Polygenic risk scores compress the small effects of hundreds of thousands of common genetic variants into a single number that stratifies people by inherited susceptibility. Their predictive power depends on how well the variant effects estimated in a discovery genome-wide association study transfer to the person being scored. Because allele frequencies, linkage disequilibrium patterns, and effect sizes differ across ancestries, scores trained almost exclusively on European data typically lose a large fraction of their accuracy in East Asian populations, a problem that has limited clinical deployment of genetic risk prediction outside Europe.

The research team tackled this transferability problem using PRS-CSx, a continuous shrinkage framework that jointly models summary statistics from multiple ancestry-specific genome-wide association studies. Rather than forcing a choice between East Asian and European effect estimates, the method borrows information across ancestries through a shared prior, allowing variant effects to be estimated with greater precision while remaining anchored to the population being scored. The authors constructed two configurations. The first, termed PRS-meta, was derived purely from the genome-wide association results and linkage disequilibrium reference panels, requiring no individual-level training data. The second, PRS-combine, was additionally optimized in a Chinese training set of 2,800 coronary artery disease cases and 2,055 controls.

For comparison, the investigators generated East Asian-specific scores using two classical approaches, LDpred and pruning-and-thresholding, and also evaluated previously published coronary artery disease scores, including European-specific ones. The top-performing score for each method in the training set was then carried forward to an external validation of remarkable scale: 72,782 participants with genetic data from the China Kadoorie Biobank, a population-based cohort followed for a mean of 12.1 years for incident coronary events.

The training results were striking. PRS-meta and PRS-combine performed comparably, with odds ratios per standard deviation of 2.21 and 2.23 respectively, meaning each standard deviation increase in score roughly doubled the odds of disease. Both trans-ancestry scores delivered approximately 60 percent increases in pseudo-R squared relative to the best East Asian-specific scores and roughly 90 percent increases relative to the best European-specific scores. In practical terms, the fused scores captured far more of the heritable component of coronary risk in Chinese individuals than any single-ancestry alternative.

Validation in the biobank cohort confirmed that the gains survived outside the training data. Both scores were similarly associated with incident coronary artery disease, with hazard ratios per standard deviation of 1.33 for PRS-meta and 1.31 for PRS-combine. When participants were stratified into quintiles, those in the top quintile faced roughly double the risk of those in the bottom quintile, with hazard ratios of 2.06 for PRS-meta and 1.84 for PRS-combine. The consistency of these associations across more than twelve years of follow-up indicates that the scores measure stable, lifelong genetic susceptibility rather than transient risk factors.

A notable technical finding is that PRS-meta, which requires no individual-level training data at all, matched the performance of PRS-combine, which was tuned on Chinese cases and controls. This matters for global health equity because many populations lack the large, well-phenotyped biobanks needed to train and optimize scores locally. A method that fuses ancestry-specific summary statistics and still achieves top performance without local training data offers a path to equitable genetic risk prediction for underrepresented groups, provided that ancestry-specific genome-wide association studies of sufficient size exist.

The study also situates genetic risk within a broader clinical context. The China Kadoorie Biobank validation allowed the authors to examine how the scores relate to conventional cardiovascular risk factors and to established risk prediction models such as the Prediction for Atherosclerotic cardiovascular disease Risk in China framework. Supplementary analyses covered model calibration, improvements in discrimination and reclassification, and subgroup and sensitivity analyses, all of which support the robustness of the trans-ancestry approach. The scores are intended to complement, not replace, conventional risk assessment, adding an inherited-risk dimension that is present from birth and independent of lifestyle factors measured in adulthood.

Coronary artery disease is shaped by the interplay of hundreds of common variants, each contributing a tiny increment of risk, together with diet, smoking, blood pressure, diabetes, and environment. The trans-ancestry scores developed here do not predict destiny; they stratify inherited susceptibility so that individuals in the upper tail of the distribution can be identified earlier for intensified prevention, from lipid management to lifestyle counseling. Because the scores are computed from standard genotyping arrays and summary statistics, they could in principle be integrated into population screening programs without exotic infrastructure.

