In the latest episode of The Genetics Podcast, Patrick Short is joined by Dr. Luke Pilling, Senior Lecturer in Genetic Epidemiology at the University of Exeter. Luke works with large datasets such as UK Biobank, linking genetics, medical records, and other data. His goal is to understand why some people develop age-related conditions and others don't.
The conversation covers hemochromatosis and the case for earlier diagnosis, the genetic drivers of multimorbidity, and drug response. One argument connects all three topics: knowing who carries a variant is the first step. The more useful question is who will become ill, and when, so intervention can start early.
Around 1 in 150 people of Northern European ancestry are homozygous for the HFE C282Y variant. Even so, hemochromatosis diagnosis is often delayed or missed.
Luke described the main clinical challenge as finding people early enough to start phlebotomy, the removal of blood. It is safe and effective when started early, and it is currently the only real treatment.
The NHS has no organized screening program for the condition. The UK National Screening Committee found insufficient evidence to support population screening.
Luke pointed to targeted case finding in higher-prevalence regions, such as parts of Ireland and Scotland, as an approach that may be justified. He drew on Jim Wilson's work on regional HFE prevalence. His hemochromatosis work, which began with his mentor David Melzer, includes a paper in The BMJ.
But a positive genetic test is only the starting point. Screening runs into a second problem: penetrance. Not every homozygote develops disease, so a genetic result alone cannot tell a clinician who needs treatment. Luke put it directly:
"Even if you have a positive genotype test, you find out you're a homozygote. Not all of them go on to develop clinical symptoms or clinical disease. So, yes, we could identify people who have the variant, but then they'd need to undergo a lot of tests to find out do they actually need the intervention."
His team's analyses show how other factors shape the outcome. In homozygotes, waist-hip ratio and polygenic risk scores add to the effect of the variant, steering people toward specific outcomes such as diabetes or osteoarthritis.
For teams designing genetically stratified studies, this changes how a population is defined. Variant status identifies who is at elevated risk, and other genetic and environmental factors shape who progresses and toward which condition.
Luke sees Our Future Health as a strong opportunity to study this. The cohort will include around 5 million people and roughly 30,000 C282Y homozygotes, most of whom are probably undiagnosed. Participants have consented to be recontacted, which allows researchers to follow up with the people most likely to need intervention.
Luke is part of GEMINI, an MRC-funded project led by Tim Frayling that studies the genetics of multimorbidity. The team analyzed 84 common long-term conditions and found obesity to be a large driver of multimorbidity. Their findings are published in eBioMedicine.
Colocalization analyses identify variants shared between diseases. Some shared signals point to drug repurposing opportunities, as with LPA. Others run in opposite directions: a CTLA4 variant that raised rheumatoid arthritis risk lowered the risk of a cancer.
"In cancer you want to promote the immune response, and in an autoimmune condition you want to suppress it... But what was surprising to us was that that might imply that the treatments for one might be increasing the risk of the other disease."
The team is also mapping the order in which diseases appear. Hypertension often comes first, and anemia and iron status show up as intermediate points where clinicians could intervene. Luke named the gaps that limit this work: primary care data, ferritin and transferrin measures, proteomics, and the accuracy of clinical codes.
Pharmacogenetics follows the same logic. In the UK Biobank, people with CYP2C19 loss-of-function variants who took clopidogrel had an increased risk of stroke. Some NHS trusts are now rolling out point-of-care tests.
Luke also discussed the TWIST framework, work with Jack Bowden that estimates the causal effects of genetic variants on drug response. He was clear about the limits. For most common drugs, he said, the evidence for genotype-guided prescribing is not there yet, though it is growing.
The field also remains European-centric. Luke wants to collaborate with Asian biobanks to broaden the populations this research covers.
Luke's earlier research covered the genetics of lifespan and frailty. Lifespan GWAS signals tend to point to specific diseases, such as cardiovascular disease, rather than geroscience hallmark pathways like senescence. He said he expected to find more of those pathways than his team has.
His reading has two parts. Living a long, healthy life depends on avoiding the common diseases researchers already know a lot about. Current datasets also probably lack the right measures, such as tissue-specific data and mitochondrial phenotypes.
Each finding in this conversation rests on cohorts that link genotypes to medical records and allow recontact. The open work sits in the gaps Luke listed: primary care records, iron biomarkers, proteomics, cleaner clinical codes, and more diverse populations. Closing them would help researchers estimate who will become ill and when, across hemochromatosis, multimorbidity, and drug response.
Listen to the full episode below.