Only six other children in the world are known to have the same condition as Lillian Yuska. (Children’s Hospital of Wisconsin)

Born prematurely, Lillian Yuska struggled to feed, and she suffered from chronic gastrointestinal problems and repeated infections. After years of shuttling her from specialist to specialist, Lillian’s parents turned to cutting-edge technology: They had their daughter’s genetic code mapped. This genomic sequencing, which began when Lillian was 4, revealed that she had tricho-hepato-enteric syndrome-2, a condition caused by a gene mutation that disrupts gut function and immunity. Only six other children worldwide are known to have the condition.

Families dealing with rare genetic disorders will be the first beneficiaries of genome sequencing. But researchers say the technique may one day help find clues to such common conditions as heart disease and diabetes, which may also be triggered by rare mutations.

“Most common diseases are merely a collection of rare diseases that happen to have the same end point,” says David Dimmock, a paediatric geneticist at the Medical College of Wisconsin (MCW) in Milwaukee. In other words, common conditions are actually rare disorders in disguise, which means that the same methods that helped the Yuskas promise to deliver personalized medicine for us all.

Over the past three years, a clinic run by MCW and the Children’s Hospital of Wisconsin, where Lillian, now 7, was treated, has sequenced the genomes of about 40 patients. It got off to a spectacular start with its first patient.

At age 4, Nic Volker’s inflammatory bowel disease was so severe that food in his gut would leak into his abdomen. The clinic’s team zeroed in on a genetic mutation that can cause a leukemia-like disorder.

Hoping to avoid that outcome and suspecting that the defect might also be causing Nic’s immune system to attack his bowel, his doctors recommended a bone marrow transplant. It worked, and while it was too late to save Nic’s colon, he is leading a healthy life today.

Not every diagnosis has led to a similar breakthrough. But for families such as the Yuskas, ending the diagnostic mystery is a huge step forward. For a bleak couple of weeks, the leading contender for Lillian’s condition was Fanconi anemia, which typically kills people in their 20s.

Lillian’s diagnosis, based on the sequencing of her genetic code, holds more hope: While her health problems won’t go away, her life will not necessarily be cut drastically short. “Now we need to find out how best to care for her,” says her mother, Danielle Yuska.

Even when the news is the worst possible, families have been grateful to know what’s wrong. Dimmock recalls telling the mother of an infant with severe liver disease that the mutations responsible would quickly lead to fatal neurodegeneration. The mother cried, then turned to Dimmock and said: “Thank you. That really helps.” The diagnosis meant that her daughter could be spared a liver transplant that would have proved useless.

The root of common problems

Geneticists had hoped that heritability of common conditions such as heart disease would be explained in large part by a few relatively common genetic variants. It hasn’t turned out that way. Instead, the genetic roots of common disorders seem to be spread across a plethora of rare mutations, some of which may even be unique to their carriers.

This transforms the process of genetic discovery.

To find common disease-causing variants, you might recruit 1,000 people with the condition in question, plus a similar number of controls, and look for variants that seem to be inherited with the disease. But if 1,000 people with the same external symptoms have hundreds of different mutations causing them, finding the gene variants responsible becomes a needle-in-a-
haystack problem.

That might sound hopeless, given that a typical genome contains several million departures from the reference human DNA sequence. But you can winnow down the list of candidates using algorithms that predict whether a variant is likely to disrupt a gene’s function. A premature “stop” instruction, for instance, which cuts a protein short, is an obvious candidate for further scrutiny.

These algorithms work in various ways. Some flag mutations if they occur in DNA sequences that differ little from species to species. The thinking goes that if evolution has preserved a gene’s sequence, then any change in it is likely to be bad.

Other algorithms are more sophisticated, for instance assessing how gene variants might change a protein’s shape, which might affect how it works.

A mutation may be damaging to a protein without causing disease. This is because many of our metabolic pathways have a degree of redundancy, which means that when individual proteins “break,” there are often others available to take the strain.

While sequencing methods and the tools used to sift for damaging mutations have advanced far enough to spot single variants causing rare diseases, the signal-to-noise problem becomes more challenging when the symptoms are common. But as our knowledge of rare variants expands and prediction algorithms become more sophisticated, the genome’s dark matter should come into view.

Even at this early stage, there are valuable nuggets to be mined from a typical genome. The most useful information surrounds carrier status for genetic disorders and clues about how commonly used drugs are likely to work.

The full benefits of the genomic revolution will come when doctors are routinely able to pick out the rare variants triggering certain symptoms and prescribe drugs that have been specifically designed to correct the biochemical pathways concerned. This may still be decades away.

This article is excerpted from a longer story originally published by New Scientist magazine.