Unveiling the Dual Nature of Biology's Genetic Secrets
The quest to understand the human genome has led to an intriguing revelation: two distinct methods, each with its own unique perspective, are shedding light on the complex interplay between genes and disease.
A recent study, published in Nature, has delved into the contrasting approaches of genome-wide association studies (GWAS) and burden tests, revealing their distinct contributions to our understanding of biology.
The human genome, a vast landscape of genetic instructions, contains not only genes but also regulatory DNA, which acts as a master controller, dictating when and how genes are activated. This new research takes a panoramic view, exploring how small variations in DNA, known as variants, can influence a myriad of traits, from physical attributes like height and hair color to the risk of developing diseases.
Led by researchers from NYU Langone Health, Stanford University, UC San Francisco, and the University of Tokyo, the study analyzed these two primary methods used to decipher the impact of genetic differences on disease biology. GWAS, a well-established technique, scans the entire genome for common variants linked to diseases, while burden tests focus on rare variants that have a more significant impact on protein function.
By examining the results of both methods for 209 traits from the UK Biobank, a vast genetic database, the researchers made a fascinating discovery. Burden tests primarily identify genes that have a specific and focused impact on the disease being studied, with minimal influence on other traits. In contrast, GWAS can identify these disease-specific genes but also reveal genes that have a broader impact, influencing multiple diseases and biological processes.
"Our study provides clarity on why these methods produce different results and highlights their unique biological significance," explained co-senior author Hakhamanesh Mostafavi, PhD, from NYU Grossman School of Medicine. "Understanding these differences is crucial for translating genetic findings into practical applications, such as drug development."
The need for a new approach becomes evident when considering the limitations of GWAS. While it has been a valuable tool for identifying disease-associated genes, the results can be overwhelming, often implicating hundreds of genes per disease, making it challenging to pinpoint the most relevant ones.
However, the emergence of massive biobanks has empowered burden tests, allowing them to paint a different picture. These tests reveal a more concise and interpretable set of genes linked to the same diseases, raising questions about which method provides a more accurate reflection of disease biology.
The researchers uncovered a key reason for the divergent results: genes vary in their impact on different traits and biological processes. Some genes primarily influence a single trait, while others have a more widespread effect, impacting multiple traits simultaneously. Variants that severely disrupt these 'multi-trait' genes have far-reaching consequences and are often weeded out by evolution due to their detrimental effects on survival and reproduction.
This evolutionary process makes these variants less common, which, in turn, makes them harder for burden tests to detect. In contrast, GWAS can still identify these genes because regulatory DNA variants often have a more limited impact on gene activity, allowing them to evade evolutionary removal.
The study authors propose that two critical gene features are essential for prioritizing genes with respect to disease risk or traits. The first is 'importance' - how significantly a gene's disruption affects disease. The second is 'specificity' - whether a gene primarily affects one disease or has a broader impact on multiple traits. Understanding both features is crucial for identifying the most promising therapeutic targets and anticipating potential side effects.
Another intriguing finding relates to the p-value, a standard measure of the reliability of study results. Surprisingly, the study reveals 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 provide insights into the core biological processes involved in disease development.
"Our results do not diminish the value of GWAS and burden tests," clarified Mostafavi. "Rather, they highlight the need for a new interpretation of these methods. We believe that developing novel approaches to infer gene importance is essential."
Looking ahead, the research team has embarked on developing methods to prioritize genes based on their importance. While GWAS and burden tests alone lack the power to accurately estimate a gene's impact on disease, combining these results with the rapidly expanding experimental data on gene function within cells can enhance our understanding. Machine learning methods can identify shared patterns across genes and improve our estimates, leveraging the wealth of cell-level experimental data to gain insights into human-level traits.
"This could revolutionize our approach to drug development," said co-senior author Jeffrey Spence, PhD, from the University of California, San Francisco. "By integrating all this experimental data, we can identify the most critical disease genes and streamline the drug discovery process."
This study opens up exciting avenues for further exploration, inviting discussion and debate among scientists and researchers. As we continue to unravel the mysteries of the human genome, these contrasting methods offer a fascinating glimpse into the dual nature of biology, each providing unique insights into the complex relationship between genes and disease.
