The type 1 diabetes genetic risk score, or T1D-GRS, is a tool that has already proven its worth in diabetes prediction and classification. It aggregates information from ten single nucleotide polymorphisms, or SNPs—individual positions in the genome where the DNA letter varies between people and where certain variants shift the probability of developing the autoimmune form of diabetes. Because the score draws on genetic variants that remain stable throughout life, it can help distinguish type 1 from type 2 diabetes in young adults and identify individuals at elevated risk long before symptoms appear. Crucially, the T1D-GRS and islet autoantibodies, the immune markers that signal the body’s attack on insulin-producing cells, act as independent predictors of diabetes, which makes pairing them particularly powerful for screening and diagnostic purposes.

The obstacle has always been practical rather than scientific. To calculate the score, laboratories traditionally need genomic DNA extracted from white blood cells, which involves purification, quantification and storage steps that add expense at every turn, before a single genotyping reaction has even been set up. The Exeter team reasoned that this entire pipeline might be unnecessary. Serum, the very material already used for islet autoantibody assays, contains extracellular DNA released from cells that burst during the clotting process. If that DNA were sufficient for genotyping, the same tube of blood could serve both the antibody test and the genetic score, eliminating the need for a separate DNA extraction altogether.

There was a catch, however, and it lay in the biochemistry of clotting. The extracellular DNA in serum comes from leukocytes that rupture as the blood clots, so the amount of genomic DNA available depends on how long the sample is allowed to clot before being spun down in a centrifuge. Many standard serum processing protocols specify clotting times of less than thirty minutes, a window that previous work suggested may be inadequate for releasing enough genomic DNA into the fluid. The researchers therefore designed their study to answer two linked questions: whether a direct-from-serum genotyping method could calculate the ten-SNP T1D-GRS accurately, and how strongly the answer depended on clotting time.

The technical setup was deliberately conventional, which is part of the method’s appeal. The team used TaqMan SNP genotyping assays, a widely adopted fluorescence-based qPCR chemistry, together with TaqPath ProAmp Master Mix, in tiny five-microliter reactions containing just one microliter of either purified DNA or raw serum. Plates, prepared on a Biomek i7 Automated Workstation and run for fifty cycles on a QuantStudio 12K Flex Real-Time PCR System, included positive and no-template controls, and genotypes were called automatically using commercial software. In other words, nothing exotic was required—only standard reagents, standard instruments and a single microliter of serum per SNP.

Accuracy came first. In a cohort of 116 matched serum and whole-blood-derived DNA samples sent to the Exeter Genomics Laboratory from across the UK and Ireland, all with clotting times exceeding 24 hours, both biospecimens achieved call rates above 95 percent across all ten SNPs on the first pass, with no repeated reactions. A complete T1D-GRS was obtained from 84.5 percent of serum samples compared with 96.6 percent of purified DNA samples, a statistically significant difference, but among the 94 samples where both approaches yielded a complete score, the results agreed perfectly—100 percent concordance. In a subset of 32 samples with sequencing data available, the serum-derived scores were again fully concordant, confirming that the shortcut does not come at the cost of correctness.

The clotting-time experiment revealed where the method’s limits lie. The team analyzed 122 serum samples with recorded clotting times ranging from 12 to 475 minutes, running every sample in duplicate on separate qPCR runs to check reproducibility. Splitting the cohort at the thirty-minute mark produced a striking contrast. Samples clotted for thirty minutes or less achieved a genotyping success rate of just 68.8 percent, a replicate discordance rate of 28.8 percent, and yielded a complete ten-SNP score in only 7.4 percent of cases. Samples clotted for longer than thirty minutes, by comparison, reached a 97 percent genotyping success rate, only 2.2 percent replicate discordance, and complete scores in 77.8 percent of samples—a difference the authors report as highly significant. Performance in this longer-clotting group closely matched that of the samples clotted for more than 24 hours.

The explanation, the researchers argue, is straightforward: insufficient clotting time leaves too little extracellular DNA in the serum for reliable amplification. This aligns with earlier findings that PCR-based genotyping from samples containing very low DNA concentrations—below roughly 56 picograms—is unreliable. The practical message for laboratories hoping to combine genetic and biochemical testing is that serum should be allowed to clot for more than thirty minutes before processing, a duration that does not compromise autoantibody assays. Where additional tests sensitive to proteolysis, glycolysis or other degradation processes are needed, the authors suggest an intermediate clotting window of more than thirty minutes but under about three hours may offer the optimal balance between DNA release and analyte preservation.

