Background

Over the past 20 years, Genome-Wide Association Studies (GWAS) have identified thousands of genetic variants involved in various complex traits, such as kidney function and lipid levels, establishing a standard methodology for identifying correlations between specific genetic variants and phenotypes in large populations. However, the situation is starkly different in the field of psychiatry.

Central nervous system (CNS)-related diseases, such as schizophrenia, bipolar disorder, and major depressive disorder, exhibit very high heritability, yet the influence of individual variants is uniquely small. Even when the same number of significant variants are identified, the effect size of brain disease variants is much smaller than that of physical traits. Cases where statistical significance barely reached the genome-wide significance level (p-value of 5×10⁻⁸) were frequent.

The academic community primarily attributed this phenomenon to clinical measurement errors. The hypothesis was that diagnostic heterogeneity, due to reliance on interviews rather than physical indicators, blurred the signal. Even after increasing study samples to hundreds of thousands, this gap did not narrow. Instead, only the structural differences between physical traits and brain diseases became prominent. The argument that this is not merely measurement noise, but the result of the evolutionary trajectory of the human brain imprinted on the genome, began to gain strength.

Key Findings

A joint research team led by Professor Jonathan Pritchard of Stanford University and Professor Guy Sella of Columbia University solved this puzzle using an evolutionary genetics model. Utilizing data from the UK Biobank and the Psychiatric Genomics Consortium (PGC), the researchers compared genomic structures after matching the effective sample sizes of CNS-related traits with non-neural complex traits.

The analysis results overturned existing conventional wisdom. The weak variant effects observed in brain diseases were not due to measurement noise, but were the result of powerful evolutionary selective constraint and a massive mutational target size acting on the genome.

The human brain is a sophisticated organ that expresses more than half of all genes. Considering the regulatory sequences involved in synapse formation and neural network development, the mutational target range affecting brain function is dispersed across the entire genome. Because the target size is so vast, influence is not concentrated on specific genes but is fragmented into numerous variants.

Here, natural selection, which is directly linked to survival and reproduction, intervened. High-impact variants that disrupt brain function critically reduce an individual's fitness. During evolution, purifying selection acted to thoroughly eliminate such high-impact variants from the population.

The variants that survive and are common in the population are those with effects so weak as not to threaten survival. Population genetics model fitting confirmed that CNS-related traits have a wider mutational target size than physical traits, along with a significantly higher intensity of negative selection on variants.

Implications and Outlook

This study demonstrates that the genomic structure of complex diseases is determined by the biological importance of the tissue in which the trait is expressed and by evolutionary constraints. For organs critical to survival and cognitive function, such as the brain, strong selective pressure is applied, reducing the effect of common variants and supporting the principle of the omnigenic model, where susceptibility is scattered across the genome.

A shift in research design is now urgent. While previous GWAS focused on quantitative expansion by gathering millions of people to find minute common variants, the focus must now shift to the discovery of rare variants and de novo mutations. To find clues to pathogenesis, it is necessary to track high-risk variants that have escaped purifying selection; although rare, these variants exert a substantial impact on the disease.

To achieve this, the rapid adoption of large-scale Whole Genome Sequencing (WGS) and Whole Exome Sequencing (WES) is required. Ensuring racial diversity in analysis subjects also remains a task. As current data is biased toward European populations, follow-up studies are needed to confirm whether the same evolutionary model holds true across diverse populations.