/RAD-error-rates

R scripts for estimating locus, allele and SNP error rates of RAD data processed with Stacks

Primary LanguageRGNU General Public License v3.0GPL-3.0

README

This repo contains R functions to estimate error rates; to generate the a list of white loci and population maps to run the populations program of Stacks; to perform a PCoA; and, to estimate the distance between individuals of the same population. The scripts assume than several datasets will be compared (e.g. the several outputs of assembling the data with different parameters).

Error rates

The R functions to estimate the loci, allele and SNP error rates are:

  • LociAllele_error.R, which uses a tsv matrix of alleles and loci (formated as the output of export_sql from Stacks but keeping only the columns CatalogID Consensus SNPs Sample1 Sample1_r Sample2 ...) to convert it to genlight object to estimate loci and allele error rates between replicate pairs. The function produces a dataframe with the following columns for each replicate pair:

  • "nloci" = total number of loci in the dataset

  • "nMissLoc" = number of missing loci for both samples of a replicate pair

  • "MissTotProp", = nMissLoc/nloci

  • "shareMissLoc" = number of loci that are lost in both samples of replicate pair

  • "loci.mismatches" = number of loci that failed in the sample or replicate, but not in both

  • "unshareMissLoc" = loci.mismatches/nMissLoc

  • "loci.error.rate" = loci.mismatches/nloci

  • "n.loci.woNA" = number of loci that were NOT missing in both samples or a replicate pair

  • "allele.mismatches" = number of alleles that mismatch between samples of a replicate pair

  • "allele.error.rate" allele.mismatches/n.loci.woNA

and

  • SNPs_error.R that uses a genlight object of SNPs, (obtained by e.g. transforming a plink.raw to genlight object) to estimate SNP error rates.

The function PairsDiff.R wraps up LociAllele_error. and SNPs_error.R to estimate loci, allele and SNP error rates and visualize the Matrix of SNPs as a plot.

White lists and Population Map

  • whitePopMap.R creates a list of the samples present in a desired df (e.g. a subsection or of the tsv output of explort_sql) and writes a file that can be read by Stacks as a Population Map of desired samples to analyse with the populations program

  • whiteRADlist.R creates a list of the CatalogID loci present in a desired matrix and write a file that can be read by Stacks as whitelist of desired loci to perform analyses with the populations program.

Compare distance between populations and PCoA

  • dist.pop.R estimate distance between individuals of the same population
  • PCoA_pop.r performs a PCoA.