Genome-wide analyses point to differences in genetic architecture of BMI between tall and short people

Bochao Lin, Benjamin H. Mullin, Scott G. Wilson, John P. Walsh, Yue Li, Roger A.h. Adan, Jurjen J. LuykxbioRxiv 2020 · DOI

Abstract: To examine differences in the genetic architecture of BMI between tall and short people, we conducted genome-wide and follow-up analyses using UK Biobank data. We identify 57 loci as height-specific, detect differences in SNP-based heritability between tall and short people and show how genetic correlations between the two rises during the lifespan. Using phenome-wide analyses (PHEWAS), a significant association between a short people-specific locus on MC4R and energy portion size was detected. We identify one locus (GPC5-GPC6) with different effect directions on BMI in short and tall people. PHEWAS indicates this locus is associated with bone mineral density. Transcriptome-wide analyses hint that genes differentially associated with BMI in short vs tall people are enriched in brain tissue. Our findings highlight the role of height in the genetic underpinnings of BMI, provide biological insight into mechanisms underlying height-dependent differences in BMI and show that in short and tall people obesity is a risk factor that differentially increases susceptibility for disease.

Introduction

Body Mass Index (BMI), calculated as $weight/height^2$, is a standard clinical metric for obesity. However, it is a relatively crude ratio that may mask significant biological variation. Shorter adults typically have higher BMIs and different metabolic requirements compared to taller individuals. Specifically, shorter people require less energy for basic metabolism; if they consume the same food portion sizes as taller individuals, they face a higher risk of obesity.

Relationship between height and energy intake: Total energy intake (KJ) increases with height (left), while portion energy intake adjusted for body weight (KJ/Kg) decreases as height increases (right).

This study investigates whether these phenotypic differences are reflected in the underlying genetic architecture of BMI. By identifying 57 genetic loci that influence BMI differently in tall versus short people, the research suggests that the biological drivers of obesity are height-dependent, linking these differences to bone mineral density and brain-regulated appetite.

Methods and Data

The researchers utilized the UK Biobank (UKBB) dataset, encompassing 502,538 participants of European ancestry. They employed a multi-stage genomic pipeline:

  • Stratified GWAS: The population was split into "short" (<33rd percentile) and "tall" (>67th percentile) groups for separate association analyses.
  • mtCOJO (Multi-trait-based Conditional and Joint Analysis): This core innovation allowed the team to condition results of the short group on the tall group (and vice versa) to isolate variants with height-specific effects.
  • PHEWAS & TWAS: Phenome-wide and Transcriptome-wide association studies linked height-specific variants to other physical traits and tissue-specific gene expression (using GTEx and Linnarsson Lab datasets).
  • Mendelian Randomization (MR): Used to determine causal links between height-specific BMI variants and various diseases.

Key Insights: The Biological Drivers of BMI

The "Gravitostat" Hypothesis

A particularly striking finding was a locus between GPC5 and GPC6 that exhibited opposite effect directions on BMI in tall vs. short people. This locus is associated with bone mineral density, supporting the "gravitostat" theory—a biological weight-sensing mechanism in the bones that helps regulate body mass.

Behavioral vs. Metabolic Drivers

A short-specific locus on the MC4R gene, a well-known appetite regulator, was significantly associated with "energy portion size." This suggests that for shorter individuals, genetic obesity risk may be more closely tied to the brain's behavioral response to food and portion control.

Age-Dependent Convergence

The study found that the genetic correlation between tall and short people’s BMI is lower in youth and increases with age. This suggests that while initial genetic predispositions differ by height, cumulative environmental factors like diet and lifestyle eventually dominate the BMI profile in older age.

Results and Disease Susceptibility

The study identified 57 regions of the genome where the impact on BMI was significantly different between height groups. Tissue enrichment analysis showed that genes associated with BMI in short people were enriched in 12 brain regions (including the hippocampus and amygdala), compared to only 4 regions for tall-specific genes.

Furthermore, Mendelian Randomization revealed distinct disease risks:

  • Tall-specific BMI alleles: Increased risk of Major Depressive Disorder; decreased risk of Chronic Kidney Disease.
  • Short-specific BMI alleles: Increased risks of Bipolar Disorder and hip osteoarthritis.

Conclusion and Future Directions

While this study provides deep insights into the height-dependent genetic architecture of BMI, it is limited by its focus on individuals of European ancestry. Future work is needed to validate these molecular pathways—particularly the bone-BMI link—in animal models and to explore these interactions across diverse global populations. Understanding how these variants interact with environmental exposures over a lifetime remains a key area for future research.

Summary updated

Abstract

Abstract To examine differences in the genetic architecture of BMI between tall and short people, we conducted genome-wide and follow-up analyses using UK Biobank data. We identify 57 loci as height-specific, detect differences in SNP-based heritability between tall and short people and show how genetic correlations between the two rises during the lifespan. Using phenome-wide analyses (PHEWAS), a significant association between a short people-specific locus on MC4R and energy portion size was detected. We identify one locus ( GPC5 - GPC6) with different effect directions on BMI in short and tall people. PHEWAS indicates this locus is associated with bone mineral density. Transcriptome-wide analyses hint that genes differentially associated with BMI in short vs tall people are enriched in brain tissue. Our findings highlight the role of height in the genetic underpinnings of BMI, provide biological insight into mechanisms underlying height-dependent differences in BMI and show that in short and tall people obesity is a risk factor that differentially increases susceptibility for disease.
BibTeX
@misc{lin2020genome,
  author = {Bochao Lin and Benjamin H. Mullin and Scott G. Wilson and John P. Walsh and Yue Li and Roger A.H. Adan and Jurjen J. Luykx},
  title = {{Genome-wide analyses point to differences in genetic architecture of BMI between tall and short people}},
  year = {2020},
  doi = {10.1101/2020.09.24.312181},
  howpublished = {bioRxiv},
}
Figure