The analysis of genetic variants in different populations often reveals that the same locus (position in the genome) may have different or even absent effects depending on the ancestry of the individuals. This phenomenon does not imply “error” in the original studies, but rather reflects variations in the genetic structure, evolutionary history and environment of each group. In this article, we discuss the main reasons why an allele proven to be associated with a trait in Europeans may not have the same impact in Africans, illustrating with two classic examples and pointing out methodological solutions.
1. Mechanisms of allelic heterogeneity
To understand why an allele A may increase the risk of disease or modulate a trait in Europeans but not in Africans, we need to consider three key concepts:
1. Allele frequency
• Refers to the proportion in which each variant (allele) appears in the population. If the “A” allele is common in Europe (e.g., 30% frequency) and rare in Africa (<1%), its statistical impact on the African continent will be difficult to detect, even if biologically real.
2. Linkage disequilibrium (LD)
• This is the correlation between nearby variants on the chromosome. In African populations, LD tends to be shorter — that is, there is less “drag” of a marker on the causal one. Thus, a SNP (single nucleotide polymorphism) that signals an effect well in Europeans may not capture the same association in Africans because it loses contact with the functional variant.
3. Local ancestry and gene–environment interactions
• In admitted individuals (ancestry mixture), genomic segments of African or European origin can modulate the effect of the same allele. In addition, environmental factors — diet, exposure to pathogens, lifestyle — interact with genetic variants differently in each population.
Technical terms
• SNP (Single Nucleotide Polymorphism): variation of a single base pair (e.g., A vs. C) at a DNA position.
• Alleles: the two (or rarely more) forms that a SNP can assume.
• Heterogeneity: variation in the effect of the same marker between populations.
2. Illustrative examples
2.1 APOE ε4 and Alzheimer’s
• In Europeans: the ε4 allele of the APOE gene substantially increases the risk of late-onset Alzheimer’s (odds ratio ~ 3–4 for heterozygotes) and tends to explain much of the genetic burden of the disease.
• In native Africans: several studies with cohorts from Nigeria and Ghana have not replicated this strong effect; ε4 carriers have much lower or inconsistent risks of dementia.
• Why? In African populations, the haplotype (set of variants) around APOE differs structurally from that in Europeans, and environmental factors (diet, infectious comorbidities) modulate gene expression differently.
2.2 LCT – lactase persistence
• In Europeans: the SNP rs4988235 (–13 910 C→T) is strongly associated with the maintenance of lactase enzyme activity in adulthood, an adaptive trait that has been positively selected over the last 7,500 years.
• In African populations: this same locus is usually monomorphic (only the C allele) and, even so, many African groups consume milk throughout their lives thanks to other mutations (rs145946881, rs41525747, etc.). These are cases of convergent evolution, in which different mutations produce the same phenotype.
3. Implications for research and clinical application
1. Trans-ethnic PGS:
• Polygenic Score (PGS) combines hundreds to thousands of SNPs to predict risks or traits (such as IQ). Developing this score in Europeans and applying it “blindly” to Africans drastically reduces its accuracy.
• Recommended strategy: use methods such as PRS-CSx, which integrate GWAS from multiple ancestries, or locally calibrate the score in African cohorts.
2. Percentile recalibration:
• Instead of comparing values to absolute extremes (minimum/maximum), define quantile limits (25th, 50th, 75th percentiles) in the target population. This minimizes distortions caused by outliers and differences in amplitude.
3. Genetic intervention and counseling:
• It is essential to communicate that a PGS or an association observed in one group does not automatically translate to another ancestry. Counseling should include information on potential biases and uncertainties.
4. Conclusion
The variation in the effect of the same genetic marker across populations does not indicate a contradiction, but rather the complexity of human inheritance. Differences in allele frequency, LD, and environment shape the strength and direction of associations. To ensure reliable predictions in diverse populations—especially in admitted individuals—it is essential to:
