A team led by researchers at Amsterdam UMC and the Wellcome Sanger Institute set out to squeeze more value from the fluid intelligence measures collected in the UK Biobank, a resource of roughly half a million British volunteers. Fluid intelligence, the capacity to solve novel problems independent of acquired knowledge, was assessed through short verbal-numerical reasoning tests administered at the baseline in-person visit and in later online questionnaires. Only about 62 percent of participants ever completed one of these tests, and earlier studies typically analyzed just a subset of those. By integrating scores from five different testing occasions and then statistically imputing scores for participants who never took any test, the team expanded the effective sample from roughly 270,000 measured individuals to more than 455,000.

The imputation itself relied on a matrix-completion algorithm called SoftImpute, which fills in missing entries by exploiting low-dimensional structure in the data. The researchers began with 154 biobank variables correlated with intelligence, then deliberately pruned that set down to 82 variables whose genetic correlation profile aligned more closely with intelligence than with a residual noncognitive component of educational attainment. This refinement step proved crucial: it modestly reduced raw prediction accuracy, from a correlation of about 0.52 to 0.46, but reduced the importation of socioeconomic and other noncognitive signal into the imputed phenotype. Notably, an accuracy of 0.46 falls below the test-retest reliability of the fluid intelligence test itself, which the team measured at roughly 0.6, meaning the imputed values add noise but, according to the authors’ theoretical framework, need not add systematic bias.

That theoretical framework is one of the study’s most distinctive contributions. The authors formally derived the conditions under which pooling measured and imputed phenotypes recovers consistent population genetic effects. Two assumptions matter most: measurement error in the observed scores must be uncorrelated with genotype, and the imputed values must be calibrated to the intended cognitive construct in the unmeasured group, both in their mean and in their covariance with each genetic variant. Although these assumptions cannot be proven directly, they yield a testable consequence: the genetic correlation profile of the combined analysis should match that of an independent GWAS targeting the same construct. The team checked this against the COGENT consortium GWAS of cognitive function and found close correspondence.

The validation also exploited within-family GWAS, a technique that separates direct genetic effects from confounding sources such as genetic nurture, assortative mating and population stratification. In population-based analyses, a negative correlation between direct effects and the so-called nontransmitted coefficient has previously been interpreted as a signature of ascertainment bias. For the baseline fluid intelligence measure alone, that correlation was strongly negative at about minus 0.54. Integrating scores across testing occasions moved it closer to zero, and adding imputed values pushed it to a point estimate that was no longer significantly different from zero, exactly what the ascertainment-bias hypothesis predicted.

The payoff for common-variant discovery was substantial. Combining measured and imputed scores raised the number of independent genome-wide significant lead SNPs from 390 to 550, and meta-analyzing the result with COGENT pushed the total to 620 lead SNPs in a GWAS of 490,700 individuals. Larger sample sizes also improved gene prioritization, strengthened enrichment of associated genes in brain tissue, and monotonically improved polygenic index prediction of measured cognitive ability in three independent cohorts: the ALSPAC birth cohort, the Millennium Cohort Study and the INTERVAL study. Crucially, direct genetic effect estimates within families also increased with sample size, indicating that the gains were not an artifact of population-level confounding.

The most striking results came from the rare variant side. Using whole-exome sequencing data, the researchers tested whether ultra-rare protein-truncating variants and damaging missense variants in individual genes were associated with fluid intelligence. With measured scores alone, only five genes passed the false discovery rate threshold of 1 percent. With the combined measured-and-imputed phenotype, that number jumped to 26, a 5.2-fold increase in gene-level discovery, while effect size estimates remained essentially unchanged, suggesting the imputation introduced power rather than distortion. Fourteen of the 26 genes are already cataloged in the DDG2P database of developmental disorder genes, a significant enrichment given that only about 13 percent of tested protein-coding genes belong to that list.

