Leveraging genomic innovations for effective coral restoration in Kenya

Methods

Summary

We collected coral tissue fragments (n=63) from two branching Acropora species — Acropora intermedia (previously referred to as A. formosa, prior to a 2024 taxonomic revision of the species complex) and Acropora tenuis — across three restoration contexts in the Shimoni-Vanga seascape, Kwale County, Kenya: natural reef colonies, nursery-reared fragments, and artificial-reef outplants. Sampling took place in January 2024 at the Mkwiro Community Managed Area and House Reef Firefly restoration sites, in partnership with Mkwiro Beach Management Unit and REEFolution Trust.

Genomic DNA was extracted from each sample using the Qiagen DNeasy Blood & Tissue kit, quantified and quality-checked by NanoDrop spectrophotometry and gel electrophoresis, then built into 3RAD sequencing libraries following the Bayona-Vásquez et al. (2019) protocol (BamHI-HF, ClaI and MspI digestion, iTru-indexed adapters, size-selected to ~300–450 bp). Libraries were sequenced as 150 bp paired-end reads on an Illumina NovaSeq X, producing over one billion paired-end reads in total.

Reads were quality-filtered and demultiplexed in Stacks v2.65. A. intermedia reads were aligned to a conspecific reference genome (BWA-MEM) and genotyped with the Stacks ref_map pipeline; A. tenuis, for which cross-species reference alignment performed poorly, was genotyped through de novo locus assembly. After filtering, this yielded 15,357 SNPs across 44 A. intermedia individuals and 1,422 SNPs across 19 A. tenuis individuals, which were used to assess genetic diversity, clonal structure, relatedness/inbreeding, and population structure across the three restoration contexts.

Challenges

The main challenges we anticipated and encountered: (1) Acropora tenuis had no suitable conspecific reference genome, and cross-species alignment to related genomes produced poor mapping quality — we resolved this by genotyping A. tenuis through de novo locus assembly instead of reference alignment, which yielded a smaller but still usable SNP panel (1,422 SNPs vs. 15,357 for A. intermedia). (2) Because individual donor colonies were not tagged when the nurseries were originally established (fragments were collected as "corals of opportunity" and propagated without a tracking system), we could not definitively separate genuine nursery-to-outplant genetic change from incomplete donor tracking — we addressed this by explicitly testing and reporting clonal structure and cross-context genotype sharing rather than assuming independence between contexts. (3) Sample sizes per restoration context were necessarily modest (as few as 6–7 individuals in some A. tenuis groups) given field and budget constraints, which limits the statistical power to detect subtle differences — we used non-parametric tests with false-discovery-rate correction throughout to guard against over-interpreting small-sample noise, and we flag this limitation directly in our reporting rather than over-claiming precision.

Pre Analysis Plan

For each species separately, we planned to: (1) estimate genetic diversity — observed and expected heterozygosity, nucleotide diversity, rarefied allelic richness, and individual inbreeding coefficients (FIS) — and compare these across natural reef, nursery, and artificial-reef contexts using Kruskal–Wallis tests, followed by pairwise Mann–Whitney tests with Benjamini–Hochberg correction for multiple comparisons where the global test was significant; (2) identify clonal replicates using pairwise identity-by-state similarity (IBS ≥ 0.95 threshold) and summarize genet richness as the ratio of unique genets to total sampled fragments per context; (3) estimate genome-wide relatedness (KING kinship) and runs of homozygosity (PLINK) to test for elevated inbreeding or restricted founder representation in restored contexts relative to the natural reef; and (4) evaluate population structure and differentiation using PCA, ADMIXTURE ancestry estimation, pairwise FST, and PERMANOVA on genetic distance matrices, to test whether nursery and outplant populations diverge from their natural-reef source. This plan was followed as specified; full results are reported under Lab Notes and in the project's published Results.