Back

Phenomic environment-wide association study (PheEWAS) models complexity in the exposome

Palmiero, N.; Gonzalez Zarzar, T.; Passero, K.; Zhou, J.; Crawford, D.; Hall, M.

2025-06-14 health informatics
10.1101/2025.06.13.25329592 medRxiv
Show abstract

Phenome-wide association studies (PheWAS) have successfully identified genomic-based interrelationships between phenotypes but seldom consider environmental exposures. Here, in a phenomic environment-wide association study (PheEWAS), we interrogated relationships between 326 exposures and 55 phenotypes for [~]19,000 participants of the National Health and Nutrition Examination Survey (NHANES). Linear regression models adjusted for age, sex, socioeconomic status, BMI, race/ethnicity, and survey year identified and replicated 106 significant exposure-phenotype associations after Bonferroni correction. The top association was for alpha-tocopherol (vitamin E) with triglycerides (Discovery p = 1.16 x 10-{superscript 1}{superscript 1}; Replication p = 8.05 x 10-{superscript 1}3). The exposure retinol (vitamin A) had the largest number of individual replicating associations (14 phenotypes including total calcium, iron-binding capacity, ferritin, albumin, transferrin saturation, creatinine, gamma-glutamyl transferase, triglycerides, uric acid, alkaline phosphatase, hemoglobin, and blood urea nitrogen). The phenotype with the greatest number of exposure associations was homocysteine (associated with thiamine; alpha- and gamma-tocopherol; dietary fiber, protein, and potassium; riboflavin; cotinine; folate; phosphorus; cadmium; iron intake; supplement count; and niacin). A race/ethnicity-stratified analysis revealed 11 unique population-specific associations. Our findings demonstrate PheEWAS a method to provide new details on the complexity of the exposome at the level of the phenome

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.