Dietary Factors and NAFLD Risk from NHANES 2017-2023 and Mendelian Randomization

A dataset by Siqi Xie, last updated in May 2026, combines observational and genetic data to investigate causal dietary risk factors for non-alcoholic fatty liver disease (NAFLD). It uses machine learning on NHANES 2017-2023 survey data and two-sample Mendelian randomization analyses with GWAS summary statistics. The 2.7 MB ZIP file contains findings highlighting alcohol and vitamin B6 intake.

Use Cases

• Training predictive models for NAFLD risk based on dietary predictors identified by random forest.
• Conforming causal inference studies in nutritional epidemiology using Mendelian randomization results.
• Investigating non-linear and threshold effects of alcohol and vitamin B6 intake on liver disease risk.
• Benchmarking machine learning feature importance methods against genetic epidemiology findings.

Strengths

• Integrates two methodological approaches: machine learning on survey data and Mendelian randomization.
• Focuses on modifiable dietary factors, offering potential for public health intervention insights.
• Data is sourced from the authoritative NHANES program, covering cycles from 2017 to 2023.

Limitations

• Row count is unknown, which may limit suitability assessment for large-scale modeling.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is small (2.7 MB), which may indicate it contains summary or processed results rather than raw observational records.

Provenance

Source

National Health and Nutrition Examination Survey (NHANES) and genome-wide association study (GWAS) summary statistics.

Collection Method

Combined random forest machine learning on survey data and two-sample Mendelian randomization analysis.

Time Range

2017–2023 (NHANES cycles 2017–2020 and 2021–2023).

Freshness

Last updated 2026-05-04 14:56:52; freshness should be verified.

Geography

United States (NHANES).