Parasites & vectors · 2025
A machine learning-driven early warning system for cryptocaryoniasis in marine aquaculture.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Xie Xiao, Zhang Bo, Wang Xingyu, Jiang Yunyan, Buchmann Kurt, Zhou Suming, Li Yuezhuo, Yin Fei, Galindo-Villegas Jorge
Аннотация
BACKGROUND: Disease outbreaks, particularly cryptocaryoniasis caused by the ciliate Cryptocaryon irritans, pose significant barriers to sustainable marine fish aquaculture, undermining productivity, profitability, and biosecurity. Despite its impact, early warning tools for parasitic diseases leveraging advanced technologies remain underdeveloped.
METHODS: We developed a machine learning (ML)-driven early warning system for cryptocaryoniasis, integrating seven years of outbreak surveillance data (n = 429 events from 2016 to 2023) with 17 high-resolution oceanographic predictors influencing parasite life cycles along China's coast. Five supervised ML models: logistic regression (LR), support vector machine (SVM), random forest (RF), XGBoost (XGB), and artificial neural network (ANN), were trained using cross-validation and benchmarked in commercial open-sea cages and recirculating aquaculture systems (RAS).
RESULTS: The RF model achieved the highest sensitivity (98.6%), with RF and XGB excelling in F1 scores (0.93 and 0.938, respectively), identifying stocking density, water temperature, salinity, pH, and novel predictors such as silicate and nitrate as key risk factors. The predictive engine was deployed as an open-source web-based platform, delivering weekly, spatially resolved outbreak forecasts. Field validation across 12 open-sea cage events and weekly RAS monitoring confirmed high predictive accuracy (91.67% in sea cages; 87.5% in RAS), revealing seasonal and latitudinal disease trends.
CONCLUSIONS: This study establishes a robust, scalable framework for real-time disease forecasting in marine aquaculture, adaptable to other aquatic pathogen-host species to support parasite surveillance and precision health management across diverse global aquaculture systems. While further validation with larger datasets and integration of pathogen and host data will enhance future models, this system provides a flexible foundation for advancing disease control in aquatic environments.
METHODS: We developed a machine learning (ML)-driven early warning system for cryptocaryoniasis, integrating seven years of outbreak surveillance data (n = 429 events from 2016 to 2023) with 17 high-resolution oceanographic predictors influencing parasite life cycles along China's coast. Five supervised ML models: logistic regression (LR), support vector machine (SVM), random forest (RF), XGBoost (XGB), and artificial neural network (ANN), were trained using cross-validation and benchmarked in commercial open-sea cages and recirculating aquaculture systems (RAS).
RESULTS: The RF model achieved the highest sensitivity (98.6%), with RF and XGB excelling in F1 scores (0.93 and 0.938, respectively), identifying stocking density, water temperature, salinity, pH, and novel predictors such as silicate and nitrate as key risk factors. The predictive engine was deployed as an open-source web-based platform, delivering weekly, spatially resolved outbreak forecasts. Field validation across 12 open-sea cage events and weekly RAS monitoring confirmed high predictive accuracy (91.67% in sea cages; 87.5% in RAS), revealing seasonal and latitudinal disease trends.
CONCLUSIONS: This study establishes a robust, scalable framework for real-time disease forecasting in marine aquaculture, adaptable to other aquatic pathogen-host species to support parasite surveillance and precision health management across diverse global aquaculture systems. While further validation with larger datasets and integration of pathogen and host data will enhance future models, this system provides a flexible foundation for advancing disease control in aquatic environments.