Machine-learning model led design to experimentally test species thermal limits: The case of kissing bugs (Triatominae)

Precision Prediction for Dengue Fever in Singapore: A Machine Learning Approach Incorporating Meteorological Data

OBJECTIVE

This study aimed to improve dengue fever predictions in Singapore using a machine learning model that incorporates meteorological data, addressing the current methodological limitations by examining the intricate relationships between weather changes and dengue transmission.

METHOD

Using weekly dengue case and meteorological data from 2012 to 2022, the data was preprocessed and analyzed using various machine learning algorithms, including General Linear Model (GLM), Support Vector Machine (SVM), Gradient Boosting Machine (GBM), Decision Tree (DT), Random Forest (RF), and eXtreme Gradient Boosting (XGBoost) algorithms. Performance metrics such as Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and R-squared (R2) were employed.

RESULTS

From 2012 to 2022, there was a total of 164,333 cases of dengue fever. Singapore witnessed a fluctuating number of dengue cases, peaking notably in 2020 and revealing a strong seasonality between March and July. An analysis of meteorological data points highlighted connections between certain climate variables and dengue fever outbreaks. The correlation analyses suggested significant associations between dengue cases and specific weather factors such as solar radiation, solar energy, and UV index. For disease predictions, the XGBoost model showed the best performance with an MAE = 89.12, RMSE = 156.07, and R2 = 0.83, identifying time as the primary factor, while 19 key predictors showed non-linear associations with dengue transmission. This underscores the significant role of environmental conditions, including cloud cover and rainfall, in dengue propagation.

CONCLUSION

In the last decade, meteorological factors have significantly influenced dengue transmission in Singapore. This research, using the XGBoost model, highlights the key predictors like time and cloud cover in understanding dengue's complex dynamics. By employing advanced algorithms, our study offers insights into dengue predictive models and the importance of careful model selection. These results can inform public health strategies, aiming to improve dengue control in Singapore and comparable regions.

Positive associations matter: Microbial relationships drive tick microbiome composition

Untangling how factors such as environment, host, associations among bacterial species and dispersal predict microbial composition is a fundamental challenge. In this study, we use complementary machine-learning approaches to quantify the relative role of these factors in shaping microbiome variation of the blacklegged tick Ixodes scapularis. I. scapularis is the most important vector for Borrelia burgdorferi (the causative agent for Lyme disease) in the U.S. as well as a range of other important zoonotic pathogens. Yet the relative role of the interactions between pathogens and symbionts compared to other ecological forces is unknown. We found that positive associations between microbes where the occurrence of one microbe increases the probability of observing another, including between both pathogens and symbionts, was by far the most important factor shaping the tick microbiome. Microclimate and host factors played an important role for a subset of the tick microbiome including Borrelia (Borreliella) and Ralstonia, but for the majority of microbes, environmental and host variables were poor predictors at a regional scale. This study provides new hypotheses on how pathogens and symbionts might interact within tick species, as well as valuable predictions for how some taxa may respond to changing climate.

Interpretable machine learning applied to on-farm biosecurity and porcine reproductive and respiratory syndrome virus

Effective biosecurity practices in swine production are key in preventing the introduction and dissemination of infectious pathogens. Ideally, on-farm biosecurity practices should be chosen by their impact on bio-containment and bio-exclusion; however, quantitative supporting evidence is often unavailable. Therefore, the development of methodologies capable of quantifying and ranking biosecurity practices according to their efficacy in reducing disease risk has the potential to facilitate better-informed choices of biosecurity practices. Using survey data on biosecurity practices, farm demographics, and previous outbreaks from 139 herds, a set of machine learning algorithms were trained to classify farms by porcine reproductive and respiratory syndrome virus status, depending on their biosecurity practices and farm demographics, to produce a predicted outbreak risk. A novel interpretable machine learning toolkit, MrIML-biosecurity, was developed to benchmark farms and production systems by predicted risk and quantify the impact of biosecurity practices on disease risk at individual farms. By quantifying the variable impact on predicted risk, 50% of 42 variables were associated with fomite spread while 31% were associated with local transmission. Results from machine learning interpretations identified similar results, finding substantial contribution to predicted outbreak risk from biosecurity practices relating to the turnover and number of employees, the surrounding density of swine premises and pigs, the sharing of haul trailers, distance from the public road and farm production type. In addition, the development of individualized biosecurity assessments provides the opportunity to better guide biosecurity implementation on a case-by-case basis. Finally, the flexibility of the MrIML-biosecurity toolkit gives it the potential to be applied to wider areas of biosecurity benchmarking, to address biosecurity weaknesses in other livestock systems and industry-relevant diseases.