Eight of the significant genes had no strong prior evidence of involvement in intelligence or neurodevelopmental conditions, yet the evidence that they matter is compelling. Across a cohort of more than 31,000 exome-sequenced probands with neurodevelopmental conditions, damaging de novo mutations in the 12 newly implicated non-DDG2P genes occurred 2.59 times more often than expected by chance. And in aggregate, protein-truncating variants in the newly discovered genes were significantly associated with lower measured cognitive ability in both ALSPAC and the Millennium Cohort Study, providing external replication that does not depend on the imputed phenotype at all. Comparisons with educational attainment showed the imputed intelligence measure outperformed that proxy, yielding 26 significant genes versus nine for educational attainment.

The findings carry a broader message about the genetics of cognition: rare variants with moderate to large effects on intelligence, estimated at roughly 0.25 to 1 standard deviations per carrier, are more pervasive than previously appreciated. The associations largely persisted even after excluding individuals with recorded diagnoses of intellectual disability, autism or epilepsy, implying that damaging variants in these genes can shape cognitive ability across the population without necessarily pushing carriers over a clinical diagnostic threshold. Polygenic background and environmental factors likely modify how such variants manifest clinically, blurring the traditional dominant-versus-recessive dichotomy into more of a gradient.

The authors are careful to frame the work as a pragmatic response to the limitations of existing biobank data rather than an endorsement of prediction as a substitute for measurement. Imputed phenotypes blur the boundary between a construct of interest and its correlates, and some residual noncognitive signal almost certainly remains in the combined measure, as reflected in slightly elevated genetic correlations with educational attainment and household income. The team argues that the most robust path forward is better phenotyping at the design stage, with routine collection of high-quality cognitive measures and deliberate strategies to minimize selection bias. Still, for researchers working with the biobanks that exist today, the study offers both a powerful tool and a principled framework for deciding when imputed traits can be trusted, and it opens a route to finding new neurodevelopmental risk genes hiding in the genomes of people who never took the test.

Subject of Research: Statistical imputation of fluid intelligence scores in the UK Biobank to reduce ascertainment bias and boost genetic discovery for cognition and neurodevelopmental conditions

Article Title: Imputation of fluid intelligence scores reduces ascertainment bias and increases power for analyses of common and rare variants

Article References: van den Berg, D. M., Malawsky, D. S., Huang, W., Danecek, P., Walter, K., Kalantzis, G., Wade, I., Palmer, D., Birney, E., Verweij, K. J. H., Smit, D. J. A., Lindsay, S. J., Hurles, M. E., Abdellaoui, A., & Martin, H. C. (2026). Imputation of fluid intelligence scores reduces ascertainment bias and increases power for analyses of common and rare variants. Nature Genetics. https://doi.org/10.1038/s41588-026-02787-5

Image Credits: AI Generated

DOI: 10.1038/s41588-026-02787-5

Keywords: fluid intelligence, UK Biobank, genome-wide association study, ascertainment bias, phenotype imputation, rare variants, polygenic scores, neurodevelopmental disorders, genetic correlations, SoftImpute, developmental disorder genes, cognitive ability

The German health-care context provides a particularly informative setting for such research. Germany maintained relatively comprehensive primary care structures throughout the pandemic, and several large national cohorts were already established before SARS-CoV-2 emerged, allowing investigators to draw on pre-pandemic baseline data. This pre-pandemic anchoring is a methodological advantage that cannot be overstated: it permits researchers to distinguish symptoms that are genuinely new after an infection from symptoms that existed beforehand but were perhaps unnoticed, unreported, or reinterpreted by patients in light of their illness. Recall bias is a persistent threat in post-acute syndrome research, and baseline assessments collected before infection substantially mitigate it.

When comparing individuals developing a post-acute infection syndrome for the first time with those experiencing a recurrence of prior symptoms, several conceptual distinctions become important. A new-onset syndrome implies that the infection triggered a qualitatively different health state, potentially through distinct mechanisms such as viral persistence, autoimmunity, microvascular injury, or dysregulation of the autonomic nervous system. A recurring syndrome, by contrast, suggests that the infection acted as a flare trigger for a pre-existing or previously quiescent condition, a pattern familiar from the way infections can precipitate relapses in autoimmune diseases or reactivate symptoms in chronic fatigue syndromes. The distinction has practical implications for prognosis, treatment selection, and the design of clinical trials, since patients whose symptoms are reactivations of an older illness may respond differently to immunomodulatory or rehabilitative interventions than those with genuinely novel pathology.

