Avian Flu and Zoonoses Overview

Podcast

Play an AI-generated podcast conversation about this lesson

Questions and Answers

What type of viruses cause Avian Flu and how are they classified?

Influenza type A viruses, based on surface proteins hemagglutinin (H) and neuraminidase (N).

Besides temperature and precipitation, what other environmental factors are included as features to understand and predict Avian Flu outbreaks?

NDVI, NDWI, and bioclimatic variables (bio1-bio19)

Besides environmental factors, what are two types of socioeconomic data included as predictive features for the Avian Flu models?

Export, Imports and GDP.

According to the provided text, what is the main goal of the One Health approach to zoonoses?

To reduce the risk at source, moving from response to prevention. Signup and view all the answers

What question is Model I primarily constructed to answer regarding AIV outbreaks?

What eco-climatic and socio-economic variables drive AIV outbreaks and how predictive are they? Signup and view all the answers

What is the term for the key step in the emergence of a zoonotic disease?

Spillover Signup and view all the answers

Name two types of intermediate hosts that can be involved in zoonotic spillover?

Vertebrate and Invertebrate Signup and view all the answers

According to the lecture, what is the total number of known zoonotic disease outbreak events?

13985 Signup and view all the answers

What term is used by the WHO to describe a placeholder for a potential future disease that requires research and development?

Disease X Signup and view all the answers

What environmental factor, according to Mora et al (2022), can aggravate over half of human pathogenic diseases?

Climate change Signup and view all the answers

According to Robinson et al., what factor is connected to increased disease transmission between animals and humans?

Land use and ecosystem degradation Signup and view all the answers

What effect does high biodiversity have on disease transmission, as illustrated by the Lyme disease example?

Dilution effect Signup and view all the answers

Besides environmental changes, what other type of factors can contribute to the increase of disease outbreaks?

Social drivers Signup and view all the answers

How many features does Model 1 use?

49 features Signup and view all the answers

What is one factor that influences the spread of poultry disease according to Model 3?

Trade Signup and view all the answers

What is the prediction accuracy of the model described according to the 'Conclusion' section?

94% Signup and view all the answers

What is the start and end year range for the models described?

2006-2021 Signup and view all the answers

What is one key climate indicator mentioned?

The temperature of the coldest month Signup and view all the answers

According to the SHAP explanation, what does it measure?

The contribution of each feature by considering all possible feature combinations Signup and view all the answers

What is one of the metrics used to compute XGBoost standard feature importance?

Weight (Frequency) Signup and view all the answers

Besides temperature, what are the other two leading variables the model is sensitive to?

Ndvi and Ndwi Signup and view all the answers

What does the 'gain' metric represent in the context of feature importance?

The average improvement in the objective function achieved by splits on a feature. Signup and view all the answers

What does the 'cover' metric in feature importance signify?

The average proportion of samples affected by splits on a feature. Signup and view all the answers

How does SHAP feature importance differ from standard XGBoost feature importance in its approach to feature interactions?

SHAP considers feature interactions, whereas standard XGBoost importance does not. Signup and view all the answers

What properties are used by SHAP to produce consistent and fair attributions?

Additivity, symmetry, and efficiency. Signup and view all the answers

How does SHAP handle correlated features compared to the XGBoost standard importance?

SHAP distributes importance fairly among correlated features, while XGBoost might overestimate the importance of highly correlated features. Signup and view all the answers

What does it mean to decompose individual predictions using SHAP values?

It means giving an explanation of why a particular prediction was made by showing each feature's contribution. Signup and view all the answers

Does XGBoost’s standard feature importance provide local or global explanations?

Only global explanations. Signup and view all the answers

What type of explanations does SHAP offer?

