README: Genomic analysis of population history for Hawaiian monk seals

https://doi.org/10.5061/dryad.djh9w0w72

This README file version was generated MAR 2024 by Samantha Hauser (samantha.hauser1@gmail.com).

General Information:

Published article in Endangered Species Research Journal: https://doi.org/10.3354/esr01308

Also note: The associated article is an expansion of Hawaiian monk seal research presented in a preliminary governmental report written by these authors, available at https://repository.library.noaa.gov/view/noaa/32349

File Information:

(a) HMS_BestRAD_SNPs.vcf:

• VCF formatted genomic data (final single nucleotide polymorphism (SNP) data) for the Hawaiian monk seal
• 7,507 SNP markers

• 169 individuals

  • 85 seals from the Northwestern Hawaiian Islands and 84 from the Main Hawaiian Islands
  • Across 14 islands (see HMS_BestRAD_PopMapInfo.csv)

• Excerpt from publication re: genomic data production:

  • "We extracted high-quality genomic DNA from tissue samples using the Qiagen DNeasy Blood and Tissue Kit. We followed the BestRAD library preparation, specifically the ‘New RAD protocol’, as per Ali et al. (2016) with no modifications and sequenced the resulting libraries on a portion of an Illumina Novaseq lane. Sequence reads (5,987,336,510 clean reads) were demultiplexed, filtered for quality, and trimmed to 140 bp using the process_radtags function with the best-rad flag in Stacks v2.0 (Rochette et al. 2019). The filtered reads were then aligned to the Hawaiian monk seal reference genome (Mohr et al. 2022) using the BWA-mem algorithm (Li & Durbin, 2009) and SNP genotypes were called using the reference-alignment pipeline in Stacks v2.0 (min_maf = 0.01, r = 0.50; Rochette et al. 2019). We identified individuals with a high proportion of missing data (50% genotyping rate) in each population in VCFTOOLS (Danecek et al. 2011), omitted them, and then re-ran only the populations function on the total dataset (min_maf = 0.01, r = 0.80) in STACKS per the “bad apples” protocol (Cerca et al. 2021). After filtering, 169 seals from 14 islands genotyped at 7,507 SNP loci remained. Exploratory analysis of filtering based on Hardy-Weinberg equilibrium and the recommendations outlined in Pearman et al. 2022 revealed that no filtering was more appropriate for our dataset than removing loci that exhibited departures in every population (“Out All”; unclear how to a priori define populations; performed similarly to “No Filter”) or in any population (“Out Any”; removed too many loci; suspected artificial population structure).  Not filtering based on Hardy-Weinberg equilibrium is often fitting in studies such as this one where assumptions are violated (i.e., large, closed population, equal reproductive variance, and random mating) and population stratification is unknown (Wittke-Thompson et al. 2005, Pearman et al. 2022)."

(b) HMS_BestRAD_PopMapInfo.csv:

• Associated island and region for each seal ID - matches with HMS_BestRAD_SNPs.vcf

• Variables:

  • Seal: unique seal ID

  • Island: primary island each seal was sampled/observed

    • Kure (KUR), Midway (MID), Pearl & Hermes (P&H), Lisianski (LIS), Laysan (LAY), French Frigate Shoals (FFS), Necker (NEC), Nihoa (NIH), Ni’ihau (NII), Kaua’i (KAU), O’ahu (OAH), Moloka’i (MOL), Kaho’olawe (KAH), and Hawai’i (HAW)

  • Region: region island can be classified within

    • NWHI: Northwestern Hawaiian Islands; MHI: Main Hawaiian Islands