Across the three billion ‘letters’ of our DNA, we each carry around six million variations. Researchers are unravelling their effects on our lives.
“The day before, we'd been scooting round Whipsnade Zoo. She was seemingly happy and healthy,” remembers Nikki Speed, of the life-changing December day just before Christmas in 2013.
That morning, Nikki went to wake her two-year-old daughter Rosie.
But Rosie was never to open her eyes again.
There isn’t a ‘big enough word’ to describe the impact on the family, says Nikki, who became one of many grieving parents who looked for the answers among the single-letter variations we all carry in our DNA.
Solving the mystery of how these variations affect our health – how we live and die – is a vital scientific quest. As well as providing answers for parents like Nikki, it could reveal the causes of inherited cancers, identify patients who could benefit from precision drugs, and help settle appalling miscarriages of justice.
And thanks to new gene-editing technologies, scientists are working out how to predict the effects of these subtle DNA changes before they even occur.
“I sat with my mum in her kitchen, and we went through the weeks leading up to Rosie’s death – every mouthful of pesto pasta she ate – trying to find what I'd missed. We were just told we needed to wait for the post-mortem. So we just waited. We were numb.”
But the post-mortem turned up nothing, nor did further investigations to rule out cardiac causes. “I also had a little boy who was nearly five. I was paralysed with fear for him too,” Nikki says.
Each year in the UK, around 40 children aged between one and 18 die in unexplained circumstances, often while asleep – that’s more than one every fortnight. The phenomenon has a bland acronym: SUDC, Sudden Unexplained Death in Childhood.
In many cases the culprits are thought to be hiding in our DNA. And for that reason, a key technique in molecular biology – DNA sequencing – is being brought to bear on SUDC.
Since the first human genome was mapped in the early 2000s, researchers have sequenced the genomes of millions of people, discovering over a billion individual variations in our genetic code. Some are common, most are rare. Some are slap-bang in the middle of vital genes, others lurk in the vast, ‘dark’ spaces between them. A landmark study in 2022, using data from the UK Biobank, estimated that we each carry around six million variants across our genome’s three billion DNA letters.
Rosie and Nikki Speed. Credit: Nikki Speed and Michael Bowles.
But these variants can have very different effects on our health. At one extreme, so-called pathogenic variants can cause inherited forms of diseases like cancer, and rare but lethal diseases that affect a handful of people worldwide. At the other end of the spectrum are benign variants – harmless genomic background noise.
After Rosie’s death, Nikki Speed turned to DNA sequencing for answers. And since it wasn’t yet routinely available in the UK, she found hope in New York, where a project called the SUDC Registry and Research Collaborative was offering bereaved families genetic testing, based on current knowledge of genes linked to sudden death.
Being accepted onto the project was a “huge moment”, Nikki says. “I felt like I was doing something”.
The results, which arrived a year later during a family holiday in Florida, were frustratingly inconclusive. They’d done all they could.
Nevertheless, Nikki remains hopeful: “Science evolves, and one day they might link more genes [to SUDC]. So we might get answers in time.”
But for other parents, the mystery surrounding their child’s death has triggered even further tragedies. And even when science has suggested answers, society has been agonisingly slow to act.
More than a decade before the human genome was deciphered, in February 1989, Kathleen and Craig Folbigg celebrated the arrival of their first child, Caleb. Less than three weeks later, in the middle of the night in their suburban flat in New South Wales, Australia, Caleb stopped breathing. Craig called an ambulance, but there was nothing the paramedics could do. Caleb was just 19 days old. The cause of death was recorded as Sudden Infant Death Syndrome, SIDS (SIDS is a term reserved for children under one; above that age, deaths are recorded as ‘SUDC’).
In 2003, Kathleen Folbigg was sentenced to forty years in jail. Credit: Fairfax Media/Getty.
Over the next decade, the couple would lose three more children, all in unexplained circumstances. Patrick, born a year later, developed epilepsy and blindness, and died aged eight months old. Two years later, the couple’s third child, Sarah, failed to wake from her sleep at 10 months old.
And in 1998, the couple welcomed their fourth child, Laura. A battery of post-natal tests came back normal, and a year later the family celebrated her first birthday. But several months later, Laura, too, failed to wake from an afternoon nap. She was 18 months and 22 days old.
