Deep-sea Gulf species show genetic signs of extinction risk, study finds

Three lanternfish and two shrimp species sampled in the Gulf of Mexico between 2011 and 2023 had effective population sizes mostly below the 500 that global conservation guidelines set for long-term viability, and one lanternfish lost about 8% of its genetic diversity, yet no population has been seen collapsing toward extinction.

Peer reviewed · Population genomics over 8 to 12 years · Climate projections (modelled) · Five species, one region

Five small animals that live hundreds of metres down in the Gulf of Mexico carry genetic warning signs that conservation guidelines link to higher extinction risk, according to a study published in Nature Communications on 12 November 2026. Pedro A. Peres and colleagues at Florida International University and Nova Southeastern University tracked the DNA of three lanternfish and two shrimp species across more than a decade. All five had small effective population sizes. These are risk indicators and model projections for one region, not observed extinctions, and the cause is still unknown.

Evidence at a glance

Is this a big deal?

A solid first step, with a serious caveat. The authors call long-term genetic records for deep-sea animals virtually non-existent, and the signal of small effective populations shows up in every species and in all three of the team’s main methods. But a genetic red flag is not a body count. Whether these populations are heading for trouble, or have simply been small in genetic terms for thousands of years, is still open.

What scientists found

The key measure is effective population size, written N_E. It is not a head count: it describes how many individuals effectively pass their genes to the next generation, and it is usually far smaller than the number alive. The UN’s Global Biodiversity Framework, adopted in 2022, uses it as a genetic yardstick: an N_E above 50 for short-term survival and above 500 for long-term viability. The paper notes that some researchers argue for higher bars of 100 and 1,000. It is a different yardstick from the Red List threat categories used, for example, in a recent assessment of reef-building corals.

Across three temporal methods, all five species had most N_E estimates below 500, and every species had estimates near or below 50 in at least one time window. A single-sample method gave values from 37.8 to “infinite”; the authors treat the infinite results as unreliable.

Genetic diversity fell in three of the five species, but by different amounts and on different paths:

• Diaphus dumerilii (a lanternfish) dropped 10.9% to 13.1% between 2011 and 2016, then partly recovered, for a net loss of about 7–8% by 2023.
• Ceratoscopelus warmingii (a lanternfish) dipped 3–5% by 2018, then returned to its 2011 level.
• Phorcosergia grandis (a shrimp) ended 1.6–3% below its 2015 level by 2023.
• Lepidophanes guentheri (a lanternfish) and Robustosergia robusta (a shrimp) changed so little that the authors judged it not biologically relevant.

The habitat models point the same way. For four of the five species, the Gulf becomes less suitable in the future, with losses of suitable habitat ranging from 5.2% to 70% depending on the species and scenario. The biggest losses appear in 2100 under the high and extreme emission pathways (RCP 6.0 and RCP 8.5). Phorcosergia grandis was the exception: its models projected the Gulf becoming more suitable, although they disagreed with each other and performed worst of the five.

How they found it

The animals came from the DEEPEND consortium, a monitoring programme set up after the 2010 Deepwater Horizon oil spill. From the research vessel Point Sur, the team towed a net system with six nets that close at chosen depths between the surface and 1,500 metres in the northern and eastern Gulf.

The three lanternfish were sampled in 2011, 2015, 2016, 2018, 2021 and 2023; the two shrimp from 2015 onwards. Some of the 2011–2016 fish DNA had been sequenced in earlier DEEPEND work. The team added the later years and the shrimp, sequencing 432 new animals for a total of 590. After quality filtering, 486 remained, with 5 to 25 animals per species per year.

To estimate N_E, they tracked how gene variants drifted in frequency between sampling years. Picture each generation drawing its genes from a jar of marbles held by the parents. Draw thousands and the colour mix barely shifts; draw a few dozen and rare colours vanish by chance. Effective population size is roughly how many marbles are really being drawn. The analogy has a limit that matters here: real fish and shrimp can arrive from outside the Gulf, bringing new marbles, and nobody yet knows how much that happens for these species.

Two further programs, Stairway Plot 2 and GADMA2, rebuilt each population’s history from its DNA. Both found declines, but they disagree on timing. Stairway Plot shows a slow slide starting roughly 5,000 to 9,000 generations ago, from 200,000 or more to under 200. GADMA2 favours a sudden drop 14 to 91 generations ago, from more than 20,000 to under 500 in three species. The team assumed one generation per year, so generations here roughly equal years.

For the habitat forecasts, the researchers trained three types of model on records of where each species has been found across the Atlantic, from the equator to 65° north, using seven surface and seafloor conditions such as temperature, salinity and chlorophyll. They then projected them to 2050 and 2100 under four emission pathways.

Why it matters to you

You will probably never see a lanternfish, but you live on a planet whose largest habitat they help run. The deep sea, below 200 metres, makes up about 90% of the ocean by the paper’s count. Many lanternfish and shrimp swim up toward the surface at night and back down by day, travelling up to 1,000 metres in 24 hours, and they are food for larger animals. A companion study in PNAS reports that Gulf managers have seen parallel declines in protected species that eat deep-sea prey, including beaked whales and the endangered Rice’s whale.

