Nature Study Finds Rare Variants Lift Extreme-Trait Risk Prediction 71.7%
Updated
Updated · BIOENGINEER.ORG · May 27
Nature Study Finds Rare Variants Lift Extreme-Trait Risk Prediction 71.7%
7 articles · Updated · BIOENGINEER.ORG · May 27
A Nature study tied rare genetic variants directly to the extreme tails of human traits, showing that adding them to polygenic risk models sharply improved prediction for outlier individuals.
UK Biobank exome and genome data showed rare coding hits clustered in trait tails, and adding rare-variant scores cut some tail-specific POPout signals by as much as 90%.
The gain was concentrated in extremes: rare-plus-common models raised tail-related odds ratios about 71.7% even though average variance explained increased only around 11.6%.
Across 74 traits, signals implicated disease-linked genes including LDLR, MC4R, JAK2 and CHEK2, reinforcing the clinical value of sequencing people at phenotypic extremes.
The findings suggest standard GWAS miss much of the biology in the deepest 0.1% tails, leaving larger sequencing cohorts to uncover ultra-rare and structural variants still unexplained.
Why do our most extreme health traits defy the standard rules of genetics?
How could a forgotten class of rare genes revolutionize personalized medicine?
Rare Genetic Variants Explain 92% of Extreme Human Trait Cases: A Paradigm Shift in Precision Medicine and Genetic Risk Prediction
Overview
A groundbreaking study published in Nature on May 27, 2026, by researchers at the Icahn School of Medicine at Mount Sinai has revealed a major shift in our understanding of human genetics. The team found that people at the extreme ends of traits like cholesterol or height often have a clear genetic cause, challenging the traditional polygenic view that many small-effect common variants shape complex traits. Instead, the study shows that rare genetic variants with large effects are more common at these extremes, offering new insights into disease risk and paving the way for more precise, personalized medicine.