Lorne special: Tracking recent human evolution

This feature appeared in the January/February 2010 issue of Australian Life Scientist. To subscribe to the magazine, go here.

Pardis Sabeti is something of a rock star in the world of genetics, both literally and figuratively. Not only has she developed a powerful statistical method for detecting regions in the human genome undergoing natural selection, she is also singer and bass player in a rock band. But when she speaks on the Comparative & Evolutionary Genomics panel at the Lorne Genome conference in February, it’ll be strictly science.

Furthering her work in scouring the human genome for signs of evolution, Sabeti, who heads up the eponymous Sabeti Lab at Harvard University, will be speaking about a new test she and her team have developed that will be able to pinpoint specific regions in the genome that are undergoing natural selection.

The test could help teach us a great deal about how humans have responded to a wide range of environmental pressures, particularly infectious diseases, over our recent evolutionary history.

The big breakthrough is the accuracy her new test affords. “There have been numerous tests developed in the last few years to detect evidence of recent local adaptation in humans,” she says. “They’ve identified hundreds of candidate regions where evolution is acting, but these regions are often very large – one megabase long. Within those regions it’s hard to identify what specific change is the driver.”

Sabeti’s new test, called composite of multiple signals, or CMS, is actually three tests in one. “We picked a bunch different tests that have already been used and created a composite framework, a way of merging them together. We then showed that when you apply the composite framework you very dramatically increase your power to assess region.”

The first step is to look at the prevalence of a particular mutation in a population. “When a new mutation rises in prevalence in a population, it leaves behind distinctive footprints in your DNA,” says Sabeti. “If the selection happens in one population and not another – because one population has a different environmental pressure – then it’ll drive significant differences in the prevalence of the mutation in one population rather than another.”

The second element of the test is to look for a high prevalence of non-ancestral alleles within a population. Ancestral alleles are those that are believed to have belonged to our last common ancestor, so non-ancestral alleles are new mutations that have occurred since the time of that ancestor.

---PB---

The final stage is to look for long haplotypes, a technique Sabeti has used in the past to scan for genes that have spread through human populations over the past few tens of thousands of years. “Essentially, it’s another way of measuring very young and common variants.”

Sabeti’s team hypothesised that a mutation that has been selected for should register on all three tests, allowing the researchers to ‘triangulate’ in on a very specific region on the genome to see where the mutation is located. So, while each individual test is relatively coarse – spanning regions about a megabase long – when combined they can narrow things down to around 50 kilobases or less.

“More dramatically, within those regions, we’re able to pinpoint a handful of mutations that are the likely causal ones rather than have the entire one megabase region with thousands of variants being implicated.”

Not all the mutations pinpointed by the test are gene coding regions, however. According to Sabeti, around half the regions she’s elucidated so far contain no genes at all, indicating the mutations are in transcriptional or regulatory functions.

Sabeti hopes CMS will enable researchers to investigate mutations that have lent resistance to certain infectious diseases, which she suggests is a common driver of natural selection. “What’s really fascinating is that a lot of the candidate regions lie within immune genes, or lie within genes known to be involved in infectious diseases.”

One example of a disease being investigated is Lassa fever. “Arguably one of the most neglected diseases,” she says. There is evidence of individuals in West Africa being exposed to the virus that causes Lassa fever, but not becoming infected, which suggests a mutation is lending them resistance.

The test will enable researchers to drill down and pinpoint the specific regions on the genome that are lending resistance, and then determine what those regions do – whether they code for proteins, or whether they regulate the behaviour of other genes. This opens up a potentially fruitful way to study diseases.

“Rather than do large disease studies, we may actually be pinpointing important functional mutations just by mining the genome,” says Sabeti. “We have about ten new novels – or candidate genes – that we’re interested in that have all been implicated in infectious diseases.”

Furthermore, as genome sequencing comes down in cost and more complete genomes are made available for research in coming years, the more we can learn from the likes of Sabeti’s latest opus. The paper featuring the CMS was published in Science early this year.

This feature appeared in the January/February 2010 issue of Australian Life Scientist. To subscribe to the magazine, go here.