Scientists Decode the Full Genome of the Ocean-Wandering Trindade Petrel

Researchers have assembled a high-quality, long-read reference genome for the Trindade petrel (Pterodroma arminjoniana), a wide-ranging seabird that breeds on remote islands in the Atlantic and Indian Oceans and regularly hybridises with closely related petrel species. The new assembly, built from Oxford Nanopore long-read sequencing of a single male bird captured on Trindade Island off the Brazilian coast, spans 1.25 gigabases and recovers 97.6 percent of expected avian genes, providing one of the most complete genomic resources yet available for the gadfly petrels, a genus of highly threatened oceanic seabirds.

The Trindade petrel is an unusual candidate for such a detailed genomic portrait. The species originates from Trindade Island, a volcanic outpost roughly 1200 kilometres from the Brazilian coast, but genetic and tracking data suggest it ranges widely across the western Indian and southern Pacific Oceans. Over the past century it has also colonised Round Island off Mauritius, where it coexists and interbreeds with several related Pacific Pterodroma species. On its home island, the species shows two temporally distinct breeding seasons, with different sub-populations nesting in the same rocky cavities around the extinct volcano at different times of year, a phenomenon that can, in other seabirds, drive the early stages of speciation.

These biological quirks are precisely why the genome matters. Hybridisation between closely related Pterodroma species is recorded across the petrel’s range, particularly with Pterodroma neglecta and Pterodroma heraldica, and the fitness consequences of that mixing remain unresolved. Temporal segregation of breeding within a single population raises a second set of questions: whether birds breeding at different times are accumulating genetic or transcriptomic differences, and whether such allochrony might ultimately split the species without any physical barrier. A contiguous reference genome gives researchers the scaffold they need to test these ideas with genome-wide data rather than the small sets of genetic markers used previously.

The sequencing effort itself reflects the logistical realities of working on a remote tropical island. A single adult breeding bird was captured with a dip net at Praia das Tartarugas on Trindade Island in April 2022, banded, and then recaptured in May 2024, when roughly 500 microlitres of blood was drawn from the brachial vein. The sample was stored in ethanol and kept frozen on the island before reaching an ultralow-temperature freezer a week later. Genomic DNA was extracted with a high-molecular-weight kit recommended by Oxford Nanopore Technologies, and three types of sequencing libraries were built from the same individual to maximise coverage from the limited material.

Sequencing was performed on a PromethION 2 Solo device at the Subtropical Biodiversity Center of the Universidade Federal do Rio Grande in Brazil, using six new and three reused flow cells. The three libraries collectively produced about 57 gigabases of sequence before filtering; after adapter trimming and read filtering, the assembly dataset comprised 55.27 gigabases, equivalent to roughly 44-fold coverage of the final genome. The ultra-long library contributed the bulk of the data, delivering reads with a mean length above 10,000 bases and an N50 of more than 22,000 bases, the long continuous molecules that make a high-quality assembly possible.

The team assembled the reads de novo with the Flye assembler, polished the result with Medaka, screened for adapter and contaminant sequence, removed a single mitochondrial contig, and purged haplotypic duplications before scaffolding the assembly against a chromosome-level reference genome of Cory’s shearwater. The final scaffolded assembly contains 1103 scaffolds with a scaffold N50 of 85.5 megabases, meaning half the genome sits in just five scaffolds, and the longest scaffold reaches 220 megabases. Gaps account for only 0.069 percent of the assembly. As a benchmark, the researchers also assembled the same reads independently with Hifiasm, which produced slightly greater contiguity but more duplicated genes; the Flye assembly was retained as the primary reference because its lower duplication provides a more conservative basis for downstream analyses.

