Unlocking Obesity's Secrets Through Candidate Genes and Microarrays
Obesity isn't just about diet or exercise—it's written in our DNA. By 2030, an estimated 250 million children and adolescents worldwide will have obesity, driven by complex interactions between genes and environment 7 . While lifestyle factors matter, genetics accounts for 40–70% of body mass index (BMI) variation 1 6 . Early approaches to find obesity's genetic roots focused on "candidate genes"—suspects chosen based on biological clues from rare syndromes or animal models. But how do we test these suspects? Enter microarray technology, a DNA detective tool that scans thousands of genes simultaneously. This article explores how scientists use this approach to crack obesity's genetic code.
40-70% of BMI variation is attributed to genetic factors, highlighting the significant role of DNA in obesity.
Powerful tool that enables simultaneous scanning of thousands of genes to identify obesity-related variants.
Rare syndromes like Prader-Willi (PWS) and Bardet-Biedl (BBS) provided the first genetic clues. PWS, caused by loss of genes on chromosome 15 (e.g., SNURF-SNRPN, MAGEL2), leads to insatiable hunger and obesity. BBS involves mutations in BBS1–BBS20 genes, causing obesity with vision loss and polydactyly 1 5 . These disorders revealed genes critical for appetite regulation and energy balance.
| Syndrome | Prevalence | Genes | Core Features |
|---|---|---|---|
| Prader-Willi | 1:10,000–30,000 | SNURF-SNRPN, MAGEL2 | Neonatal hypotonia, hyperphagia, developmental delay |
| Bardet-Biedl | 1:13,500–160,000 | BBS1–BBS20 | Obesity, retinal degeneration, kidney defects |
| Alström | 1–9:1,000,000 | ALMS1 | Obesity, cardiomyopathy, insulin resistance |
Beyond syndromes, "non-syndromic" obesity includes:
This method targets genes pre-selected based on:
For example, a 2012 study tested 547 candidate genes for BMI associations using GWAS data from 123,564 adults 4 .
Hypothesis: If candidate genes are truly linked to obesity, they should be enriched for BMI-associated variants compared to random genes.
547 candidate genes were compiled from:
Used SNP microarrays to genotype ~2.5 million DNA variants across the genome. Genes were defined as the coding region ±10 kb flanking sequences 4 5 .
Applied three methods to test for enrichment of BMI-associated variants:
| Test | P-value (All Genes) | P-value (Genes with ≥2 Evidence Sources) |
|---|---|---|
| Hypergeometric (25% quantile) | 0.015 | 0.053 |
| Gene-set enrichment (GSEA) | 0.035 | 0.132 |
| Rank tail-strength | 0.042 | 0.011 |
Analysis: This modest enrichment suggests candidate genes retain some validity but explain only a fraction of obesity risk. Most heritability lies in non-candidate genes or complex interactions 4 6 .
| Reagent/Method | Function | Example in Obesity Research |
|---|---|---|
| SNP Microarrays | Genotypes 100,000s of variants genome-wide | Identified FTO as the strongest common obesity locus 1 |
| Chromosomal Microarray (CMA) | Detects copy-number variants (CNVs) | Found pathogenic CNVs in 22% of syndromic obesity cases (e.g., 16p11.2 deletions) 5 |
| Polygenic Risk Scores (PRS) | Combines multiple risk variants into one score | Predicts childhood BMI trajectories as early as age 2.5 2 |
| Gene Expression Arrays | Measures RNA levels across the genome | Linked retinol-binding protein 4 (RBP4) to insulin resistance in fat tissue 3 |
Microarrays paved the way for integrative approaches:
Obesity genetics isn't binary—it spans a spectrum from rare single-gene mutations to complex polygenic risk. The candidate-gene approach, empowered by microarrays, identified key players like MC4R and FTO. But as massive GWAS and polygenic scores reveal, most genetic risk flows through thousands of subtle variants acting in concert. The future lies in merging these approaches: using candidate biology to interpret genome-wide data, and ensuring diverse populations benefit from these discoveries. As one study aptly notes, there's a "continuum between rare monogenic and common polygenic obesity" 8 —and cracking this code will require every tool in the genetic toolkit.