AI-Powered Environmental Exposomics: PFAS, Chemical Mixtures & Exposure Risk Analysis

Aim

Workshop Objectives

Workshop Structure

Day 1: Environmental Exposomics, PFAS Biomonitoring & Python-Based Exposure Profiling

Session Focus: Understand environmental PFAS exposure and analyze public biomonitoring datasets using Python.

Topics Covered

• Environmental Exposomics: Exposome concepts, environmental exposures and public health applications.
• PFAS & Emerging Contaminants: PFOA, PFOS, PFHxS, PFNA, environmental persistence and exposure pathways.
• Environmental & Human Biomonitoring: Water, soil, air and biological exposure matrices.
• Public Exposure Datasets: CDC NHANES, PFAS concentration measurements and exposure indicators.
• Data Preprocessing: Missing values, detection limits, concentration units and data transformation.
• Python-Based Exposure Analysis: Concentration distributions, exposure variability and chemical correlations.

🛠️ Hands-On: PFAS Biomonitoring & Exposure Data Visualization

Analyze a curated CDC NHANES PFAS dataset using Python to clean exposure measurements, examine concentration distributions and generate comparative plots and chemical correlation heatmaps.

Tools: Google Colab, Python, Pandas, NumPy, Seaborn, CDC NHANES.

Day 2: AI-Driven Chemical Mixture Analysis, PCA & Exposure Pattern Discovery

Session Focus: Investigate multi-chemical exposure patterns using statistical analysis and unsupervised machine learning.

Topics Covered

• Chemical Mixture Exposomics: Co-exposure, cumulative exposure and mixture-related toxicological challenges.
• Chemical Exposure Relationships: Correlation patterns, chemical co-occurrence and exposure variability.
• Mixture Data Preprocessing: Missing-value handling, normalization, log transformation and feature scaling.
• Principal Component Analysis (PCA): Dimensionality reduction and identification of major exposure patterns.
• K-Means Clustering: Identification of exposure groups with similar chemical profiles.
• Exposure Pattern Interpretation: PCA loadings, cluster characteristics and chemical mixture visualization.

🛠️ Hands-On: AI-Based Chemical Mixture Profiling & Exposure Clustering

Apply PCA and K-means clustering to a curated multi-PFAS exposure dataset, identify major co-exposure patterns and generate PCA plots, cluster visualizations and chemical exposure heatmaps.

Tools: Google Colab, Python, Pandas, Scikit-learn, Seaborn, Matplotlib.

Day 3: Predictive Exposure Modeling, Toxicity Evidence & Chemical Risk Prioritization

Session Focus: Apply supervised machine learning and computational toxicology resources to investigate exposure-related predictions and prioritize environmental contaminants.

Topics Covered

• Environmental Risk Assessment: Hazard identification, exposure characterization, dose-response concepts and risk interpretation.
• PFAS Risk Evaluation: Exposure indicators, matrix-specific screening values and uncertainty considerations.
• Computational Toxicology: EPA CompTox, ToxCast bioactivity indicators and chemical hazard evidence.
• Predictive Machine Learning: Random Forest modeling, feature selection and train-test validation.
• Model Evaluation & Explainability: Prediction accuracy, appropriate performance metrics and permutation feature importance.
• Chemical Prioritization: Integration of hazard, exposure and evidence-quality indicators for exploratory screening.
• Emerging Research Applications: Exposome-Wide Association Studies (ExWAS), explainable AI and computational environmental health research.

🛠️ Hands-On: Machine Learning-Based Exposure Prediction & Chemical Prioritization

Develop and evaluate a Random Forest model using a curated exposure-related dataset with a clearly defined outcome. Interpret influential predictors and generate an exploratory chemical-prioritization table using selected exposure and toxicological evidence.

Tools: Google Colab, Python, Pandas, Scikit-learn, Matplotlib, EPA CompTox.

Who Should Enrol?

Important Dates

Workshop Outcomes

Fee Structure