The broader significance of the work lies in its generalizable strategy. As genome-wide association studies accumulate for diverse populations, continuous shrinkage methods that share information across ancestries promise risk scores that work everywhere rather than only where the data were collected. The authors highlight PRS-meta as particularly promising for this purpose, and the validation in one of the world’s largest Chinese cohorts provides strong evidence that the approach is ready for wider testing in other East Asian and admixed populations. For a field long criticized for concentrating its benefits in European ancestry populations, the study marks a concrete step toward genomic medicine that serves everyone.

Subject of Research: Development and validation of a trans-ancestry polygenic risk score for coronary artery disease in the Chinese population

Article Title: Development and validation of a trans‑ancestry polygenic risk score for coronary artery disease in the chinese population

Article References: He, J., Yang, S., Cui, Q., Sun, D., Lin, Z., Li, J., Liu, F., Yu, C., Sun, D., Pang, Y., Huang, K., Chen, S., Cao, J., Pei, P., Shen, C., Hu, D., Yang, L., Millwood, I. Y., Liu, X., … Lu, X. (2026). Development and validation of a trans‑ancestry polygenic risk score for coronary artery disease in the chinese population. Genome Medicine. https://doi.org/10.1186/s13073-026-01800-0

Image Credits: AI Generated

DOI: 10.1186/s13073-026-01800-0

Keywords: polygenic risk score, coronary artery disease, trans-ancestry, genomic risk prediction, China Kadoorie Biobank, PRS-CSx, genome-wide association studies, East Asian population, cardiovascular genetics, health equity, risk stratification, Genome Medicine

The gene at the center of the study, POU4F3, sits on chromosome 5 and encodes a transcription factor known as Brn-3c, a protein that acts as a master switch for the survival and function of hair cells, the delicate sensory receptors inside the cochlea that convert sound vibrations into electrical signals for the brain. Because hair cells do not regenerate in humans, any genetic insult that compromises their maintenance translates into progressive, irreversible sensorineural hearing loss. The gene was first linked to inherited deafness through the DFNA15 locus, and since then a growing catalogue of pathogenic variants has been assembled worldwide, though its mutational landscape in Chinese populations had remained incompletely charted until now.

To fill that gap, the team evaluated 83 unrelated families diagnosed with autosomal dominant non-syndromic hearing loss at two tertiary referral centers in China. Each family underwent whole-exome sequencing, a technique that reads the protein-coding regions of the genome and allows researchers to sift through tens of thousands of variants in search of the causal culprit. Candidate variants in POU4F3 were then validated by Sanger sequencing, the classical gold-standard method, and tested for co-segregation, meaning that the variant had to be present in affected relatives and absent in unaffected ones across each pedigree, exactly as an autosomal dominant disease gene would predict.

The search paid off. Nine distinct POU4F3 variants were identified across nine families, and four of them had never been described before: a duplication designated c.149_152dup, two small deletions called c.687_688delCA and c.709_710delTC, and a single-letter substitution, c.704 C greater than T. The remaining five variants, including c.371 C greater than A, c.592 C greater than A, c.706 C greater than T, c.952 G greater than A, and a deletion spanning the end of exon 1 and the start of exon 2, had been reported previously. All variants were classified according to the guidelines of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology, the internationally accepted framework for judging whether a genetic variant is pathogenic.

Perhaps the most striking statistic is the fraction of families explained by this one gene. POU4F3 variants accounted for 10.8 percent of the cohort, or nine of 83 families. The authors compared this with figures from selected European, Japanese, and Korean cohorts, which reported 6.5 percent, 2.5 percent, and 2.4 percent respectively. They were careful, however, to caution that such direct comparisons should be interpreted conservatively, because differences in how patients are recruited, which testing strategies are used, and the nature of the referral settings can all inflate or deflate apparent prevalence. Even so, the message is clear: in Chinese patients with familial progressive hearing loss, POU4F3 deserves a prominent place on any diagnostic gene panel.