Four unexpected deaths in a single family might flag at least a suspicion of an inherited condition. But a pathologist recorded Laura’s death as “undetermined” and raised the idea of homicide. After a police investigation, in 2001 Kathleen – then 34 – was arrested and charged with the murder of her four children.
A lengthy trial followed, and despite no direct evidence of smothering any of the children, in 2003 a jury convicted her of three counts of murder, one count of manslaughter and one count of grievous bodily harm.
Vilified in the media as “Australia’s worst mother”, the judge sentenced her to forty years in jail.
When immunologist Carola Vinuesa, a principal group leader at the Crick, talks about Kathleen, it’s with an intensity born of a deep commitment to justice.
“Right now, there are hundreds of women around the world facing criminal sanctions for hurting or killing their children,” she says, but for many the courts have either failed to consider genetics, or, if they have, have only carried out limited testing. It’s not hard to understand her frustration, given her experience: Carola led the team of researchers and legal experts who challenged Kathleen's conviction using evidence buried within the family’s genome.
Carola’s involvement began in 2018, while working at the Australian National University in Canberra. A former student in her department, now a lawyer, rang her to ask whether Kathleen's case might be helped by genetic testing, and, if so, whether Carola could help.
Before committing, Carola asked to look over their medical records, and was struck by several details. Caleb had laryngomalacia (‘floppy larynx’ – known to cause obstruction of the airways). Laura’s post-mortem had revealed inflamed heart muscle, myocarditis. Her sister was being treated for a respiratory infection at the time she died. Together with Patrick’s epilepsy, she felt there were alternative explanations. She took on the case.
Carola Vinuesa and Kathleen Folbigg at the Crick in 2025. Credit: Michael Bowles.
Three months later, she and her colleague Todor Arsov had sequenced Kathleen’s genome, and begun looking for variants in around 350 genes they had identified as either linked, or potentially linked, to sudden deaths in children.
They spotted something almost immediately: tucked away on the short arm of chromosome 2, a single-letter change in a gene called CALM2.
It would take years more to build proof sufficient to convince a court, but Carola is clear that, even back then, they strongly suspected this variation was harmful, thanks to several unique features of CALM2 and the protein it makes, calmodulin.
As researchers have sequenced DNA from organisms across the major branches of life – including higher organisms like plants, fungi and animals – they’ve discovered that certain genes are far less variable than others. And that’s because these so-called conserved genes are fundamental to life.
CALM2 is among these genes, and it's “extraordinary” for two reasons, says Carola.
First, humans have not one but three separate CALM genes – prosaically called CALM1, CALM2 and CALM3. Usually, genes that exist in multiple copies have evolved subtly different roles, with subtly varying sequences. But, uniquely, our three CALM genes seems to make proteins that are 100% identical to each other, molecular triplets, unlike any other family in the human genome. In other words, says Carola: “calmodulin is absolutely vital.”
That’s backed up by medical science: rare CALM gene variants cause serious conditions, like ‘long QT’ syndrome, which causes fast, chaotic heartbeats linked to fainting, seizures, and – occasionally – sudden death. In fact, whenever variants in CALM2 have been spotted, they’re never benign – it doesn’t seem to tolerate any changes to its sequence.
And second, across the entire tree of life, the equivalents of the CALM gene are strikingly similar. What’s truly remarkable, says Carola, is how perfectly preserved the calmodulin protein itself is – right down to its molecular building blocks. One amino acid in particular, glycine-114, stands out. “It’s not just unchanged in all animals – it’s conserved in plants and fungi too,” she explains. “Glycine-114 is an essential part of an essential protein.”
And back in 2018, the CALM2 variant she and Todor had spotted in Kathleen Folbigg’s gene altered that essential, conserved glycine-114. That meant it was highly likely to be harmful.
A few months later a legal inquiry was launched, with Carola involved, and samples from the Folbigg children’s DNA were analysed: both Sarah and Laura Folbigg carried the variant too.
But there was a disagreement among the geneticists on the team: this precise CALM2 variant had never been seen in a human before. Some considered it a ‘variant of unknown significance’, whereas Carola and her team considered it was likely to be pathogenic.
Frustratingly, the inquiry judge concluded there was no possibility that the variant had caused Sarah and Laura’s deaths.
However, through painstaking laboratory work led by a Danish team, and analysis of other children with glycine-114 variants in CALM3, in 2021 Carola and her colleagues published overwhelming, peer-reviewed evidence that the Folbigg family’s CALM2 variant was pathogenic.
