a genind or genclone object
pop: either a formula to set the population factor from the strata slot or a vector specifying the population factor for each sample. Defaults to NULL.
res: either "list" (default), "vector", or "data.frame".
by_pop: When this is TRUE, the calculation will be done by population. Defaults to FALSE
correction: a logical indicating whether or not zero-valued allele frequencies should be corrected using the methods outlined in correcting rare alleles. (Default: TRUE)
...: options from correcting rare alleles. The default is to correct allele frequencies to 1/n

Value

a vector or list of allele frequencies

Details

Calculating allele frequencies for clonal populations is a difficult task. Frequencies calculated on non-clone-corrected data suffer from bias due to non-independent samples. On the other hand, frequencies calculated on clone-corrected data artificially increases the significance of rare alleles. The method of round-robin allele frequencies as presented in Parks and Werth (1993) provides a method of calculating allele frequencies in a way that minimizes both of these effects.

Rare Alleles

Allele frequencies at a given locus are calculated based on samples that are clone corrected without that locus. When this happens, rare alleles have a high likelihood of dropping out, giving them a frequency of "0". For some analyses, this is a perfectly fine outcome, but for analyses such as pgen and psex, this could result in undefined values. Setting correction = TRUE will allow you to control how these zero-valued allele frequencies are corrected. For details, please see the documentation on correcting rare alleles and examples.

Note

When by_pop = TRUE, the output will be a matrix of allele frequencies. Additionally, when the argument pop is not NULL, by_pop is automatically TRUE.

References

Arnaud-Haond, S., Duarte, C. M., Alberto, F., & Serrão, E. A. 2007. Standardizing methods to address clonality in population studies. Molecular Ecology, 16(24), 5115-5139.

Parks, J. C., & Werth, C. R. 1993. A study of spatial features of clones in a population of bracken fern, Pteridium aquilinum (Dennstaedtiaceae). American Journal of Botany, 537-544.

Author

Zhian N. Kamvar, Jonah Brooks, Stacy A. Krueger-Hadfield, Erik Sotka

Examples


data(Pram)

# Round robin allele frequencies, correcting zero-valued frequencies to 1/nInd(Pram)
rraf(Pram)
#> $PrMS6A1
#> PrMS6A1.165 PrMS6A1.168 
#>   0.5360825   0.4639175 
#> 
#> $Pr9C3A1
#> Pr9C3A1.216 Pr9C3A1.226 
#>   0.5357143   0.4642857 
#> 
#> $PrMS39A1
#> PrMS39A1.130 PrMS39A1.213 PrMS39A1.214 PrMS39A1.242 PrMS39A1.246 PrMS39A1.250 
#>  0.500000000  0.001371742  0.007575758  0.001371742  0.121212121  0.363636364 
#> PrMS39A1.254 
#>  0.007575758 
#> 
#> $PrMS45A1
#> PrMS45A1.154 PrMS45A1.166 PrMS45A1.186 
#>  0.001371742  0.505747126  0.494252874 
#> 
#> $PrMS43A1
#> PrMS43A1.213 PrMS43A1.281 PrMS43A1.285 PrMS43A1.345 PrMS43A1.349 PrMS43A1.357 
#>  0.001371742  0.001371742  0.001371742  0.001371742  0.001371742  0.001371742 
#> PrMS43A1.360 PrMS43A1.364 PrMS43A1.368 PrMS43A1.372 PrMS43A1.376 PrMS43A1.380 
#>  0.001371742  0.001371742  0.076923077  0.115384615  0.153846154  0.038461538 
#> PrMS43A1.384 PrMS43A1.388 PrMS43A1.465 PrMS43A1.469 PrMS43A1.473 PrMS43A1.477 
#>  0.115384615  0.001371742  0.001371742  0.001371742  0.001371742  0.001371742 
#> PrMS43A1.481 PrMS43A1.485 PrMS43A1.489 PrMS43A1.493 PrMS43A1.497 PrMS43A1.505 
#>  0.076923077  0.384615385  0.038461538  0.001371742  0.001371742  0.001371742 
#> 


if (FALSE) { # \dontrun{

## Round robin allele frequencies will be different than observed

# Compare to without round robin:
PrLoc <- seploc(Pram, res = "mat") # get locus by matrix
lapply(PrLoc, colMeans, na.rm = TRUE)

# Without round robin, clone corrected:
Pcc    <- clonecorrect(Pram, strata = NA) # indiscriminantly clone correct
PccLoc <- seploc(Pcc, res = "mat")
lapply(PccLoc, colMeans, na.rm = TRUE)

## Different methods of obtaining round robin allele frequencies

# Get vector output.
rraf(Pram, res = "vector")

# Getting the output as a data frame allows us to use ggplot2 to visualize
(Prdf <- rraf(Pram, res = "data.frame"))
library("ggplot2")
ggplot(Prdf, aes(y = allele, x = frequency)) +
  geom_point() +
  facet_grid(locus ~ ., scale = "free_y", space = "free")

## Round Robin allele frequencies by population (matrix only)

# By default, allele frequencies will be corrected by 1/n per population
(Prbp <- rraf(Pram, by_pop = TRUE))

# This might be problematic because populations like PistolRSF_OR has a
# population size of four.

# By using the 'e' argument to rare_allele_correction, this can be set to a
# more reasonable value.
(Prbp <- rraf(Pram, by_pop = TRUE, e = 1/nInd(Pram)))



} # }