The research, led by Oğuzhan Türksayar of Başakşehir Çam and Sakura City Hospital in Istanbul and colleagues at institutions across Türkiye, focused on 17 brain metastases treated in 12 consecutive patients who received single-session Gamma Knife radiosurgery. Each patient underwent dynamic contrast-enhanced MRI before treatment and again at the four-to-eight-week mark. Rather than simply observing whether lesions shrank, the team used pharmacokinetic modeling to quantify how the blood-brain barrier and tumor vasculature were behaving, extracting parameters that reflect the movement of contrast agent between blood vessels and tumor tissue.

The technical core of the study lies in the Tofts pharmacokinetic model, a mathematical framework that describes how gadolinium-based contrast material leaks from blood plasma into the extravascular extracellular space of tissue. From this model, the researchers derived several quantitative parameters. The volume transfer constant, known as Ktrans, measures how quickly contrast crosses from the vascular compartment into tissue, serving as a proxy for vascular permeability and blood flow. Kep, the rate constant, describes how rapidly contrast returns from tissue back to the plasma, while Ve represents the fraction of tissue volume occupied by the extravascular extracellular space. The team also measured the initial area under the concentration-time curve, or iAUC, a semiquantitative indicator of early contrast enhancement that captures the first blush of agent entering the lesion.

In addition to these model-based parameters, the researchers extracted semiquantitative features directly from the signal-intensity curves generated during the dynamic scan, including the upslope and the initial gradient of enhancement. These metrics do not require pharmacokinetic assumptions, making them quicker to compute, but they are also more sensitive to variations in scanning protocol. By combining both approaches, the study aimed to determine which measurements best captured the biological consequences of radiation in the earliest phase of treatment.

The results showed a consistent pattern. After radiosurgery, Ktrans, Kep, and iAUC all decreased significantly across the treated lesions, and these reductions remained statistically meaningful even after the researchers applied Benjamini-Hochberg false discovery rate correction, a conservative statistical adjustment designed to guard against false positives when many comparisons are made. The initial gradient also decreased in unadjusted analysis, but this finding did not survive correction. In practical terms, the tumors’ permeability-surface characteristics declined measurably within weeks of treatment, suggesting that radiosurgery was already altering the abnormal vasculature that sustains metastatic growth.

Response classification followed the RANO-BM criteria, the standardized framework developed by the Response Assessment in Neuro-Oncology working group for evaluating brain metastases. Of the 17 lesions, six, representing 35.3 percent, were classified as partial response, while eleven, or 64.7 percent, showed stable disease. Notably, no lesion progressed during this early window, and none achieved complete response, which is typical at such an early checkpoint since tumor shrinkage lags behind vascular change.

The most provocative finding emerged when the researchers examined whether the magnitude of imaging change tracked with the degree of response. In unadjusted analyses, the percentage change in Ktrans was associated with concurrent early radiologic response, with a Spearman correlation coefficient of negative 0.553, and the percentage change in upslope showed a similar association with a coefficient of negative 0.503. Lesions classified as partial response also showed a greater absolute reduction in Ktrans than stable-disease lesions before correction for multiple comparisons. However, the authors are careful to note that none of these between-group or correlation findings survived false discovery rate correction, a caveat that colors the entire study’s conclusions.

This statistical honesty is a defining feature of the work. The researchers label their findings as hypothesis-generating and explicitly state that they require validation in larger, adequately powered, and methodologically standardized cohorts before any clinical application. Recognizing that their small sample involved multiple lesions per patient, which can inflate statistical confidence, they performed a sensitivity analysis using only one lesion per patient. The direction of the Ktrans difference remained consistent with the lesion-level analysis, offering some reassurance that the signal, while not statistically robust in a sample of this size, was not an artifact of clustering within individual patients.

The implications, if confirmed, could be substantial. Early identification of nonresponding lesions would allow oncologists to adjust treatment sooner, whether by adding systemic therapy, planning repeat radiosurgery, or intensifying surveillance, rather than waiting months for conventional scans to reveal failure. Dynamic contrast-enhanced MRI is already available on most modern clinical scanners and can be added to routine follow-up protocols, though its pharmacokinetic modeling demands careful attention to acquisition parameters. The prospective design of this study, with baseline and early post-treatment imaging performed under controlled conditions, strengthens its foundation compared with retrospective analyses, and the lesion-level approach provides fine-grained detail about how individual metastases behave rather than averaging across patients.