Speaking at an event at the Crick this year, Kathleen recalled the day she heard the news. “It was bittersweet,” she said. “I cried for a long time, and there was a lot of soul searching. Because I’d been telling everyone ‘I didn’t do anything, I didn’t do anything’ ... but, genetically, I did.”
The discovery led to the launch of another inquiry, and this time the judge was convinced. In December 2023, 20 years after her conviction, and five years after the CALM2 variant was first spotted, Kathleen’s case was referred to the court of appeal, and her conviction was quashed.
She was, finally, free.
The thread that weaves Nikki and Kathleen’s stories together – beyond the obvious tragedy – is the quest to unpick the human consequences of genetic variation.
Nikki is still waiting for answers that only research can provide. And for Kathleen, the uncertainty around a single, never-before-seen genetic variant held up justice for five years.
But what if we could systematically test large numbers of variants in the human genome before tragedies occur? What if we could sketch out the map in advance?
And that is what one of Carola’s colleagues, Greg Findlay, is trying to do.
It was a sunny day in Boston, Massachusetts in 2012, when “something just clicked,” says Greg. “I was sitting in a Harvard Medical School courtyard reading a paper and I just thought, wow, that’s cool, that’s where I want to take my career”.
Greg Findlay in his lab at the Crick. Credit: Michael Bowles.
Then a lab technician, Greg had just read how researchers had used DNA sequencing in a novel way. Sequencing costs had plummeted from millions of dollars to thousands, allowing them to go beyond interrogating whether individual DNA sequences were present in a sample – i.e. a qualitative measurement – and instead quantitatively measuring how much of each different sequence was present, by sequencing the same sample many times and counting how many times each sequence was detected.
In Greg’s mind, this opened the door to studying genetic variation on an unprecedented scale.
“Historically, it’s been too laborious to test individual variants one by one,” he says, so variants get tested after they turn up in someone’s genome.
Moving from Harvard to the University of Washington in Seattle for his PhD and using the new technique he read about that day, Greg went on to develop a way to catalogue human variants by making every possible single-letter DNA change along a given gene, then measuring all their effects at once – something he called ‘saturation genome editing’, or SGE. “With SGE, we can test every potential variant up front, then see whether they are spotted clinically afterwards. It's prospective instead of retrospective testing.”
SGE relies on a technique called CRISPR-Cas9 to precisely edit individual letters in a stretch of a cell’s DNA. Whereas previous techniques relied on artificial and limited systems to study variants, says Greg, “CRISPR changed everything. It allowed us to precisely place variants into an authentic context – a living cell’s genome – and to study a much wider range of variants.”
After multiple variants have been edited into cells' DNA, the next step in SGE is to use qualitative DNA sequencing to measure their effects, and so reveal potentially harmful variants on an unprecedented scale.
Greg first published a proof-of-concept of this method in 2014 and, at first, “no one really seemed to notice”. That changed in 2018, when Greg and his colleagues carried out a landmark analysis of the infamous BRCA1 cancer gene.
Our genetic code tells our cells how to make proteins out of building blocks called amino acids.
The order of amino acids determines how the protein works. Variants are DNA changes that alter the sequence of letters. They can be insertions, deletions, or substitutions.
Variants can occur naturally, through mistakes as cells replicate their DNA, or can be caused by cigarette smoke, UV radiation or other exposures that damage DNA.
Sometimes, variants can change order of animo acids in a protein, causing it to malfunction. They can also be experimentally induced by gene-editing techniques such as CRISPR.
"CRISPR changed everything. It allowed us to precisely place variants into an authentic context – a living cell's genome – and to study a much wider range of variants," says Greg Findlay.
First linked to the disease in 1990, people who inherit certain BRCA1 mutations have a greatly increased risk of a range of cancers – notably breast and ovarian cancer. Since then, tens of millions of people have had their BRCA1 genes sequenced, turning up more than 4,000 pathogenic variants and thousands more of unknown effect. Greg’s 2018 study correctly, and independently, identified many of the pathogenic variants – but it went further, flagging hundreds of ‘unknown’ variants as potentially harmful.
“That's when people really took note, and started to use it in their own labs,” he says. In 2020, Greg relocated from the US to the Crick to establish his own lab and refine the technology, while SGE programmes have been established at major centres like the Wellcome Sanger Institute in Cambridge.
