Population genomics analysis with RAD, reprised: Stacks 2

Abstract

Restriction enzymes have been one of the primary tools in the population genetics toolkit for fifty years, being coupled with each new generation of technology to provide a more detailed view into the genetics of natural populations. Restriction site-Associated DNA protocols, which joined enzymes with short-read sequencing technology have democratized the field of population genomics, providing a means to assay the underlying alleles in scores of populations. More than ten years on, the technique has been widely applied across the tree of life and served as the basis for many different analysis techniques. Here, we provide a detailed protocol to conduct a RAD analysis from experimental design to de novo analysis – including parameter optimization – as well as reference-based analysis, all in Stacks version 2, which is designed to work with paired-end reads to assemble RAD loci up to 1000 nucleotides in length. The protocol focuses on major points of friction in the molecular approaches and downstream analysis, with special attention given to validating experimental analyses. Finally, the protocol provides several points of departure for further analysis.

Reference genome choice and filtering thresholds jointly influence phylogenomic analyses

“…We further demonstrate that minor allele filters greater than ∼4% reduce topological accuracy in our 51-taxon phylogenies and increase both imbalance and γ statistics, suggesting that they are removing important signal above this threshold rather than only removing sequencing or alignment errors. Thus, our analyses support other recommendations suggesting that filtering for a minor allele count of 3 allows for removing true sequencing errors without removing true signal in the data in most situations (e.g., Rochette et al, 2019;Rivera-Colón and Catchen, 2022;O'Leary et al, 2018). The use of minor allele filters of 5% or higher is common in many empirical studies; these high thresholds may be inadvertently biasing downstream results in these studies.…”

Chromosome‐scale genome assembly of Poa trivialis and population genomics reveal widespread gene flow in a cool‐season grass seed production system

“…The level of expected heterozygosity ( H E ), nucleotide diversity (π), and inbreeding coefficient ( F IS ) was calculated for each populations using 73,506 single nucleotide polymorphisms (SNPs) with the Stacks pipeline (Rivera‐Colón & Catchen, 2021 ) and Tajima's D with vcftools (Danacek et al, 2011 ). All P. trivialis populations exhibited similar H E , that ranged from .030 to .035, while the outgroup ( P. annua ) had a greater H E .…”

Chromosome-scale genome assembly ofPoa trivialisand population genomics reveals widespread gene flow in a cool-season grass seed production system

“…The level of expected heterozygosity (H E ), nucleotide diversity (π), and inbreeding coefficient (F IS ) was calculated for each populations using 73,506 SNP’s with the Stacks pipeline (Rivera-Colón & Catchen, 2021), and Tajima’s D with vcftools (Danacek et al, 2011). All P. trivialis populations exhibited similar H E , that ranged from 0.030 to 0.035, while the outgroup ( P. annua ) had a greater H E .…”