Epidemiological studies across multiple pathogens have established that acute infections of many kinds can be followed by prolonged symptom burden. Gastrointestinal infections caused by Campylobacter, Salmonella, and related organisms have long been linked to reactive arthritis and irritable bowel syndrome that persist for months or years. Epstein-Barr virus infection is a recognized precipitant of myalgic encephalomyelitis and chronic fatigue syndrome, a connection strengthened by studies of university students followed prospectively from the time of acute mononucleosis. Q fever, caused by Coxiella burnetii, produces a well-documented chronic fatigue syndrome in a substantial minority of patients, as demonstrated in outbreak investigations in the Netherlands. Influenza, dengue, and Ebola have all been associated with post-acute sequelae as well. The breadth of this list suggests that post-acute syndromes are not an idiosyncrasy of SARS-CoV-2 but rather a general biological phenomenon whose mechanisms may be shared across pathogens.

Against this backdrop, the question of whether long COVID differs from these older post-infectious syndromes has generated considerable debate. Some studies report that the symptom profile of long COVID overlaps substantially with myalgic encephalomyelitis and chronic fatigue syndrome, with post-exertional malaise, unrefreshing sleep, cognitive difficulties, and orthostatic intolerance dominating the clinical picture in both conditions. Other investigations emphasize features that appear more characteristic of SARS-CoV-2, including prominent dyspnea, loss of smell and taste, and distinctive patterns of lung and cardiac imaging abnormalities. Population-based designs that include non-infected controls are essential for resolving this debate, because many symptoms attributed to long COVID, such as fatigue and headache, are also common in the general population and fluctuate over time for reasons unrelated to infection.

The inclusion of non-infected controls deserves particular emphasis. Early in the pandemic, numerous studies reported alarmingly high prevalence figures for long COVID, sometimes exceeding fifty percent of infected individuals, but many of these studies lacked appropriate comparison groups. When matched controls are introduced, the excess symptom burden attributable to infection typically shrinks, though it rarely disappears entirely. Well-conducted studies generally find that infected individuals report certain symptoms, notably loss of smell, shortness of breath, and post-exertional malaise, at rates meaningfully above control levels, while other symptoms converge with background population prevalence. This pattern implies that long COVID is real but heterogeneous, comprising a core of infection-specific pathology surrounded by a larger halo of nonspecific symptom reporting that would have occurred anyway.

Reinfection adds another layer of complexity. As the pandemic progressed and immunity from vaccination and prior infection accumulated, most populations experienced multiple SARS-CoV-2 exposures. Evidence from several cohorts suggests that the risk of long COVID declines with each subsequent infection, plausibly reflecting milder acute illness, partial immune protection, and the selection of individuals who did not develop post-acute sequelae after their first infection. Nevertheless, reinfections are not risk-free, and a subset of patients reports symptom onset or worsening after a second or third episode. Distinguishing whether such symptoms represent a new post-acute syndrome or a recurrence of unresolved prior symptoms is precisely the analytical challenge that motivates comparisons between new and recurring cases.

Vaccination status is another critical covariate. Multiple observational studies and meta-analyses indicate that vaccination before infection reduces the subsequent risk of long COVID, with estimates of risk reduction generally in the range of thirty to fifty percent. The mechanism is presumed to involve blunting of acute viral replication and modulation of the early immune response, though whether vaccination also influences the course of established long COVID remains contested, with trials of therapeutic vaccination yielding mixed results. Any population-based comparison of post-acute syndromes must therefore account for vaccination timing and dose number, since these factors shifted dramatically across pandemic waves and could confound apparent differences between groups infected at different times.

