Environmental Exposomics in R: PFAS, Chemical Mixtures and Metabolomics

To equip participants with an integrated understanding of environmental exposomics by connecting chemical exposure profiling, exposure-associated metabolic features, and joint mixture analysis through reproducible, guided analytical workflows.

• Introduce environmental exposomics, biomonitoring, and the interpretation of chemical exposure measurements.
• Explain approaches for handling detection limits, missing values, skewed distributions, and correlated exposures.
• Develop familiarity with metabolomics feature tables, quality-control principles, preprocessing, and exploratory analysis.
• Demonstrate covariate-adjusted exposure–metabolite association analysis and false discovery rate correction.
• Introduce linear and nonlinear exposure–response modeling and major chemical mixture analysis methods.
• Apply quantile g-computation to estimate a joint mixture association and interpret component contributions.
• Strengthen scientific interpretation and reporting through consideration of confounding, uncertainty, annotation confidence, and study limitations.

📅 Day 1: Environmental Exposure Profiling & Chemical Mixture Characterization

• Focus: Understanding environmental exposures, biomonitoring data, and chemical co-exposure patterns through a reproducible analytical workflow.
• Introduction to environmental exposomics and its role in environmental health research.
• External exposures, internal chemical burden, and biomarkers of biological response.
• Priority contaminants: PFAS, metals, pesticides, and plastic-associated chemicals.
• Exposure routes, biological sampling matrices, and exposure timing.
• Principles of targeted biomonitoring, suspect screening, and non-targeted analysis.
• Understanding exposure-data tables, sample metadata, and relevant covariates.
• Handling detection limits, missing values, skewed distributions, and measurement units.
• Data transformation and standardization for exposure profiling.
• Correlation analysis, co-exposure patterns, and multicollinearity.
• Formulating research questions for single-chemical and mixture-based investigations.

🛠️ Hands-on: Profile Environmental Chemical Exposure Patterns

• Explore a prepared biomonitoring dataset in a guided notebook.
• Apply selected data-cleaning and transformation steps.
• Generate exposure-distribution plots and a correlation heatmap.
• Identify correlated chemicals and define a mixture-analysis question.

🧰 Tools Covered: R, tidyverse, ggplot2, corrplot, EPA CompTox Chemicals Dashboard.

📌 Output: An environmental exposure profile with distribution plots, a correlation heatmap, and identified co-exposure patterns.

📅 Day 2: Metabolomics Data Analysis & Exposure-Associated Metabolic Signatures

• Focus: Assessing metabolomics data quality and identifying metabolic features associated with environmental exposures.
• Principles of targeted and untargeted metabolomics in exposomics research.
• Distinguishing endogenous metabolites, xenobiotics, and xenobiotic metabolites.
• Understanding LC–MS feature tables, sample metadata, and metabolite annotations.
• Quality-control concepts: blanks, pooled QC samples, missingness, and analytical variability.
• Normalization, transformation, and scaling of metabolomics data.
• Recognizing batch effects, technical variation, and sample outliers.
• Principal component analysis for exploring metabolic variation.
• Covariate-adjusted exposure–metabolite association analysis.
• Effect-size interpretation, confidence intervals, and false discovery rate correction.
• Metabolite identification confidence, pathway context, and candidate biomarker interpretation.

🛠️ Hands-on: Identify Exposure-Associated Metabolic Features

• Analyze a prepared metabolomics dataset linked to exposure metadata.
• Apply predefined quality-filtering and preprocessing steps.
• Generate a PCA visualization of sample patterns.
• Run adjusted association models and correct for multiple testing.
• Interpret selected findings using a supplied metabolite-annotation table.

🧰 Tools Covered: R, limma, tidyverse, ggplot2, MetaboAnalyst 6.0.

📌 Output: A PCA plot and a ranked exposure–metabolite association table with effect estimates, adjusted significance values, and available annotations.

📅 Day 3: Dose–Response Modeling & Joint Chemical Mixture Analysis

• Focus: Modeling exposure-related biological responses and interpreting joint mixture associations with appropriate statistical uncertainty.
• Distinguishing administered dose, measured concentration, and exposure biomarkers.
• Dose–response analysis in experiments versus exposure–response analysis in observational studies.
• Selecting appropriate metabolic or health-related response variables.
• Linear and nonlinear response relationships.
• Spline-based modeling and visualization of exposure–response curves.
• Interpreting confidence intervals, data coverage, and apparent response patterns.
• Overview of weighted quantile sum regression, quantile g-computation, and Bayesian kernel machine regression.
• Guided application of quantile g-computation to estimate a joint mixture association.
• Interpretation of positive and negative component contributions.
• Model assumptions, confounding, association versus causation, and reproducible reporting.

🛠️ Hands-on: Model Chemical Mixtures & Exposure–Response Relationships

• Fit a quantile g-computation model using a prepared teaching dataset.
• Visualize the joint mixture estimate and component contributions.
• Compare linear and spline-based response curves for one selected exposure.
• Prepare a concise interpretation of the findings and their limitations.

🧰 Tools Covered: R, qgcomp, splines, ggplot2.

📌 Output: A joint mixture-association plot, an exposure–response curve, and a concise scientific results statement.

• Postgraduate students, Ph.D. scholars, and researchers in environmental science, toxicology, public health, epidemiology, and exposure science.
• Researchers studying PFAS, metals, pesticides, plastic-associated chemicals, or environmental chemical mixtures.
• Scientists working in metabolomics, analytical chemistry, biomonitoring, and environmental health.
• Bioinformaticians, biostatisticians, and data analysts interested in exposure-linked omics and mixture modeling.
• Academicians and research professionals seeking practical experience in reproducible environmental data analysis.

By the end of the workshop, participants will be able to:

• Explain the relationships among external exposures, internal chemical burden, and biomarkers of biological response.
• Prepare biomonitoring data for guided analysis using appropriate cleaning and transformation steps.
• Generate exposure-distribution plots and correlation heatmaps to characterize co-exposure patterns.
• Assess prepared metabolomics data using quality-control information, missingness, and analytical variability.
• Apply predefined preprocessing steps and use PCA to explore metabolic variation.
• Run covariate-adjusted exposure–metabolite models and interpret effect estimates, confidence intervals, and FDR-adjusted significance values.
• Interpret selected metabolic features using available annotations and biological context.
• Fit a guided quantile g-computation model and interpret the joint mixture estimate and directional component contributions.
• Compare linear and spline-based exposure–response curves while considering uncertainty and data coverage.
• Prepare a concise scientific results statement that distinguishes statistical associations from causal conclusions.