Published today in Nature, a team of researchers from the UK, USA, Norway, Australia, and the Netherlands is calling for a shift towards family-based sampling to advance genetic research.
Biobanks have been instrumental in advancing our understanding of population health and disease by collecting and managing vast amounts of genetic and other biological data, traditionally using “population-based” samples that treat individuals as independent units. However, as biobanks grow to include millions of participants, the benefits of this approach are reaching a point of diminishing returns. A recent Nature Perspective article advocates for a shift toward collecting data not just from individual participants but also from their family members. Such “family-based” data could significantly enhance data quality, leading to more reliable genetic findings and providing insights beyond those possible using traditional approaches. This shift could also help answer key questions in social sciences, population genetics, and biology.
Professor Neil Davies, Professor of Medical Statistics at University College London, commented: “Family-based biobanks can provide some of the most compelling evidence about the causes of physical and mental health. Collecting data from entire families rather than individuals alone can help us better understand the links between genetics, environment, and disease, which could lead to more effective interventions and treatments.”
Although family-based sampling may increase analytical complexity, the article argues that the benefits outweigh these challenges.
Firstly, family-based data--for example, samples of siblings or of parents and their children--can provide more reliable insights than traditional population-based samples. It can help overcome confounding in genetic research by exploiting the natural experiment of the randomised transmission of genetic material within families, thereby clarifying true genetic effects. For instance, estimates of genetic effects from family-based Genome-Wide Association Studies (GWASs) are significantly smaller than those derived from population-based samples for several traits, including depression and educational attainment.
Secondly, a family-based sampling strategy enables the collection and analysis of biological and phenotypic information spanning multiple generations, which can allow investigators to understand the impacts of families and how relatives influence one another.
Thirdly, family-based samples can improve data quality and can shed new light on fundamental questions across various fields that are difficult to explore using population-based samples. These topics range from the effects of sociodemographic factors on health and well-being to biological issues such as the frequency of genetic mutations and the rate of recombination.
Professor Matthew Keller, Director of the Institute for Behavioral Genetics and the University of Colorado at Boulder, said: “This paper emphasises the current lack of sufficient family data and calls for a change. We argue that the costs of collecting such data are minimal compared to the substantial benefits it could provide. Incorporating family data has the potential to transform our understanding of the causes and effects of mental health conditions, addressing long-standing limitations across disciplines such as clinical medicine and the social sciences.”
The article also addresses the practical challenges of implementing family-based biobanks. These include strategies like recruiting entire households or linking existing biobank data with administrative records. While such approaches may require more resources, the long-term scientific rewards are expected to be substantial.
Professor Gibran Hemani, Professor in Statistical Genetics in the Integrative Epidemiology Unit of Bristol Medical School, said: "Population-based analysis will continue to be a very important strategy but it is designed to maximise the efficiency of these studies. However, what a lot of recent research has shown is that we may be at an interesting inflection point in genetic research, where the efficiency problem has received a lot of attention, but this is sometimes at the cost of reliability. The Bristol-based study 'Children of the 90s' is a family-based study which is renowned for the broad scope of research questions it is able to address reliably. Learning from this and many other such studies - we can combine the two ingredients of very large samples with a family-based design, which is an exciting recipe for getting the best of both worlds."
The full article, The importance of family-based sampling for biobanks, is available in Nature: https://www.nature.com/articles/s41586-024-07721-5 (doi:10.1038/s41586-024-07721-5).
This paper was published by researchers at UCL, University of Bristol, Pennsylvania State University, Wellcome Sanger Institute, University of Oxford, University of Queensland, UCLA, and the University of Colorado.