For a fish living near a busy coastline, finding enough food does not necessarily mean living in a healthy environment. A new study of reef fish around Okinawa, Japan, suggests that animals near urbanised shores can show molecular signs of plentiful nutrition at the same time as their cells display responses associated with physiological stress.
Published on 9 October 2026 in Nature Communications, the research compared gene activity in blue damselfish collected from 18 coastal sites, spanning relatively natural reefs and heavily developed shores. The scientists combined field sampling with controlled feeding experiments to distinguish signals related to food availability from those associated with human activity. Their results offer a more complicated picture than the familiar assumption that more food necessarily means a healthier habitat.
The findings do not show that every urban reef is harmful or that the fish were starving, poisoned or dying. Instead, they reveal how a coastal habitat can simultaneously provide resources and impose biological costs, and why traditional water-quality measurements may miss some of those trade-offs.
Why the condition of a reef fish can be difficult to judge
Coastal development changes marine environments in several ways at once. Shoreline construction modifies habitat structure, wastewater and runoff can alter nutrients and contaminants, and urban waters can be warmer or otherwise different from nearby natural habitats. Fish experience these changes as a combined environment rather than as one isolated pollutant.
Conventional environmental monitoring often focuses on conditions outside the animal, such as water temperature, salinity or chemical concentrations. Ecologists may also count fish and record which species are present. These approaches are valuable, but a species can remain visible and abundant even when individual animals are making physiological adjustments to cope with difficult conditions.
The researchers therefore asked a different question: what can patterns of gene expression, the activity of genes inside an organism, reveal about how fish are responding to the environments they actually inhabit? Gene expression is not the same as a change to the fish’s DNA. It is a dynamic indication of which biological processes are being activated or reduced under particular conditions.
How the study compared 18 Okinawa coastal sites
Emma Gairin, Saori Miura, Zoé Chamot and colleagues studied Chrysiptera cyanea, a common blue damselfish found around Okinawa. The team sampled five juvenile fish at each of 18 sites, giving 90 juvenile field samples, as well as four or five adults per site. Sampling took place from May to July 2023, during the species’ reproductive season.
The researchers analysed whole-body gene activity in juveniles and liver gene activity in adults. These are not identical biological measurements. Whole-body profiles offer a broad picture of developing fish, while adult liver tissue is particularly informative about metabolism, detoxification and longer-term physiological adjustment. Comparing them can identify patterns that differ with life stage.
At the same time, the team measured environmental characteristics including water temperature, salinity and chlorophyll-related conditions. To estimate nearby human development, they calculated the proportion of natural land cover within one kilometre of each site. This was a proxy for urbanisation, not a direct measurement of every contaminant or activity to which an individual fish had been exposed.
The researchers also used field surveys of fish communities and seabed habitat. Their analysis considered factors that might otherwise be mistaken for urbanisation effects, including sampling time, fish size and adult sex. This matters because a change in gene activity could reflect a hot day, a smaller animal or its feeding condition rather than proximity to a city.
Gene activity differed strongly between locations
Fish collected at the same locations had relatively similar gene-expression patterns, and those patterns varied markedly across sites. In a statistical analysis of the main variation in the transcriptome data, site differences accounted for a substantial proportion of the observed pattern: the researchers reported R² = 0.76 for juvenile whole-body samples and R² = 0.67 for adult liver samples, with p = 0.001 in each case.
These R² figures describe variation in the particular multivariate analysis, not the percentage of fish made ill by urban development. They establish that location was strongly associated with molecular profiles, while leaving open the question of which environmental factors produced each difference.
Juveniles and adults also showed different patterns of consistency across sites. The team found 2,410 genes in juvenile samples and 251 genes in adult livers that differed in more than one-quarter of the 153 possible pairwise site comparisons. The researchers interpreted this as evidence that young fish may exhibit broader molecular plasticity, whereas adult liver responses can be more specialised and location-specific.
A feeding experiment helped interpret the wild fish
One of the study’s strongest features was the effort to give the field data a biological reference point. The scientists conducted separate controlled feeding experiments in which juvenile and adult fish were either fed normally or fasted. Juveniles were fasted for three days and adults for 14 days, with the different periods selected for their physiological tolerance.
The resulting gene-expression profiles separated the fed and fasted groups. The researchers then identified the genes most strongly distinguishing those conditions and compared that reference pattern with fish collected in the wild. In this way, they developed a molecular proxy for feeding status.
Crucially, this is not a direct record of what each wild fish ate or a measurement of calories consumed. It is an inference based on how closely its gene activity resembled that of experimentally fed or fasted animals. That distinction is central to interpreting the headline finding.
Adult fish near developed coasts showed stronger feeding signals
Adult fish from more urbanised locations had gene-expression profiles closer to the well-fed laboratory reference. In the researchers’ mixed-effects model, the association between the proportion of nearby natural land and the feeding-status proxy was β = −0.066, with a 95% confidence interval from −0.105 to −0.027 and p = 0.0023. The negative coefficient means that lower nearby natural land cover, used here to indicate more urban influence, was associated with a higher inferred feeding status.
The result is consistent with the possibility that developed coastlines can offer accessible food resources, perhaps through changes to nutrients or food webs. It does not establish the exact food source or prove that urban runoff improved the fish’s diet. The authors noted gene patterns involving energy balance and metabolic regulation that were consistent with the feeding interpretation.
For juveniles, the picture was different. The team did not find a significant overall relationship between the measured environmental parameters and the feeding-status proxy. However, 46 pairs of sites differed significantly in that proxy after the reported statistical comparisons, suggesting that local conditions still mattered. The researchers also examined non-fish sequence reads as supporting evidence about possible dietary organisms in juvenile samples.