Among the 35 confirmed variant carriers in the study, penetrance was complete, meaning every individual carrying a pathogenic variant showed clinical hearing loss at some point. The audiometric profile followed a characteristic pattern: predominantly mid-to-high-frequency sensorineural hearing loss that worsened with age. Cross-sectional analysis across the families revealed a gradient, with younger carriers showing milder, high-frequency impairment, often invisible in everyday conversation, while older individuals had progressed to severe-to-profound broadband loss affecting virtually all frequencies. This age-dependent progression is a hallmark of many dominant deafness genes and underscores why extended high-frequency audiometry, which tests frequencies above the standard clinical range, was an essential tool in the study, catching early damage that conventional pure-tone audiometry would miss.

Tinnitus, the perception of sound without an external source, emerged as a surprisingly informative companion symptom. Sixty percent of affected individuals, 21 of 35, reported subjective tinnitus. More remarkable still, 45 percent of those assessed for timing, nine of 20, developed tinnitus before any measurable hearing loss could be detected on an audiogram, with lead times ranging from zero to 27 years. In practical terms, a young adult from one of these families who begins to experience persistent ringing in the ears may be perceiving the earliest distress signals of dying or malfunctioning cochlear hair cells, decades before a standard hearing test would flag a problem. The authors propose that the triad of familial non-syndromic hearing loss, progressive sloping sensorineural loss, and accompanying tinnitus could serve as a practical clinical clue to trigger earlier genetic evaluation, though they emphasize that this framework requires prospective validation.

The study also explored how different classes of variants shape the clinical picture. Exploratory genotype-phenotype analysis suggested that truncating variants, those that introduce a premature stop signal and chop the Brn-3c protein short, may be associated with earlier onset and more severe hearing loss, consistent with the idea that a nonfunctional protein causes a loss of protective function in hair cells. Missense variants, which substitute a single amino acid, showed greater phenotypic variability, particularly when they fell within the protein’s core DNA-binding domains, the POU-specific and homeobox regions that grip target genes and regulate their expression. The authors integrated their nine variants with 43 previously published ones to build these correlations, and compiled literature and ClinVar data identifying 82 pathogenic or likely pathogenic POU4F3 variants worldwide, of which 59, or 72 percent, were associated with East Asian reports. They note that this apparent geographic concentration may partly reflect ascertainment and reporting practices rather than true population genetics alone.

The clinical implications extend beyond diagnosis. As gene therapy for inherited deafness moves from concept to clinic, with several trials targeting other deafness genes already underway, knowing which gene underlies a patient’s condition becomes the gateway to future precision treatment. Identifying POU4F3 carriers early, ideally through the tinnitus-first presentation described here, would allow families to receive accurate genetic counseling, understand the roughly 50 percent transmission risk to each child of a carrier, plan monitoring schedules with extended high-frequency audiometry, and position themselves for intervention as therapies mature. The study was approved by the Medical Ethics Committee of Chinese PLA General Hospital and funded by China’s National Key R&D Program and the National Natural Science Foundation of China.

For now, the researchers urge that POU4F3 be prioritized within hearing-loss gene panels, particularly for Chinese and East Asian patients presenting with compatible phenotypes. The work also carries a broader lesson for auditory medicine: symptoms that patients often dismiss as a minor nuisance, a faint ringing after a loud concert or a persistent hum in quiet rooms, can be the first audible thread in a genetic story that unfolds over decades. Listening carefully to tinnitus, and pairing it with family history and modern sequencing, may allow clinicians to catch hereditary hearing loss at its very beginning, when the window for preserving hearing, and eventually for repairing it, is widest.

