Substantial differences in bias between single-digest and double-digest RAD-seq libraries: A case study.

Area of Science:

• Population genomics
• Molecular ecology
• Bioinformatics

Background:

• Restriction site-associated DNA sequencing (RAD-seq) is a powerful tool for population genomics.
• Single-digest RAD-seq (sdRAD-seq) and double-digest RAD-seq (ddRAD-seq) are common protocols with debated trade-offs.
• Direct empirical comparisons of sdRAD-seq and ddRAD-seq are lacking.

Purpose of the Study:

• To empirically compare single-digest RAD-seq (sdRAD-seq) and double-digest RAD-seq (ddRAD-seq) protocols.
• To identify sources of bias contributing to differences in allele frequency estimates between the two methods.
• To assess the implications of these differences for evolutionary studies.

Main Methods:

• Genotyping of 444 Gulf pipefish (Syngnathus scovelli) using both sdRAD-seq (n=60) and ddRAD-seq (n=384).
• Analysis of Illumina sequencing data, comparing results from pooled and separate analyses.
• In silico digestion of the Gulf pipefish genome to model potential sources of bias.

Main Results:

• Significant differences in coverage, heterozygosity, and allele frequencies were observed between sdRAD-seq and ddRAD-seq protocols.
• Polymorphic restriction sites, asymmetric sampling, and PCR duplicates were identified as key contributors to allele frequency discrepancies.
• Selection components analysis yielded different results depending on the protocol used.

Conclusions:

• sdRAD-seq and ddRAD-seq protocols can produce different allele frequency estimates.
• These differences are influenced by technical factors including restriction site variation and sampling strategy.
• Researchers must consider protocol choice and potential biases when comparing population genomic data across studies.