A fine way to find disease associations in Africa

 

African populations are so genetically diverse, both within and between ethnic groups, that finding genetic links to disease has been a problem. But a new study of children with severe malaria in The Gambia by the Malaria Genomic Epidemiology Network (MalariaGEN) shows that combining genotyping (searching for known sequence variants) with limited sequencing could pinpoint the variants that play a causal role in disease.July 09 Kenya

The members of the international collaboration, led by Professor Dominic Kwiatkowski of the WTCHG, conducted a genome-wide association study (GWA) of almost 1000 young children admitted to hospital with severe malaria, and a control group of healthy newborn babies from the same area. The aim was to discover which of around half a million known variations in the genome sequence (known as single nucleotide polymorphisms, or snps) were associated with protection against the disease.

The strongest signal of such an association came from a region of the genome known to carry the mutation that causes sickle cell disease in people who inherit it from both parents. Scientists have known for more than 50 years that children with one copy of the sickle cell variant, known as HbS, are ten times less likely to suffer from severe malaria.

Overall the study showed a high level of false positive signals, because of the underlying genetic diversity: children in the study came from at least eight of the ethnic groups that make up the Gambian population. Controlling for ethnic identity helped to reduce this risk; overall, the GWA study showed up more than 100 regions that were potentially contributing to disease susceptibility at a small but significant level.

However, even the signal from the HbS region did not seem strong enough to predict its known high impact on disease risk. To explore this region further, the researchers sequenced 111,000 bases from 62 Gambian controls, and found many more snps. Using the pattern of diversity in the HbS region in Gambians revealed by this fine-resolution analysis, they were able to 'impute' the presence or absence of variants in the whole study population statistically, and test them for association with severe disease. The analysis greatly enhanced the signal from the HbS mutation, and accurately pinpointed its location.

In future the same technique may be able to locate previously unknown variants that are having a causal effect in disease susceptibility, something that is much harder to do in European populations where many variants tend to be inherited together. 'In Africa, once you've narrowed in on the gene, you're much more likely to be able to locate the causal variant, because there is less correlation with other variants', says Kwiatkowski.

M. Jallow et al. (2009) Genome-wide and fine-resolution association analysis of malaria in West Africa. Nature Genetics 41, 657 - 665

For more information on Prof Kwiatkowski 's research, click here.