Methodological considerations extend to the definition and measurement of the syndromes themselves. Unlike myocardial infarction or stroke, post-acute infection syndromes lack a biomarker-based diagnostic test. Case definitions have relied on symptom clusters, with the World Health Organization’s clinical case definition of long COVID requiring symptoms persisting at least two months that cannot be explained by an alternative diagnosis. Operationalizing such definitions in population surveys involves trade-offs between sensitivity and specificity. Broad definitions capture more true cases but also more misclassified individuals whose symptoms stem from other causes; narrow definitions do the reverse. Studies that apply identical symptom instruments to infected and control participants and then model the excess attributable to infection offer the most defensible prevalence estimates.

Statistical approaches such as latent class analysis and clustering have been deployed to identify symptom subgroups within long COVID populations. These analyses consistently suggest heterogeneity, with clusters resembling a cardiopulmonary phenotype dominated by breathlessness and chest symptoms, a neurocognitive phenotype centered on brain fog and headache, and a systemic phenotype featuring fatigue, fever, and post-exertional malaise. Whether these clusters correspond to distinct pathophysiological pathways, such as persistent viral reservoirs versus autoimmune dysregulation versus microvascular dysfunction, is an active area of laboratory investigation. Small studies have reported abnormalities in T-cell exhaustion markers, elevated autoantibodies against G-protein-coupled receptors, reduced capillary density, and evidence of hypometabolism on PET imaging in subsets of patients, but findings await replication in larger, well-controlled samples.

The economic and societal burden of post-acute syndromes compounds their clinical importance. Estimates from several countries suggest that a meaningful fraction of patients with long COVID experience reduced work capacity, with a smaller proportion leaving employment altogether. Health-related quality of life scores among severely affected patients rival those reported in advanced chronic diseases. Health systems have responded by establishing specialized post-COVID clinics, though access remains uneven and the evidence base for specific treatments remains thin. Rehabilitation programs must be designed cautiously, because graded exercise approaches that benefit many deconditioned patients can provoke worsening in those with post-exertional malaise, a phenomenon documented in myalgic encephalomyelitis research long before the pandemic.

Comparing new and recurring cases within a single population-based framework also illuminates the natural history of these conditions. Follow-up studies of long COVID indicate gradual symptomatic improvement for most patients over the first one to two years, though a persistent minority remains substantially impaired. Comparable trajectories have been described for post-Q-fever fatigue syndrome and post-infectious fatigue following mononucleosis, where recovery curves flatten after the first year. Understanding whether recurring cases follow a different trajectory from new cases could inform prognostic counseling and identify subgroups who might benefit from more intensive monitoring or earlier intervention.

Ultimately, the value of a rigorous population-based comparison lies in its capacity to reframe long COVID not as an isolated novelty but as part of a continuum of infection-triggered chronic illness. If new-onset and recurring post-acute syndromes share risk factors, symptom architecture, and trajectories, this convergence would argue for common research infrastructure, shared biobanks, and unified diagnostic frameworks that could accelerate progress across all post-infectious conditions. Conversely, demonstrable differences would sharpen the search for SARS-CoV-2-specific mechanisms and justify tailored therapeutic development. Either outcome advances the field, and either underscores the enduring lesson that acute infections can cast long shadows over the health of populations, shadows that deserve systematic measurement rather than anecdote.

Subject of Research: Comparison of new and recurring post-acute infection syndrome – results from a German population-based cohort

Article Title: Comparison of new and recurring post-acute infection syndrome – results from a German population-based cohort

Article References: Frost, J., Peter, F., Glaser, N., Pfrommer, L. R., Fasshauer, J. M., Opel, N., Gekle, M., Behrens, T., Tüscher, O., & Mikolajczyk, R. (2026). Comparison of new and recurring post-acute infection syndrome – results from a German population-based cohort. Scientific Reports, 16(1), Article 28492. https://doi.org/10.1038/s41598-026-70451-3

Image Credits: AI Generated

DOI: 10.1038/s41598-026-70451-3

Keywords: Comparison, recurring, post-acute, infection, syndrome, results, German, population-based, cohort, scientific research