In 2024, Greg’s team mapped another gene, VHL, linked to cancers including kidney cancer and a rare adrenal gland cancer called pheochromocytoma. As with BRCA1, they found “perfect accuracy” in identifying variants known to cause kidney cancer but also shed new light on how different VHL variants were linked to different cancers.
These discoveries have been hugely influential, says Clare Turnbull, a clinical geneticist at London’s Royal Marsden Hospital. “The idea that you could pre-annotate every possible variant using functional data means that, if you encounter a variant clinically, you’d be more confident in classifying it as pathogenic or benign.” And that would ultimately mean better information for people potentially at risk of inherited disease.
What’s particularly hopeful, says Clare, is the prospect of blending Greg’s methods with software that predicts variant effects ‘in silico’. “This would help us interpret rare variants of unknown significance found in people tested in our clinics,” she says.
It also starts to unpick a particularly thorny issue. “A lot of our information on genomics comes from large-scale population sequencing of predominantly White people,” Clare says. That means we lack information on harmful variants that are more common in people of other backgrounds. Proactively generating data on all possible variants could level the playing field for people of all ethnicities.
Integrating SGE with computational data is on Greg’s mind too, but in a different way. Rather than testing every single variant in each of the 20,000 genes in the human genome – something that would take “around 50 years or more” - he sees real promise in using SGE data to train computer models to make much more accurate predictions. “But that’s going to take a lot of experimental data,” he says.
As well as continuing to focus on cancer – including an analysis funded by Cancer Research UK of 18 key cancer genes – Greg’s team is further refining SGE. Recently, they adopted a newer gene editing technique called prime editing, allowing them to overcome a major limitation of traditional CRISPR editing. CRISPR can only introduce variants across a 200-DNA-letter-long stretch of a gene, requiring multiple experiments to ‘walk’ along a gene's length to induce all possible variants. Prime editing allows SGE experiments that target much larger regions of the genome. “It’s massively sped things up,” he says.
The implications of SGE go beyond identifying inherited conditions. It could help more patients get new targeted cancer drugs, which rely on spotting particular defects that arise in their cancer. “For example, if we can say that a previously uncategorised BRCA1 variant likely causes loss of function, that suggests the patient’s tumour might be sensitive to drugs called PARP inhibitors,” says Greg, opening the way for more people to benefit from these drugs.
This digram shows the results of a large-scale experiment, testing over 11,000 genetic variants (across 143 genes) observed in rare disease patients from the Genomics England 100,000 Genomes Project. Variants identified as pathogenic by the experiment are highlighted in bright yellow at the centre – 492 of these have been classified as loss-of-function. Credit: Chloé Terwagne.
Carola thinks technologies like SGE could hold the key to liberty for individual women around the world. And she hopes that the justice system can take a pragmatic view of the evidence they can provide.
“We're not talking about the greater than 90% certainty you need for a clinical decision like whether to have a mastectomy,” she says.
“We're talking about the fact there's a woman in jail for having killed her children, and she’s got a mutation in a gene that is likely to cause early infant death.”
Since her release, Kathleen has taken a pragmatic approach to rebuilding her life. “It's a good thing to be able to stand up and say, science did this, science gave me peace of mind. And science helped me sit here today to be able to talk to everyone,” she told the audience at the Crick. “I'd like to think that the legal system could listen to scientists a bit more often.”
Rachel Sylvester, Carola Vinuesa, Nikki Speed and David Wallace at the Falsely Accused: Justice by Genomics event at the Crick in 2025. Credit: Michael Bowles.
Meanwhile, Nikki has co-founded a charity, SUDC UK, to raise awareness, fund research and support affected families. Thanks in part to their campaigning, in 2023, the NHS made genomic testing routinely available to these families, something that felt “really validating, like people finally understood.”
And like Kathleen, she has huge hopes for what research can offer, both personally – in that it may one day reveal the cause of Rosie’s death – but also for others affected.
“The ideal outcome would be that we identify biomarkers that can help predict and prevent SUDC,” she says, pointing out that since 1990, research on SIDS has been accompanied by an 80% reduction in deaths in children under 1 year old.
“Since the 90s, there have been over 13,000 publications for SIDS, but there have been fewer than 100 for SUDC,” she says. “We know what we need to do, we just need to do it. If we do the research, we’ll be able to save children's lives.”
Science from inside the Crick.
Browse this issue