The implications extend well beyond the immediate assay. A direct-from-serum approach could unlock cost-effective genotyping in the many large frozen serum biobanks around the world where no extracted DNA exists, allowing researchers to mine decades of stored samples for genetic risk information without retrieving new blood draws. Although the study tested only ten SNPs, the low input volume of one microliter per SNP suggests the same logic could scale to larger genotyping panels or alternative genotyping chemistries. For clinical laboratories that lack DNA-extraction capability, the proof-of-principle result points toward direct, automated assays that run antibody and genetic tests in parallel from a single serum sample. The work was supported by the National Institute for Health and Care Research Exeter Biomedical Research Centre, and the authors note existing research funding and licensing arrangements with Randox Laboratories to develop a T1D-GRS assay, underscoring the translational momentum behind the approach. If the method holds up in broader validation, the humble serum tube—already the workhorse of clinical chemistry—may become a one-stop source for both immune and genetic markers of diabetes risk.

Subject of Research: A direct-from-serum genotyping method for calculating a type 1 diabetes genetic risk score

Article Title: A direct-from-serum method for calculating a type 1 diabetes genetic risk score

Article References: Spurrier, B., Oram, R. A., Thomas, N. J., & Locke, J. M. (2026). A direct-from-serum method for calculating a type 1 diabetes genetic risk score. Journal of Translational Medicine, 24(1), Article 1148. https://doi.org/10.1186/s12967-026-08930-4

Image Credits: AI Generated

DOI: 10.1186/s12967-026-08930-4

Keywords: type 1 diabetes, genetic risk score, serum, genotyping, TaqMan, qPCR, SNP, islet autoantibodies, DNA extraction, clotting time, biobanks, precision medicine

The scientific rationale behind the programme rests on a conceptual shift in how type 1 diabetes is understood. Earlier work by members of the consortium, including Dr Peter Achenbach and Professor Anette-Gabriele Ziegler of the Institute of Diabetes Research at Helmholtz Munich, established that the disease progresses through distinct presymptomatic stages before clinical onset. What was traditionally recognised as symptomatic diabetes is now classified as stage 3, while stages 1 and 2 describe earlier phases in which the immune system has already begun attacking the insulin-producing beta cells of the pancreas but blood glucose remains normal or only subtly abnormal. Identifying children during these presymptomatic windows matters because preventive and disease-modifying therapies are now becoming available. One such therapy, teplizumab, has already been approved to delay progression, and others are in development. Early detection also carries an immediate clinical benefit: children whose condition is known before symptoms appear face a substantially reduced risk of presenting with diabetic ketoacidosis, a life-threatening metabolic emergency that often marks the first diagnosis of type 1 diabetes.

The EDENT1FI consortium was designed to advance population-based screening and to harmonise early detection approaches across national borders. Using a common protocol, participating centres set out to assess the feasibility of screening for islet autoantibodies, to measure participation rates, to determine the prevalence of early-stage type 1 diabetes in the general child population, and to evaluate the performance and comparability of laboratory assays across different countries. Screening sites operated in the Czech Republic in Prague; in Denmark and Sweden in the Copenhagen and Öresund Region; in Germany in Munich, Hanover and Dresden; in Italy in Milan; in Poland in Warsaw and Katowice; in Portugal in Lisbon, Matosinhos and Beja; and in the United Kingdom in Birmingham.

The technical workflow followed a tiered design intended to balance sensitivity with practicality. Initial testing relied on capillary blood samples, collected as dried blood spots in the United Kingdom, and used a combined assay for three autoantibodies: GADA, directed against glutamic acid decarboxylase; IA-2A, directed against insulinoma-associated antigen-2; and ZnT8A, directed against zinc transporter 8. Samples that tested positive were then analysed for four autoantibodies, adding IAA, insulin autoantibodies, at two central laboratories in Munich and Milan. Participants with at least two confirmed autoantibodies underwent further confirmatory testing, and a diagnosis of early-stage type 1 diabetes was made only after this confirmation. Individuals diagnosed with early-stage disease were invited to receive education about their condition and to undergo metabolic staging through oral glucose tolerance testing and measurement of HbA1c, which distinguishes stage 1 disease, where glucose metabolism remains normal, from stage 2, where dysglycaemia has emerged, and from stage 3, where clinical symptoms develop.

