Stefania Benonisdottir, a DPhil candidate at the University of Oxford's Big Data Institute, and Professor Augustine Kong from the Leverhulme Centre for Demographic Science, have made significant strides in understanding participation bias in genetic studies. Their findings, published today in Nature Genetics, shed light on the impact of ascertainment bias on genetic studies and the role of footprints in genetic data to overcome it.
Read the full news story here.
Ascertainment bias occurs when collected data does not accurately represent the intended study population, leading to misleading results. For genetic studies, this bias is particularly challenging to identify and address due to limited information on non-participants.
Stefania Benonisdottir said, ‘Currently, most genetic studies are based on genetic databases which contain large numbers of participants and a wealth of information. However, some people are more likely to be included in these databases than others, which can create a problem called ascertainment bias, where the genetic data collected is not representative of the intended study population.’
In the video below, Stefania explains a new method developed with Augustine, that leverages genetic data to explore the genetic component of participation.
Their research, conducted using data from the UK Biobank, revealed a correlation between the genetic component of participation and variables like BMI and educational attainment. These findings validate the method's ability to capture genetic associations with participation and highlight the complex nature of participation in genetic studies.
Professor Melinda Mills, Director of the Leverhulme Centre for Demographic Science adds, ‘As our GWAS Diversity Monitor shows, the road to improve diversity in genome-wide association studies is long. However, this statistical framework is a huge step in the right direction to mitigate the risk of incomplete or inaccurate data analysis and ensure that genetic research truly benefits everyone.’
The study demonstrates that participation bias leaves detectable genetic footprints in datasets. By incorporating participation behaviour into data analysis and study designs, future research can improve accuracy and reduce bias.
Professor Augustine Kong concludes, ‘Ascertainment bias poses a statistical challenge in genetics research, particularly in the era of big data. Our study identifies detectable footprints of participation bias in the genetic data of participants, which can be exploited statistically to enhance the investigation of the relationships between genes and various human traits.’
