Related Experiment Video
Updated: May 7, 2026
08:58
Optimized Analysis of In Vivo and In Vitro Hepatic Steatosis
Published on: March 11, 2017
15.8K
Trans-ancestral rare variant association study with machine learning-based phenotyping for metabolic
Robert Chen1,2,3, Ben Omega Petrazzini1,2,4, Áine Duffy1,2,4
1The Charles Bronfman Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA.
Genome Biology
|March 11, 2025
Summary
This study identified rare genetic variants linked to metabolic dysfunction-associated steatotic liver disease (MASLD) using large, diverse biobanks. Machine learning improved the detection of genetic associations for this challenging-to-phenotype liver condition.
Area of Science:
- Genetics
- Genomics
- Metabolic Diseases
Background:
- Genome-wide association studies (GWAS) have identified common variants for metabolic dysfunction-associated steatotic liver disease (MASLD).
- Studies on rare coding variants are limited by phenotyping challenges and small sample sizes.
- This research investigates rare and ultra-rare coding variants in relation to MASLD and liver fat content.
Purpose of the Study:
- To identify rare and ultra-rare coding variants associated with MASLD and liver fat.
- To leverage trans-ancestral meta-analysis and machine learning for enhanced genetic discovery.
- To increase statistical power in the genetic investigation of MASLD.
Main Methods:
- Trans-ancestral meta-analysis of rare/ultra-rare coding variants with proton density fat fraction (PDFF) and MASLD case-control status in over 736,000 participants.
- Development of machine learning models to predict PDFF and MASLD status.
- Association testing using predicted phenotypes to boost statistical power.
Main Results:
- Identified 27 genes associated with MASLD, including known monogenic causes and previously implicated genes.
- Discovered ancestry-specific variants, particularly in African ancestry populations.
- Found significant heterogeneity in ancestry and sex-stratified analyses, highlighting complex genetic architecture.
Conclusions:
- Trans-ancestral association analyses are effective for discovering ancestry-specific rare variants in MASLD.
- Machine learning enhances genetic studies of difficult-to-phenotype diseases in large, diverse biobanks.
- The findings contribute to understanding the genetic basis of MASLD and identify potential therapeutic targets.
