CIEXPO-Intake: Chemical Insights Exposure Portfolio: Intake Database

In SEEM, measured concentrations of chemicals in human blood and urine (specifically from the U.S. National Health and Nutrition Examination Survey) are converted to population intake inferences using toxicokinetic models (Stanfield et al. <i>J Expo Sci Environ Epidemiol</i> <b>32</b>, 833–846,2022), providing empirical exposure benchmarks against which high-throughput exposure models can be evaluated, calibrated, and weighted in the consensus prediction.

SEEM uses machine learning to predict whether chemicals are likely to be associated with major exposure pathways, including consumer, dietary, industrial, and pesticide-related pathways, and propagates this pathway uncertainty into the final exposure estimate. SEEM provides high-throughput exposure predictions for thousands of chemicals and provides exposure estimates that can be integrated with toxicological and toxicokinetic data to support chemical prioritization and risk assessment.

CIEXPO-Intake provides schema for organizing predictions from multiple versions of SEEM, including v2 (demographic-specific), v3, and v3.1 predictions (for select chemicals, as described in the below submitted publication), along with exposure inferences and pathway probabilities.</p><p dir="ltr">CIEXPO-Intake is distributed in portable DuckDB format. DuckDB databases can be accessed programmatically using freely available DuckDB packages for R and Python, with data queried using SQL or higher-level interfaces such as dplyr/dbplyr in R and SQLAlchemy or Ibis in Python.

A data dictionary is provided.</p><p dir="ltr">Version History:</p><p dir="ltr">V0: Version accompanying submission of the following manuscript: Isaacs, KK, Wambaugh JF, Charest N, Sinclair G, Shapiro AJ, Whitehead H, Boatman A, Ferland TM, Wall JT, Wilson M, Sams RL, Edwards S, Thomas RS, and Paul Friedman K. Characterizing Exposure Pathways for a Global Collection of Chemicals. 2026. Submitted to <i>Environmental Science and Technology.</i></p><p><br></p>