For now, the study stands as an early but carefully executed proof of concept. Brain metastases affect a large proportion of patients with advanced cancer, and stereotactic radiosurgery is one of the most common treatments delivered to them, making any tool that accelerates response assessment clinically valuable. The Turkish team’s work demonstrates that measurable vascular changes appear within four to eight weeks of Gamma Knife treatment and that Ktrans, the volume transfer constant, was the most consistent parameter associated with early radiologic response in unadjusted analyses. The road from exploratory association to clinical decision-making runs through larger, standardized, multi-center trials, but this study charts a plausible path toward a future in which the effectiveness of radiation against brain tumors can be read not just in their size, but in the chemistry of their blood vessels.

Subject of Research: Early dynamic contrast-enhanced MRI biomarkers of radiologic response after stereotactic radiosurgery for brain metastases

Article Title: Early DCE-MRI changes and concurrent radiologic response after stereotactic radiosurgery for brain metastases: a prospective exploratory study

Article References: Türksayar, O., Öner, A. Y., Cindil, E., Erpolat, P., & Emmez, H. (2026). Early DCE-MRI changes and concurrent radiologic response after stereotactic radiosurgery for brain metastases: a prospective exploratory study. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02756-8

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02756-8

Keywords: DCE-MRI, stereotactic radiosurgery, brain metastases, Ktrans, Tofts pharmacokinetic model, Gamma Knife, RANO-BM criteria, treatment response assessment, blood-brain barrier, MRI biomarkers, false discovery rate, radiation oncology

The research, published in BMC Endocrine Disorders, drew on the China Health and Retirement Longitudinal Study, a nationally representative cohort that has tracked thousands of middle-aged and older Chinese adults since 2011. The team, led by Kan Chen and Yi Wang of the First Affiliated Hospital of Jinzhou Medical University, identified 5,871 participants who were free of knee osteoarthritis at the start of the study. These individuals were followed across four subsequent survey waves in 2013, 2015, 2018, and 2020, giving the investigators up to nine years of observation. During that window, 2,143 participants, or 36.5 percent of the cohort, developed symptomatic knee osteoarthritis as reported on standardized questionnaires. That strikingly high incidence rate alone underscores why the condition represents such an enormous public health burden in rapidly aging populations.

What made this study distinctive was its systematic approach to measuring insulin resistance. Because directly measuring insulin sensitivity with a glucose clamp is impractical in large population studies, researchers rely on surrogate indices calculated from routine blood tests and anthropometric measurements. The team evaluated six of them simultaneously. The estimated glucose disposal rate, or eGDR, models how quickly the body can clear glucose from the bloodstream. The triglyceride-glucose index, or TyG, combines fasting triglycerides and glucose into a single number that tracks insulin resistance. Variants of TyG that incorporate body mass index or waist circumference capture both metabolic and adiposity dimensions. The Chinese visceral adiposity index, or CVAI, was specifically developed to estimate harmful visceral fat in Chinese populations. The metabolic score for insulin resistance, or METS-IR, and the atherogenic index of plasma, or AIP, round out the panel. Each index offers a slightly different window into the same underlying metabolic state, and testing all six together guards against the cherry-picking that has plagued smaller studies in this field.

The statistical machinery was equally thorough. The investigators used multivariable Cox regression, the standard tool for survival analysis, to estimate hazard ratios for knee osteoarthritis across quartiles of each index while adjusting for a battery of potential confounders. They then applied restricted cubic splines, a flexible modeling technique that allows the relationship between an index and disease risk to take on any shape, including U-shaped or threshold effects, rather than assuming a straight line. Sensitivity analyses probed whether early incident cases, overlapping information about body size in the adiposity-based indices, or selection bias might be distorting the results. And crucially, the team corrected for multiple testing using the false discovery rate procedure, a step that many observational studies skip and that often separates robust signals from statistical noise.