Subject of Research: Prevalence and clinical characteristics of POU4F3 gene variants in Chinese patients with autosomal dominant non-syndromic hearing loss and tinnitus

Article Title: POU4F3 variants are prevalent in Chinese patients with autosomal dominant non-syndromic hearing loss and concurrent tinnitus

Article References: Zhang, Q., Zeng, L., Wang, R., Lu, Y., Wu, X., Balan, V., Zhang, L., Chen, G., Xie, L., Zhang, Y., Guo, F., Ning, F., Wang, D., Han, D., Yuan, H., & Wang, Q. (2026). POU4F3 variants are prevalent in Chinese patients with autosomal dominant non-syndromic hearing loss and concurrent tinnitus. Journal of Translational Medicine. https://doi.org/10.1186/s12967-026-08915-3

Image Credits: AI Generated

DOI: 10.1186/s12967-026-08915-3

Keywords: POU4F3, DFNA15, hearing loss, tinnitus, autosomal dominant, genetics, whole-exome sequencing, sensorineural hearing loss, genotype-phenotype correlation, East Asian population, genetic counseling, cochlear hair cells

The drug at the center of the study, roxithromycin, is a macrolide antibiotic related to erythromycin, used against respiratory tract and other bacterial infections. Like many macrolides, it is given orally and its absorption into the bloodstream depends on a chain of events that includes dissolution of the tablet in the gastrointestinal tract, permeation across the intestinal wall, and distribution into tissues. Any change in the formulation, the manufacturing process, or even the excipients used to bind and coat the tablet can in principle alter how quickly and completely the drug reaches the circulation. That is why regulators require bioequivalence testing whenever a generic version or a modified formulation is brought to market. The standard metrics are the area under the concentration-time curve, which reflects total drug exposure, the maximum plasma concentration, which reflects peak exposure, and the time at which that maximum occurs. If the ratios of these parameters between test and reference products fall within a ninety percent confidence interval of eighty to one hundred twenty-five percent, the products are considered bioequivalent.

Conducting such trials in humans is expensive, slow, and ethically burdensome, since volunteers are exposed to a drug without any personal therapeutic benefit. This is where physiologically based pharmacokinetic modeling, or PBPK, enters the picture. Unlike simpler empirical models that treat the body as a set of mathematical compartments with fitted parameters, a PBPK model attempts to represent the actual anatomy and physiology of the body: organs and tissues are connected by blood flows, and drug movement between them is governed by physicochemical properties such as solubility, permeability, lipophilicity, and protein binding, together with physiological parameters like organ volumes, blood flow rates, and enzyme abundances. Because the underlying structure reflects real biology, a well-constructed PBPK model can be adapted to different populations, doses, and even formulations simply by swapping in the relevant inputs, making it a powerful tool for predicting drug behavior in scenarios that have never been tested clinically.

The research team, led by Danfeng Yu, Qin Li, and colleagues including corresponding authors Jiyu Chen and Yan He, assembled roxithromycin’s physicochemical, in vitro, and in vivo parameters from established databases and the published literature, and constructed an adult PBPK model using the open-source software platform PK-Sim. The initial model was developed and calibrated against clinical pharmacokinetic data from a 150 milligram dose of the reference formulation. The critical test of any model, however, is not how well it fits the data used to build it, but how well it predicts data it has never seen. To address this, the researchers performed a comparative evaluation using clinical data from the test formulation, checking whether the model could reproduce concentration-time profiles for a product whose behavior had not informed the model’s construction.

To quantify the agreement between prediction and observation, the team applied the mean fold error approach, a standard metric in pharmacokinetic modeling that measures the ratio of predicted to observed values. For the three key parameters, the area under the curve from zero to the last measurable time point, the maximum concentration, and the time to maximum concentration, more than ninety percent of the mean fold error ratios fell within the range of 0.5 to 2. This means that for the overwhelming majority of comparisons, the model’s predictions were within a factor of two of the measured values, and in most cases considerably closer. With the model thus validated, the researchers imported formulation-specific in vitro dissolution data for both products, allowing the simulations to distinguish between the two tablets based on how rapidly each released its drug payload in the digestive tract.

