History’s genetic legacy
Three Oxford professors are investigating how genetic structure varies geographically in the UK.
If you are from northern parts of the British Isles, particularly Scotland and the Orkneys, there may well be traces of Viking ancestry in your DNA. Or if you are from central England or East Anglia, you could be a distant descendant of the Saxons and Angles. The invasions that, for thousands of years, swept across these islands not only changed the political and cultural landscape but also left their imprint in the genes of modern Britons (see Box).
At the University of Oxford, a new £2.3 million project, funded by the Wellcome Trust and led by cancer and population geneticist Professor Sir Walter Bodmer, aims to decipher the genetic structure of the UK. But the historical insights are a secondary concern; rather, the focus is on creating a tool for researchers tracking down genes associated with common human diseases.
Finding that a particular variant of a gene is more common in people with, say, diabetes than in those without is a clue that the gene may be increasing susceptibility to the disease. But is the association real, or are the different genetic make ups of the people involved in the study clouding the results?
To tackle the problem, Professor Bodmer has joined forces with two other Oxford researchers – Professor Peter Donnelly, an expert on population genetics and statistics, and gene-hunter and Wellcome Trust Principal Research Fellow Professor Lon Cardon.
Their joint goal is a freely available set of genetic tools to aid researchers searching for genes linked to disease.
A classic method of tracking down genes associated with disease is the ‘case–control’ study. The idea is to examine whether a particular piece of DNA, a marker, is more frequent in a group of people with a disease than a similar group without the disease.
The first successes of this approach came in the late 1960s and early 1970s and involved the HLA genes, which are important in immune responses. The HLA variant B27, for example, is found in at least 90 per cent of people with ankylosing spondylitis (a form of arthritis primarily affecting the spine). “The associations between HLA variants and certain diseases are striking,” says Professor Bodmer. “Even if there is some genetic mixture in the population, it doesn’t mask the association because the real genetic effects are so large.”
Applying the same approach to complex diseases such as heart disease has been less straightforward. “Association analysis for small effects is problematic,” says Professor Bodmer. “There may be several different genes involved, and each gene may have one or more variants that differ in frequency. Some variants may be more common and have very small effects, others may be rare but have slightly more significant effects.”
Professor Cardon agrees, and adds that many studies have simply been too small. “People recognized that many genes are involved in complex diseases, but they expected a few to have large effects. The studies were small because they were looking for ‘low hanging fruit’; with a few exceptions, that has been a failed assumption.”
Time and again, an association between a gene and a disease was reported, but when other researchers looked for the same effect, they could not reproduce the result. “About 10 years ago, geneticists got very nervous about these kinds of studies,” says Professor Cardon. “They were afraid that differences in population structure were leading to the irreproducible results – and there are examples where population structure is known to have caused problems. But it was not only lack of attention to broad ethnic differences, but small samples that led to the confusion. We now know that if we are to find the genes that have small effects on common diseases, such as heart disease, diabetes, and osteoporosis, we need to study large numbers of people and we need to know, at a fine level, the genetic structure of that sample of people.”
Searching for structure
The last 50–100 years have seen remarkable flux in the UK population. The Oxford team is therefore looking for volunteers who come from rural areas, and whose parents and grandparents were from the same area.
The goal is to gather DNA from 3500 people, from 30 locations in the UK. “We have collaborators throughout the country: Scotland, Newcastle and Cumbria, Leeds, London, Wales, Northern Ireland and so on,” says Professor Bodmer. “They are absolutely key: they make it possible for the project to happen, as they know their own area.”
To understand the range of differences between different parts of the population, each participant’s DNA will be tested for 2000 single nucleotide polymorphisms (SNPs). The data and the DNA samples will become a national resource for other researchers to use. Researchers will be able to use the samples and the data on their genetic marker distribution to match their disease samples to appropriate controls and so reduce the chance of finding spurious associations.
With a number of very large population studies, such as the UK Biobank, getting underway, a knowledge of the UK’s genetic structure will be particularly timely. Although the team expect there to be extensive interest in the historical insights that the project is likely to bring, they are firmly focused on the scientific benefits: “If we are to use population studies effectively to find genes involved in disease,” says Professor Cardon, “this is information we just have to have.”
Image courtesy of the Jorvik Viking Centre, York