In thierrygosselin/assigner: Assignment Analysis with RADseq Data

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Prepare working environment

rm(list = ls())
if (!require("pak")) install.packages("pak")
pak::pkg_install("thierrygosselin/assigner")
library(assigner)
assigner::install_gsi_sim(fromSource = TRUE)

Prepare whitelists

• Gather your whitelist in a single folder
• Get the list of these file in a string with the command below:

all.whitelists <- list.files(path = "~/Desktop/whitelists_salmon", pattern = "whitelist")
all.whitelists # to see if the whitelists are all there

Create a function to run your assignment

• Here the goal is to have a the argument you need selected
• The only argument that will change will be the whitelist
• you could have multiple arguments changing at the same time, but for simplicity, let's start with just one

whitelists_assigner <- function(x) {
  res <- assigner::assignment_ngs(
    data = "subset_melville_salmon.tsv", #change for what you want
    whitelist.markers = x, # don't change this one,
    assignment.analysis = "adegenet", #change for what you want
    common.markers = TRUE,#change for what you want
    marker.number = "all",# I think you get the idea...
    pop.levels = c(1,2,3,4,5,6,7,8,9,10,11),
    sampling.method = "random",
    iteration.method = 1,
    keep.gsi.files = FALSE
  )
}

Run the assignment

• The trick here is to map these whitelists to your new function whitelists_assigner
• The results are in a list easily managed by R
• The purrr package is awesome for this, it's very similar to the lapply function, only better.
• If you have access to a MAC or a LINUX cluster, you could just replace the purrr::map function to parallel::mclapply

salmon.assignment.res <- purrr::map(
.x = all.whitelists, # your whitelist created above
.f = whitelists_assigner # your new function
)

thierrygosselin/assigner documentation built on Nov. 9, 2024, 3:38 a.m.