Before that correction, the raw results looked tantalizing. The trend tests for eGDR and CVAI showed nominal P values of 0.046 and 0.044 respectively, just under the conventional 0.05 threshold. The atherogenic index of plasma told a slightly different story: participants in the second and third quartiles of AIP had nominally elevated hazard ratios of 1.14 and 1.15 compared with the lowest quartile, with confidence intervals that excluded the null value of one. Interestingly, the highest AIP quartile did not show a significant elevation, and the overall trend test for AIP was far from significant at 0.798, a pattern that would be difficult to reconcile with a simple dose-response relationship between lipid-derived insulin resistance and joint degeneration.

Then came the false discovery rate correction, and the picture changed. When the six fully adjusted trend tests were corrected for the fact that six independent hypotheses were being tested, none of them remained statistically significant. The quartile comparisons for AIP, which were not included in the multiplicity correction, are reported by the authors as nominal, exploratory findings. The restricted cubic spline analyses found no evidence of statistically significant non-linearity for any of the six indices. In plain terms, once the analysis accounted for the reality that testing six markers makes it likely that at least one will cross the 0.05 threshold by chance alone, the apparent signals vanished into the statistical background.

The authors are refreshingly candid about what this means. They describe the observed patterns as small in magnitude and explicitly state that they should be regarded as hypothesis-generating rather than as evidence that any of these indices can stratify knee osteoarthritis risk in clinical practice. This kind of restraint is worth pausing on, because it runs counter to a well-documented tendency in metabolic and musculoskeletal research, where surrogate indices are frequently promoted as screening tools on the basis of uncorrected associations. A nominal P value of 0.046, in a field where dozens of similar analyses have been published, is exactly the kind of finding that the false discovery rate framework was designed to temper.

The study does have limitations that the authors and independent readers will want to weigh. The outcome was self-reported symptomatic knee osteoarthritis rather than radiographically confirmed disease, which means the findings speak most directly to clinically apparent disease but may miss asymptomatic structural damage. Residual confounding remains possible in any observational cohort, and the sensitivity analyses, while consistent, showed attenuation of estimates after full adjustment, suggesting that some of the raw association may have been driven by factors such as body weight that lie on the causal pathway or correlate closely with it. The adiposity-based indices in particular share substantial mathematical overlap with body mass index, making it difficult to disentangle a specific insulin resistance effect from the well-established mechanical load of excess weight on knee joints.

Yet the null result is itself informative. If insulin resistance were a strong independent driver of knee osteoarthritis, one might expect at least one of six complementary indices, each capturing a different facet of metabolic dysfunction, to survive correction in a cohort of nearly six thousand people followed for nine years. The absence of such a signal suggests that whatever contribution metabolic dysfunction makes to knee osteoarthritis is likely modest, indirect, or mediated through pathways such as adiposity and systemic inflammation that are already captured by conventional risk factors. It also cautions against the growing enthusiasm for repurposing metabolic indices as universal predictors of diseases well beyond their original scope.

For the millions of people living with creaking, aching knees, the practical takeaway is unchanged: maintain a healthy weight, stay physically active, protect the joints from injury, and manage metabolic health for its own well-established benefits. For researchers, the study offers a methodological template worth emulating, one that pre-specifies multiple indices, applies rigorous multiplicity correction, and reports null findings with the same transparency as positive ones. In an era when flashy associations often travel faster than careful verification, a nine-year cohort study that concludes its own hypothesis is not yet supported is a quiet but valuable contribution to the science of aging joints.

Subject of Research: Prospective associations between surrogate insulin resistance indices and incident symptomatic knee osteoarthritis in a Chinese cohort

Article Title: Associations between insulin resistance surrogate indices and risk of incident symptomatic knee osteoarthritis: insights from a 9-year Chinese cohort

Article References: Chen, K., Wang, X., Zheng, H., Zhang, X., Wu, J., Huang, J., & Wang, Y. (2026). Associations between insulin resistance surrogate indices and risk of incident symptomatic knee osteoarthritis: insights from a 9-year Chinese cohort. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02526-1

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02526-1

Keywords: insulin resistance, knee osteoarthritis, TyG index, eGDR, CVAI, CHARLS, prospective cohort study, false discovery rate, Cox regression, metabolic syndrome, joint health, epidemiology

The research, published as an open-access brief report in Discover Plants, was led by José Carlos Páez-Franco and Manuel Gutiérrez-Aguilar of the Universidad Nacional Autónoma de México together with Hilda Sánchez-Vidal and Imelda Cecilia Zarzoza-Mendoza of the Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán. Rather than testing whether one strain tolerates heat better than the other, the team set out to determine whether the two Columbia derivatives show conserved or divergent metabolic responses under normal growth conditions and after an acute heat shock. The distinction matters, because metabolomic baselines can differ between accessions and quietly shape how stress experiments are interpreted.