The virtual bioequivalence trial itself was conducted in a simulated cohort of ninety-nine East Asian subjects, a population chosen to match the demographic context in which the formulations are intended to be used. Virtual subjects in PBPK platforms are not identical clones; the software samples from realistic distributions of physiological parameters such as body weight, gastric emptying time, and intestinal transit, so that the simulated population exhibits the same variability that a real trial population would. Each virtual subject received both formulations, and the resulting pharmacokinetic profiles were analyzed exactly as they would be in a conventional bioequivalence study. The outcome was unambiguous: the ninety percent confidence intervals for the geometric mean ratios of the primary pharmacokinetic parameters all fell within the regulatory acceptance range of eighty to one hundred twenty-five percent, indicating that the two formulations would be judged bioequivalent.

The implications of this result extend well beyond a single antibiotic. Bioequivalence studies are among the most numerous clinical trials conducted worldwide, required for every generic approval and for many post-approval manufacturing changes. If validated PBPK models can reliably substitute for some of these trials, the savings in cost, time, and human exposure could be substantial. Regulators, including the United States Food and Drug Administration and the European Medicines Agency, have already issued guidance on when modeling and simulation can support or replace certain clinical studies, and virtual bioequivalence is an active area of regulatory science. The present study adds to a growing body of evidence that, for drugs with well-characterized pharmacokinetics and formulations whose dissolution behavior can be measured reliably in vitro, a computer can indeed stand in for the clinic.

The study also illustrates the methodological discipline that makes such virtual trials credible. The researchers did not simply assume the model was correct; they built it from one dataset, tested it against an independent dataset from a different formulation, and reported a formal quantitative measure of predictive performance before trusting it for the bioequivalence question. The work was registered retrospectively at ClinicalTrials.gov under number NCT06798051, the underlying clinical study was approved by the Medical Ethics Committee of the Affiliated Hospital of Guizhou Medical University, and all participants in the source trials provided informed consent. The modeling was funded by the Science and Technology Planning Project of Guizhou Province and related institutional programs, and the authors declared no competing interests.

Certain caveats remain. A virtual trial is only as good as its underlying model, and PBPK predictions for drugs with complex metabolism, food effects, or narrow therapeutic windows demand even more extensive validation than was possible here. The ninety-nine subject cohort, while adequate for this comparison, is a simulation of a specific population rather than a demonstration across all patient groups. Nevertheless, the study offers a concrete template for how virtual bioequivalence can be conducted and documented: assemble the drug’s parameters, build and validate the model against real data, import formulation-specific dissolution profiles, and simulate a demographically realistic cohort under the same statistical criteria that govern human studies. As more models of this kind accumulate and are independently verified, the day when a generic drug can win approval on the strength of a computer trial, with human studies reserved for cases the models cannot yet handle, moves measurably closer.

Subject of Research: Virtual bioequivalence assessment of two roxithromycin formulations using a physiologically based pharmacokinetic model

Article Title: Virtual bioequivalence assessment for two roxithromycin formulations using a physiologically based pharmacokinetic model

Article References: Yu, D., Li, Q., Zeng, L., Liao, M., Lu, X., Chen, J., & He, Y. (2026). Virtual bioequivalence assessment for two roxithromycin formulations using a physiologically based pharmacokinetic model. BMC Pharmacology and Toxicology. https://doi.org/10.1186/s40360-026-01244-7

Image Credits: AI Generated

DOI: 10.1186/s40360-026-01244-7

Keywords: roxithromycin, PBPK modeling, virtual bioequivalence, PK-Sim, pharmacokinetics, generic drugs, drug dissolution, macrolide antibiotics, clinical trial simulation, regulatory science, East Asian population, BMC Pharmacology and Toxicology