Unveiling the Dual Nature of Gene Discovery: A New Study Challenges Conventional Wisdom
The Unseen Duality of Gene Discovery Methods
The world of gene discovery is not as straightforward as it seems. A groundbreaking study has revealed that the two primary methods used to uncover disease genes, genome-wide association studies (GWAS) and burden tests, each reveal a distinct and often contrasting picture of biology. This finding has significant implications for drug development and our understanding of disease risk.
The human genome, a complex tapestry of genes and regulatory DNA, is the focus of this research. It contains the instructions for making proteins and the regulatory elements that control gene expression. The study takes a genome-wide approach to understanding how small DNA variations, known as variants, can influence traits such as height, hair color, and disease susceptibility.
Led by researchers from NYU Langone Health, Stanford University, UC San Francisco, and the University of Tokyo, the study analyzed the results of GWAS and burden tests for 209 traits from the UK Biobank, a vast genetic database containing information from hundreds of thousands of individuals. The findings were striking: burden tests primarily identify genes that significantly impact the specific disease being studied, with minimal effect on other traits. In contrast, GWAS can pinpoint both disease-specific genes and genes that influence a wide range of diseases and biological processes.
The Controversy of Gene Discovery Methods
But here's where it gets controversial. The study authors propose that the difference in results between the two methods is due to the varying impact of genes on different traits and biological processes. Some genes primarily influence a single trait, while others affect multiple traits simultaneously. The variants that severely disrupt these 'multi-trait' genes have broad consequences and are often removed by evolution, making them harder to detect in burden tests. GWAS, on the other hand, can still identify these genes because regulatory DNA variants may have more limited effects, allowing them to escape evolutionary pressures.
The study authors suggest that two gene features are crucial for ideal gene prioritization in disease risk and trait analysis. The first is 'importance' - how much a gene affects disease when disrupted. The second is 'specificity' - whether a gene primarily influences one disease or many traits. Understanding both features would enable researchers to identify the best therapeutic targets and anticipate potential side effects.
The Limitations of p-Values
A related finding involved the p-value, a standard measure of statistical significance. The study shows that p-values from GWAS and burden tests are not reliable indicators of a gene's importance. This is significant because identifying important genes can reveal the central biological processes involved in disease. The authors suggest that new methods are needed to accurately infer this key biological gene feature.
The Future of Gene Discovery
Looking ahead, the research team has begun developing methods to prioritize genes based on their importance. While GWAS and burden tests alone may not have sufficient power to accurately estimate gene impact, combining these results with experimental data describing gene function inside cells can improve estimates. Machine learning methods can then identify shared patterns across genes, leading to a more comprehensive understanding of disease biology and streamlined drug development.
'This would be revolutionary,' said co-senior author Jeffrey Spence, 'because it would allow us to leverage all the cell-level experimental data to learn about human-level traits, identify the most important disease genes, and streamline drug development.'
The study's findings challenge conventional wisdom and highlight the need for a more nuanced understanding of gene discovery methods. As the field of genomics continues to evolve, it is essential to consider the dual nature of gene discovery and its implications for personalized medicine and disease prevention.