“This challenges the idea that deep-sea species are naturally more resilient because they occupy such vast environments,” Peres said in FIU’s news release. Nothing here changes what you should do this year.

What the evidence does not show

It does not show that any of these species is going extinct. Low N_E and lost diversity are indicators that conservation biologists associate with higher risk; the paper itself notes that the universal use of the 50/500 thresholds is debated. Low N_E also does not mean few animals. Marine species with huge numbers can have low N_E when a few parents produce most of the young, an idea the authors consider and judge unlikely to explain the whole pattern.

It does not identify a cause. The study has no samples from before the 2010 spill, so the authors say a direct link “cannot be definitively established”. Nor does it show that the habitat losses will happen. The seven model inputs did not include oxygen, depth-specific midwater conditions were unavailable, all species shared one Atlantic training region, and for Diaphus dumerilii the risk that projections stray outside the conditions the models learned from was rated moderate to high.

How this fits what we already knew

The Gulf’s midwater animals were already in trouble by another measure. A PNAS paper published on 19 October 2026 by Tracey Sutton, Bracken-Grissom and colleagues reported that abundances of deep-pelagic fishes, crustaceans and squid fell by 60–85%, depending on the group, over a 3.5-year period, and stayed low for more than a decade. No natural factor they checked (temperature, oxygen or food supply) tracked the decline, and they called the spill, which released 4 to 4.9 million barrels of oil at about 1,500 metres, the most parsimonious explanation.

The new paper adds a genetic layer to that picture. It also offers a possible reason why Diaphus dumerilii fared worst: it migrates less than the others and spends its adult life near the seabed, close to where a deep oil plume about 400 km long lingered at 1,000–1,300 metres and later settled. The authors present that as a possible explanation, not a demonstrated one.

What the peer reviewers said

SciQuest has not yet obtained independent comment. The clearest outside views are in the paper’s open peer review file, from anonymous referees. One wrote in the first round that the N_E results “seem very hard to believe” given the small samples; the authors responded by adding the demographic reconstructions and a second filtering threshold, and that referee was satisfied with the revision. Another judged some conclusions in the second round “overly ambitious and not fully supported by the data presented”, pointing out that Diaphus dumerilii diversity rose again after 2016 while its inbreeding fell, which could suggest recovery; after a further revision, that referee had no more comments. A niche-modelling referee called the single training region for all species “a serious concern”; the published paper discusses it as a source of uncertainty.

How other headlines framed it

• FIU’s own release, republished by Phys.org, spoke of “hidden extinction risks” and focused on the 8% loss in one lanternfish.
• Environment Coastal & Offshore said Gulf deep-sea species are “facing” extinction risk, a firmer framing than the paper, which reports indicators “associated with” higher extinction risk.
• Earth.com widened the result to deep-ocean animals in general, though its body text kept the main caveats.
• Nature World News stayed closest to the evidence, calling them warning signs in five Gulf species.

What happens next

DEEPEND sampling continues and is funded until at least 2029, according to the paper, so later years can test whether Diaphus dumerilii keeps recovering. The authors call for studies of how much these species mix with populations outside the Gulf, which would show whether outside migrants can top up lost diversity. The genetic data are public on NCBI and the habitat-model files on Zenodo, so other teams can rerun the analysis.

In one sentence

Up to 12 years of DNA from five Gulf midwater species show effective population sizes below the usual long-term safety line and real diversity loss in one lanternfish, which is a reason to keep watching the deep sea rather than proof that these animals are disappearing.

Read next

• 44% of assessed reef-building corals are threatened, Nature study finds
• Severe drought raises projected habitat loss for most cacti, study finds
• How deep is the ocean? Average depth, deepest point and each ocean

Sources and disclosure

1. Peres, P. A., Sellers, K., Sutton, T. T. & Bracken-Grissom, H. Extinction risks in deep-sea populations revealed by genomic data and ecological niche modeling. Nature Communications 17, 11243 (2026), published 12 November 2026. SciQuest read the full open-access text, figure legends, methods and data statements.
2. Transparent peer review file for the paper (CC BY 4.0; anonymous referees).
3. Florida International University. FIU scientists uncover hidden extinction risks in the deep sea, 12 November 2026.
4. Sutton, T. T. et al. Immense but not immutable: A sudden and sustained decline in deep-pelagic life in a large oceanic ecosystem. PNAS (2026), published 19 October 2026 (abstract).
5. Convention on Biological Diversity. Kunming-Montreal Global Biodiversity Framework.

The authors declare no competing interests. The work was funded by NOAA’s RESTORE Science Program, the Gulf of Mexico Research Initiative and the NOAA Office of Response and Restoration.

Disclosure: This article was drafted with AI assistance and checked against the sources above. The graphic plots only percentages stated in the paper; the comparison species are those the paper itself cites. Spotted an error? See our corrections policy.