Diversity, genomics and global health
26 August 2019
The 2019 annual report of the UK Chief Medical Officer– Health, our global asset: partnering for progress – highlights the UK’s engagement with global health, through a series of letters from Global Health Leaders.
Of particular interest was the letter from Charles Rotimi (Director of the Centre for Research on Genomics and Global Health, US National Institute of Health) entitled The growing power of genomics. Dr. Rotimi brings a welcome focus on genomics as a technology that can have a positive impact in a global health context. The letter includes a discussion of polygenic risk scores (PRS) and how they will ’allow us to target individuals with the greatest genetic risk for additional monitoring or lifestyle changes well before the appearance of abnormal clinical results‘. Concerns on PRS have been discussed in previous blogs, and an important report on the topic will be released this autumn. There are however issues of equity in genetic data that need to be considered before genomics, and PRS, can be considered a truly global health tool.
Generalisability of genomics
Genome wide association studies (GWAS) are driving the development of PRS. These studies identify genetic variants linked with common diseases such as asthma, cancer or heart disease by using array technology. Arrays can be used to genotype hundreds of thousands to millions of common single nucleotide polymorphisms (SNPs, also known as single nucleotide genetic variants) and are a cheaper alternative to obtaining this information than via sequencing. Polygenic risk scores use the collective predictive power of each of the small-effect SNPs identified using GWAS - which on their own aren’t very useful - and combines them into a single score. The hope is that these scores could be used to predict an individual’s risk of developing disease. However, 78% of individuals in all GWAS are of European ancestry, resulting in a lack of generalisability of findings to people from other populations, and limiting our understanding of how genetics influence disease.
With the recent hype around PRS, attention has been drawn to the fact that not just GWAS but most genomic research studies have been undertaken in European ancestry populations only. This is a long-standing concern for the genomic research community. The homogeneity of the populations in these studies is thought to be one reason why PRS have so far shown to perform less well in predicting disease risk in non-European populations. PRS’s developed in studies that use a diversity of population groups are proving more accurate in predicting disease, and once more of these studies are completed, other underlying reasons for poorer performance will become clearer.
Diversity in genomics
The lack of diversity in genomic data has consequences both for research and clinical applications. To fully understand the uniqueness of (and similarities between) individual human genomes, genomic research needs to be global. For example, the current standard reference genome – a template genome that incorporates all the knowledge we have to date on the human genome – was constructed from a small number of people from a small number of countries. And although it is updated, the human reference genome isn’t complete, in part, due to this lack of diversity, but also because there are regions in the genome that are difficult to sequence. A recent study looking at 910 individuals of African ancestry compared their genomes to the reference genome – and showed that the current reference genome lacked almost 10% of the information contained within the genomes of these individuals. Attempts to fill in these gaps continue.
Global efforts, such as GenomeAsia 100K and H3Africa, are applying genomic science and associated technologies to improve understanding of health and disease in diverse populations. The under-representation of whole genome data and therefore an incomplete understanding of genetic variation globally is hampering large scale genome-wide studies. For example, to better capture the genetic variation across the major ethnolinguistic group’s on the African continent, the H3Africa program developed a 2.3 million SNP array more suited to investigate disease and populations in Africa. These efforts will ultimately inform our understanding of susceptibility to disease on a global scale.
As some direct to consumer testing companies, such as 23andMe, are already offering these services to the public, individuals should be made aware of the current limitations of PRS and how genomic tests apply to them and their ancestry.
Meeting UK population needs
The UK has also been working to address this problem with a concerted drive to include more people from different populations to the UK Biobank and the 100,000 genomes project. The existence of projects such as East London Genes & Health, which focuses on people of South East Asian ancestry is helping. Born in Bradford is another study. However, more such studies that engage and involve underrepresented communities are needed.
Similarly, the government’s new Accelerating Detection of Disease challenge is seeking to enrol under-represented groups ’to enable a better understanding of disease and preventative measures for every individual in society and reduce existing health inequalities‘. The challenge also wants to ’offer as many participants as possible their PRS‘. However, evidence shows the current PRS will have much poorer predictive power in non-European populations and until they are fully evaluated in all populations, it will not be possible to offer an accurate PRS in these under-represented groups.
Creating a global genomic future
Continued support for better representation of all populations in genomic studies will give us a clearer picture of the role of genetics in disease. By better understanding genomics in all populations on a global scale can we pinpoint disease causes quicker and more effectively, improving disease risk prediction for people of all ancestries.