Crucially, the screening targeted children from the general population, not only those with a known family history of the disease. Children and adolescents aged 2 to 10 years were eligible in Germany, while other countries screened those aged 2 to 17 years owing to site-specific variations in screening protocols. This general-population emphasis is scientifically important because the vast majority of children who develop type 1 diabetes have no first-degree relative with the condition, meaning that family-history-based risk stratification alone would miss most future cases.

In total, 140,568 young people, 49.2 percent of them female, took part in screening, with a median age of 7.6 years. Of these, 10,273 children, or 7.3 percent, had a first-degree family history of type 1 diabetes, while 127,754, or 90.9 percent, had no such relative, and family history was not reported for 2,541 participants. Among children with a first-degree relative with the disease, 219 were found to have early-stage type 1 diabetes, equivalent to 2.13 percent or roughly one in 45. Among children without a family history, 408 cases were detected, or 0.32 percent, approximately one in 350. These figures confirm that although genetic risk is elevated in relatives, the sheer size of the general population means that most presymptomatic cases are found among children with no family connection to the disease.

Country-level results revealed meaningful variation in prevalence. Among children without a first-degree relative with type 1 diabetes, frequencies ranged from 0.22 percent in Germany to 0.42 percent in the United Kingdom and 0.47 percent in Sweden. Prevalence was also associated with age: it was lowest, at 0.22 percent, among children younger than 4 years, rising to 0.31 percent in the 4 to under 8 age group and 0.40 percent among those aged 8 to under 13. In Italy, where screening took place in the EASD host city of Milan, 4,507 participants with a median age of 6.5 years were screened, including 3,641, or 80.8 percent, without a first-degree family history, of whom five, or 0.14 percent, had early-stage type 1 diabetes. No cases were detected among the 167 Italian children who did have a first-degree relative with the condition.

The larger national cohorts provided further detail. In the United Kingdom, 17,409 children with a median age of 8.4 years were screened; 59 cases, or 0.42 percent, were found among those without a first-degree relative with type 1 diabetes, and 67 cases, or 3.43 percent, among those with one. In Sweden, 7,926 children with a median age of 9.4 years were screened, yielding 36 cases, or 0.47 percent, among those without a family history and 4 cases, or 1.21 percent, among those with one. In Germany, where 43,470 children with a median age of 4.3 years were screened, case rates were notably lower, at 90 cases, or 0.22 percent, in children without a first-degree relative and 35 cases, or 1.36 percent, in those with one. The lower German figures may partly reflect the younger age distribution of the German cohort, given the observed rise in prevalence with age.

Of the 634 participants diagnosed with early-stage type 1 diabetes across the consortium, 386, or 61 percent, completed metabolic staging. The results showed that 300, or 78 percent, had stage 1 disease, 70, or 18 percent, had stage 2, and 16, or 4.1 percent, had already reached stage 3. This distribution underscores the value of antibody screening: the overwhelming majority of detected cases were identified at the earliest presymptomatic stage, when intervention opportunities are greatest. All participants identified with early-stage type 1 diabetes have been included in the European pre-T1D Registry, creating a harmonised continental cohort that can support future studies of disease progression and prevention.

The authors conclude that screening principles originally developed in Germany can be successfully applied to regions with different healthcare systems, and that the EDENT1FI initiative demonstrates that screening for early-stage type 1 diabetes can be effectively implemented throughout Europe. The next steps, they say, are aimed at integrating screening for early-stage type 1 diabetes in the general population into standard medical care. Similar screening initiatives are under way worldwide, including in the United States and Australia. Policy momentum is already visible within Europe: Italy has introduced legislation to establish nationwide screening for type 1 diabetes, although the government and regional health systems are still working to implement the law. The consortium recommends that screening be rolled out across Europe with three testing windows, at ages 2 to 4 years, 6 to 8 years and 10 to 15 years, an approach designed to capture the rising prevalence observed across childhood age groups while embedding early detection into routine paediatric care.

Subject of Research: Population-based screening of children for presymptomatic type 1 diabetes using islet autoantibody testing across eight European countries

Article Title: Study across eight European countries shows that screening for type 1 diabetes can be effectively implemented in children across Europe (EDENT1FI)

Article References: Study across eight European countries shows that screening for type 1 diabetes can be effectively implemented in children across Europe (EDENT1FI). (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: type 1 diabetes, EDENT1FI, islet autoantibodies, screening, EASD, presymptomatic stage, teplizumab, pediatrics, Europe, autoimmunity, diabetic ketoacidosis, public health