DNA methylation at 109 sites helped estimate haddock age without removing the fish’s ear stones.
Researchers collected pectoral-fin samples from haddock during two bottom-trawl surveys in 2022 and 2023. They measured DNA methylation across the genome and used statistical models to identify sites associated with age, while accounting for fish length and sex.
The final age-prediction model used 109 methylation sites. It estimated haddock age with a mean absolute error of 0.17 years, offering a non-lethal alternative to ageing fish by removing and examining their otoliths, or ear stones.
Findings from fin tissue
The researchers analyzed whole-genome DNA methylation data from 140 haddock selected across the species’ age range. After quality filtering, 133 fish were used for methylation analysis, and the final model identified 109 CpG sites whose methylation patterns predicted age.
The model had an R² of 0.98 and a mean absolute error of 0.17 years. Age-related sites included both increases and decreases in methylation and were spread throughout the genome. They were more common in parts of genes involved in development, immunity, gene regulation and cell signaling.
Overall DNA methylation increased slightly during the first five years of life and then leveled off through age 10.2 years. This increase was more pronounced in males, although the researchers’ age-prediction model was described as independent of sex and length.
Why non-lethal ageing matters
Fisheries population models use age information to estimate stock size, population changes and sustainable catch recommendations. The usual method for haddock—counting growth rings in otoliths—requires killing the fish and can be labor-intensive or inaccurate, especially when rings are difficult to distinguish.
A reliable age estimate from a fin sample could allow researchers to collect age information without sacrificing sampled fish. The study presents the method as a potentially cost-effective and scalable tool for haddock fisheries management, although it is specific to this species and still requires appropriate sampling and analysis.
Evidence and caveats
The study analyzed fish collected in fall 2022 and spring 2023 bottom-trawl surveys from the Georges Bank–Gulf of Maine region. Otolith annuli counts supplied the age estimates used to develop and assess the methylation model. The researchers used a 70/30 training-test split; the reported median mean absolute error was 0.11 years in the training set and 0.55 years in the testing set, while the final model’s overall reported mean absolute error was 0.17 years.
The work therefore provides evidence for a species-specific prediction method in the sampled haddock, rather than proof that the same accuracy will apply to all haddock populations or other fish. The model also depends on laboratory methylation measurements and was developed from fish collected in one broad survey region and two sampling seasons.