Read-depth analysis confirmed the sampled bird was male, with similar coverage of the Z chromosome and the autosomes, as expected for a bird carrying two Z chromosomes. Repeat analysis showed that 11.55 percent of the genome consists of repetitive elements, dominated by LINE transposons, broadly in line with other bird genomes. Structural annotation with the deep-learning gene predictor Helixer produced 18,099 protein-coding gene models, 97.3 percent of them multi-exonic, and functional annotation assigned putative functions to 94.9 percent of the predicted proteins, mostly through similarity to manually curated Swiss-Prot entries. The annotation also includes 333 predicted transfer RNA genes and 66 ribosomal RNA genes. Completeness checks on the protein set recovered 93.6 percent of avian orthologues, and a separate quality tool found no contaminant sequences.

One unexpected finding concerns the mitochondrion. The mitochondrial contig identified in the assembly measures about 20,300 bases, unusually long for a bird mitochondrial genome, and an independent Hifiasm assembly recovered a highly similar contig about 1100 bases longer. The two sequences align over nearly their entire length with only 0.24 percent divergence, showing the expanded region was recovered consistently by both assemblers, although its precise architecture is not fully resolved. The authors suggest the extra length may reflect duplicated or repetitive mitochondrial sequence, a pattern reported in other procellariiform seabirds, but they caution that resolving its structure will require a dedicated mitogenomic analysis.

To place the new genome in context, the team compared it with eight genomes from the Birds 10 K consortium and the recently published Galápagos petrel genome. The new assembly is far more contiguous than the two existing short-read Trindade petrel assemblies, with a scaffold N50 of 85.5 megabases against 18 and 17 megabases, while all three show similarly high gene completeness. Phylogenomic analysis of more than 6000 shared single-copy loci produced a striking result: the two existing B10K genomes labelled as Trindade petrel formed a strongly supported pair in every analysis, but the new reference genome consistently fell outside that pair, so the three genomes named P. arminjoniana did not form an exclusive group. The researchers stress this pattern is hypothesis-generating rather than proof of species-level paraphyly; disentangling incomplete lineage sorting, introgression, deep population structure and taxonomic uncertainty will require denser population-level sampling and genome-wide tests of admixture.

The genome arrives at a critical moment for the species. The Trindade petrel is classified as Vulnerable globally but Critically Endangered in Brazil, and introduced house mice remain an ongoing threat on Trindade Island, where deforestation driven by fires and goats has already contributed to the extinction of other native seabirds. The authors say the reference genome will support future work on population structure, gene flow, hybridisation, colour polymorphism and the species’ distinctive breeding seasonality, as well as improving read mapping and variant discovery in resequencing studies. Because chromosome-scale organisation was achieved by reference-guided scaffolding rather than direct long-range data, the team notes that scaffold order and orientation should be interpreted with caution until validated with independent methods such as Hi-C. Even so, the resource moves the study of this taxonomically complex group from marker-based inference toward genuinely genome-wide analysis, and adds to a rapidly growing set of genomic tools for some of the ocean’s most imperilled wanderers.

Subject of Research: A high-quality long-read reference genome for the Trindade petrel and its phylogenomic context within Pterodroma

Article Title: The Genome of an Ocean Wanderer: A High‐Quality Long‐Read Reference Genome for the Trindade Petrel (Pterodroma arminjoniana)

Article References: Mazzochi, M. S., Herrera‐Trujillo, O. L., Norris, K., Nicoll, M. A. C., Plaza, P., Stewart, D. A., Dias, M. P., Silva, M. C., & Bugoni, L. (2026). The Genome of an Ocean Wanderer: A High‐Quality Long‐Read Reference Genome for the Trindade Petrel ( Pterodroma arminjoniana ). Ecology and Evolution, 16(10), Article e74568. https://doi.org/10.1002/ece3.74568

Image Credits: AI Generated

DOI: 10.1002/ece3.74568

Keywords: Trindade petrel, Pterodroma arminjoniana, reference genome, Oxford Nanopore sequencing, genomics, seabird conservation, hybridisation, allochrony, phylogenomics, genome assembly, BUSCO, Procellariiformes