Among 424 patients who remained undiagnosed after complete genome sequencing, a team of researchers measured nearly 1,500 proteins circulating in their blood. The result: thirteen diagnoses confirmed and twenty-three additional genetic leads, where DNA alone had not provided a verdict.
The starting observation can be boiled down to one sentence: introducing genome sequencing for the diagnosis of rare diseases did not permit identifying a genetic cause for the majority of patients.
That is the starting point for the work carried out by Queen Mary University of London, the Berlin Institute of Health at Charité, and Genomics England. The study analyzed blood samples from people with rare diseases who remained genetically undiagnosed after an assessment via Genomics England’s 100,000 Genomes Project. These patients had already benefited from the best available tool, whole-genome sequencing, without a single variant being definitively incriminated. Their DNA file existed, but it remained silent on what was truly causing their illness.
That is where the second layer of information comes in: biology in action, not just the code that programs it.
- The measurement of 1,463 blood proteins resolved 13 diagnostic cases that had remained unanswered after whole-genome sequencing.
- Proteins reveal the actual functional effect of a mutation where DNA alone leaves doubt about its function.
- This approach particularly benefits missense variants and splice-site variants, the most challenging to interpret genetically.
What DNA Cannot Say by Itself
A genetic variant is not evidence in itself. It can be a harmless variation, a true disease-causing culprit, or a complete unknown, what we call a variant of uncertain significance. These uncertain-significance variants are particularly difficult to classify because sequencing data alone may not reveal their biological effects. The genetic text describes an instruction, but it does not guarantee that it has been executed in the patient’s cells.
Proteins, on the other hand, are the tangible result of that execution. As Claudia Langenberg, director of the Precision Health Research Institute at Queen Mary University of London and corresponding author of the study, puts it, “the genome gives us the plan, but proteins can tell us something about how that plan is translated into biology in a given patient.” A damaged gene often, but not always, produces a protein in abnormal quantity. Detecting this anomaly in the blood provides functional evidence that the suspected variant truly has an effect, rather than remaining in theoretical doubt.
Thirteen Diagnoses Resolved, Twenty-Three Leads Opened
Concretely, the researchers employed a protein profiling test, the Olink Explore 1536, capable of measuring a panel of 1,463 proteins in the serum of each patient. For thirteen of them, the detection of abnormally low protein levels (a z-score below -2) enabled confirmed genetic diagnoses, either by resolving variants of uncertain significance or by guiding targeted reanalysis of the genome toward specific genes. The blood thus served as an arbiter where the sole genetic text left room for doubt.
For twenty-three other patients, the approach did not provide a definitive answer but opened a serious lead. The researchers identified candidate links between genes and diseases as well as variants, thanks to a convergence of evidence linking abnormally low proteins and variants ranked by the Exomiser prioritization tool. A single example nicely illustrates the mechanism: a rare heterozygous variant in the TIE1 gene was found only in a patient with an abnormally low serum level of the corresponding protein, and in his father, both affected by the same monogenic heart condition, while no other individual in the 100,000 Genomes Project carried this variant. Two converging clues, one family, a diagnosis that becomes credible.
Technically, two types of variants benefited most from this approach. Missense variants (52.5%) and splice-site variants (27.5%) accounted for the majority of the diagnostic or candidate variants identified. These are precisely the categories most difficult to interpret from the DNA sequence alone, because their functional impact is not always evident on paper.
A Complement, Not a Universal Test
It would be tempting to imagine this protein dosage as a new blood test available for every patient drifting in diagnostic limbo. This is not the case, and the authors state it plainly. The current technology measures only a fraction of proteins encoded by the human genome, and not all proteins relevant to a disease can be reliably detected in the blood; moreover, not all pathogenic variants change the quantity of circulating protein. Concretely, thirteen diagnoses out of 424 patients represent a real advance, but far from a universal solution.
Athanasios Kousathanas, principal data scientist at Genomics England and corresponding co-author, contextualizes the challenge: “one of the biggest challenges in rare disease genomics is understanding which genetic variants can contribute to a person’s condition.” Protein data do not replace sequencing; they illuminate it. Before a clinical broadening, several technical hurdles must still be overcome. Advances in the number of measurable proteins, the sensitivity of the assays, and the establishment of robust reference ranges will be necessary before this type of approach can be more widely integrated into clinical diagnosis.
One fact remains worth keeping in mind: this study mobilized the database of the UK 100,000 Genomes Project, a large-scale research platform, not a routine hospital tool. For the moment, no laboratory offers this combined dosage in routine practice, and patients seeking a diagnosis will largely continue to rely on conventional sequencing and its successive reanalyses.
Sources: ncbi.nlm.nih.gov | labmedica.com | science.org