These distinctions matter because it would be misleading to say that urbanisation improved feeding for all ages of fish. The stronger urbanisation-related pattern was reported in adults, while juvenile feeding signals varied more between individual sites.
The hidden cost appeared in immune and inflammatory pathways
The researchers then examined which genes were associated with nearby human activity after accounting for other measured influences. They identified 425 genes in juvenile whole-body profiles and 585 genes in adult liver profiles associated with the urbanisation proxy, without a detected association with the other parameters tested in their models.
Those genes included processes related to immune defence, inflammation, cell signalling and programmed cell death. Near urbanised sites, some pathways associated with inflammation and reactive oxygen responses were more active, while some processes involved in tissue maintenance and normal cellular function showed lower activity. The authors interpreted these patterns as potential molecular signs of chronic physiological pressure.
These gene counts should not be mistaken for 1,010 damaged genes, nor do they prove that any particular chemical caused disease. Gene-expression changes can reflect adaptation and regulation as well as harmful exposure. The important observation is the combination of apparent nutritional advantage with molecular pathways that suggest stress, a trade-off that might not be obvious from a visual inspection of the animals.
Temperature complicated the pollution story
One striking result was that many familiar molecular indicators of stress did not map neatly onto urban development. The researchers found that temperature and feeding status influenced several processes commonly used as indicators of pollution or physiological strain.
Temperature alone was associated with 1,854 genes in juvenile whole-body samples and 1,312 genes in adult livers that were not associated with the other modelled factors. Removing the four warmest sites, where temperatures ranged from approximately 29.6°C to 32.4°C, reduced the number of temperature-associated genes by almost two-thirds.
That result illustrates why an apparent pollution signal must be interpreted carefully. If a gene responds strongly to heat, it may change in a developed coastal area even without being a specific marker of chemical contamination. Likewise, a fish with less available food may alter immune activity because maintaining an immune response requires energy.
The researchers’ approach tried to separate these overlapping influences instead of assigning every molecular difference to a single environmental cause. The study therefore provides a case for using groups of biological indicators, supported by environmental measurements and controlled experiments, rather than relying on one supposedly universal stress marker.
What this means for environmental monitoring
The findings suggest that monitoring marine ecosystems could eventually move beyond measuring the water around an animal to examining how the animal is responding to its combined environment. This field, sometimes called landscape transcriptomics, uses gene-activity profiles collected across different locations to identify biological patterns linked to environmental conditions.
For conservation agencies, such methods could complement traditional surveys. A reef might still support fish while exposing them to physiological demands that are not visible in population counts. Conversely, a site that looks degraded may offer food resources that help explain why animals remain there. Both facts are relevant to understanding ecosystem resilience.
However, molecular monitoring is not a replacement for direct water-quality testing, habitat surveys or long-term population tracking. The strongest approach would combine them. Readers interested in the broader evidence linking biodiversity to ecosystem functions can also see Research Today’s coverage of a global synthesis of 423 biodiversity studies.
Why the results matter beyond Okinawa
Coastal urbanisation is a global issue. Ports, tourism, housing, agriculture and transport infrastructure bring multiple pressures to marine habitats, often at the same time. The Okinawa study offers a way to ask how those pressures register inside wildlife, rather than assuming that one environmental measurement captures the full experience of an organism.
The authors also highlight the importance of age and life stage. A young fish may respond quickly to its immediate nutritional circumstances, while an adult liver can reflect a different balance of metabolic and immune processes. Conservation monitoring that samples only one life stage could miss these differences.
The practical implication is not that managers should encourage nutrient pollution because some fish appear well fed. Additional food resources can coexist with physiological costs, and the study did not evaluate whether the net result improved reproduction, survival or population growth. Better feeding signals should not be equated with better long-term ecological health.
Important limitations of the evidence
The field component was observational and involved one fish species around one Japanese island. The 18 sites offered a useful range of environments, but they do not represent all reefs, coastlines or marine animals. The urbanisation measure was based on nearby land cover, so it could not identify the exact mixture of contaminants, nutrients or habitat changes experienced by each fish.
Gene expression is a snapshot influenced by timing, temperature, nutrition and individual biology. The researchers modelled several of these influences, but unmeasured factors and complex interactions remain possible. Their laboratory feeding experiments strengthened interpretation of the nutritional signal, yet the proxy does not replace direct measurement of wild fish diets. Nor do stress-related gene signatures by themselves establish clinical illness, reduced lifespan or population decline.
The results also require caution when comparing juveniles and adults because different tissues were analysed. A whole-body juvenile profile and an adult liver profile are informative but not directly interchangeable measurements of the same biological processes.
The wider lesson
The study challenges a simple equation between food availability and environmental quality. Fish living near human activity may gain access to resources while making costly biological adjustments to their surroundings. Across 18 Okinawa reef sites, gene activity revealed both sides of that possibility.
For environmental researchers, the advance is methodological as well as ecological: combining wild-animal molecular data with controlled reference experiments can help interpret the pressures acting on organisms in complex real-world habitats. For the public, the lesson is that wildlife can appear to be coping while its physiology tells a more complicated story.
Source Information
Study: Landscape transcriptomics reveals ecological constraints on fish under anthropogenic coastal change.
Authors: Emma Gairin, Saori Miura, Zoé Chamot, Camille A. Sautereau, Jann Zwahlen, Hiroki Takamiyagi, Yann Gibert, Marcela Herrera and Vincent Laudet.
Journal: Nature Communications, volume 17, article 10218 (2026).
Publication date: 9 October 2026.
Research type: Peer-reviewed observational field study with controlled fasting and feeding experiments and transcriptomic analysis.
Study location and sample: 18 coastal sites around Okinawa, Japan, with five juvenile blue damselfish and four or five adult fish sampled per site.
DOI: 10.1038/s41467-026-77751-2.