Both local explanations for individual predictions and global feature importance. Signup and view all the answers

Study Notes

Lecture 8: Integrated Data Analysis and Outbreak Prediction: Avian Flu

The lecture is about integrated data analysis and outbreak prediction, focusing on avian flu.
The presenter is Marina Treskova, PhD, Head of Research Group, Eco-Epidemiology.
The research team is associated with the Heidelberg Institute of Global Health & Interdisciplinary Centre for Scientific Computing, Heidelberg University.
Website: hei-planet.com

Spillover and Zoonotic Disease Emergence

Spillover is a key step in zoonotic disease emergence.
Spillover cascade: The process starts with infection of wildlife, then shedding of the pathogen (animal-to-animal transmission), spillover to humans or other animals, followed by spread within human populations.
Various pathways for zoonotic spillover:
- Direct transmission
- Vertebrate intermediate host
- Invertebrate intermediate host
- Environment

Known Zoonotic Diseases

The number of included diseases is 203.
Of these, 70 are viral, 47 bacterial, and 74 parasitic.
The total number of outbreak events is 13985.
Zoonotic spillover is a key public health challenge

WHO Priority Diseases

Diseases highlighted by WHO include Ebola virus, Lassa fever, Crimean-Congo haemorrhagic fever, Middle East respiratory syndrome coronavirus (MERS-CoV), Nipah virus infection, Rift Valley fever, and Zika virus.

Clusters of Zoonotic Spillover

Various factors are associated with zoonotic spillover.
Factors include viral adaptation, climate and vectors, food & livestock, environmental contact, global movement, socioeconomics, and land conversion.

Climate Change and Human Pathogenic Diseases

Over half of known human pathogenic diseases can be aggravated by climate change.
Key factors associated with climate change are warming, precipitation, floods, droughts, storms, and natural cover change.

Land Use and Ecosystem Degradation

Ecosystem degradation and human encroachment contribute to zoonotic spillover risk.
Ecosystem restoration and biodiversity conservation are crucial to reduce spillover risk.

Biodiversity Loss and Dilution Effect

Lyme disease is used as an example
High biodiversity reduces reservoir density. This makes zoonotic disease transmission less likely.
Low biodiversity increases reservoir density making zoonotic disease transmission more likely.

Globalization and environment: Deforestation, mining and dams, and ecosystem degradation can change vector and non-human host habitats. International travel and trade, urbanization, and population displacement spread pathogens and increase risk factors.
Sociodemographic factors: Children, the elderly, and pregnant women are more vulnerable. Quality of housing, exposure to vectors, baseline incidence of disease, population health status, and humanitarian crises impact vulnerability.

Public Health Systems

Surveillance, early warning systems and vector control are important to prevent disease.
Access to and quality of healthcare can affect severity of infections.
Research is key to effectively control and prevent these diseases.

One Health Approach to Zoonoses and Prevention

This approach emphasizes reducing disease risk at the source, not just responding to outbreaks after they have happened.

Avian Flu Model

The avian flu model includes several features, and focuses on how well these can predict future outbreaks.
3 models were built for this study.
Model 1: focuses on eco-climatic and socio-economic variables, and determining predictability levels.
Model 2: focuses on wild bird outbreaks.
Model 3: focuses on wild bird outbreaks but removes bird labels

Model Metrics

Evaluation metrics for different models. Logloss was used to measure the effectiveness of models.

Main Results

Model 1, 2, and 3 results are highly accurate when predicting Avian Influenza outbreaks in Europe.
The model is climate sensitive.
The key climate indicators are the temperature of the coldest month, the mean temperature of quarter two, and the minimum temperature.
Climate, Environmental, and Bioclimatic features are most important for the model.
Trade and Economic factors are also significant.

SHAP vs Importance

Differences between SHAP feature importance and XGBoost standard feature importance
SHAP considers feature correlations
XGBoost does not
SHAP provides both local and global feature importance. XGBoost's feature importance is mainly global.

Studying That Suits You

Use AI to generate personalized quizzes and flashcards to suit your learning preferences.

Description

Test your knowledge on Avian Flu, its classification, and the environmental and socioeconomic factors that influence its outbreaks. This quiz also covers the One Health approach and terminology related to zoonotic diseases. Ideal for students and professionals interested in infectious diseases.