The experimental design was deliberately straightforward. Seeds of Col-0 and Col-4 were obtained from the Arabidopsis Biological Resource Center at Ohio State University, sterilized, and germinated on half-strength Murashige and Skoog medium before transfer to soil. Plants were grown at 22 degrees Celsius under a 16-hour light, 8-hour dark photoperiod. When the plants reached four weeks of age, half of them were subjected to an acute heat-stress regime of 45 minutes at 42 degrees Celsius, while the control plants remained at 22 degrees under the same lighting conditions. Leaves were then harvested for untargeted metabolomic profiling by gas chromatography coupled to mass spectrometry, a technique that separates volatile derivatives of small molecules and identifies them by their mass fragmentation patterns.

The analytical workflow was built around rigorous quality control. Roughly 0.15 grams of fresh leaf tissue from each sample was homogenized in cold methanol and water, spiked with two internal standards—tridecanoic acid and 5α-cholestane—and dried before chemical derivatization with methoximation and trimethylsilylation to make the metabolites amenable to gas chromatography. A quality control sample was injected every four runs to track instrument stability, and only metabolites showing less than 30 percent relative standard deviation across those QC injections were retained. The relative standard deviations for the two internal standards came in at 8.3 and 7.6 percent, indicating a stable analytical platform. Data were processed in MZmine, and putative annotations were assigned by comparison with the NIST spectral library, with the authors careful to note that these assignments, made without authentic standards or retention-index measurements, should not be treated as definitive metabolite identifications.

From 23 biological samples—six Col-0 controls, six heat-stressed Col-0 plants, six Col-4 controls, and five heat-stressed Col-4 plants—the profiling pipeline detected 54 metabolic features. Under control conditions, the two accessions proved strikingly similar. Only a small set of features met the nominal statistical thresholds of a P value below 0.05 and an absolute log2 fold change greater than 1 in the direct comparison between Col-0 and Col-4. Among these, features putatively annotated as maleic acid and a sinapic acid derivative were more abundant in Col-4, with log2 fold changes of +1.40 and +1.19 respectively, while a feature putatively annotated as phytol, a chlorophyll-derived metabolite, was lower in Col-4 with a log2 fold change of −1.11 and a nominal P value of 0.0053. In other words, the baseline metabolic fingerprints of the two Columbia strains overlap almost completely.

Heat stress told a more interesting story. When the researchers compared each accession against its own unstressed control, they found partly overlapping sets of features meeting the nominal criteria. Some features, including putative maleic acid, Arabino-Hexos-2-ulose, and a sinapic acid derivative, met the criteria only in Col-0, while others—putatively threonine, a second threonic acid feature, ribitol, and psicose—met the criteria only in Col-4. Crucially, the authors emphasize that membership in only one within-accession list does not demonstrate that the two accessions respond differently to heat. Ten features, however, met the nominal criteria in both accessions and shifted in the same direction: succinic acid, serine, 3,4-dihydroxytetrahydro-2-furanone, 4-ketoglucose, an aspartic acid feature, threonic acid, glutamic acid, indole-3-acetonitrile, fructose, and galactinol. This shared panel spans organic acids, amino acids, sugars, and phenylpropanoid-related compounds, sketching a common metabolic signature of acute heat response in the Columbia background.

Principal component analysis reinforced that picture of shared response. In this descriptive visualization, the first principal component explained 32.2 percent of total variance and the second explained 13.4 percent, and control and heat-stressed samples separated primarily along the first component. That pattern identifies heat treatment, not genotype, as the dominant source of metabolic variation in the dataset. The heat-stressed Col-0 and Col-4 groups showed partially overlapping distributions in the PCA space, consistent with substantial overlap in their global metabolic profiles, though the authors caution that a descriptive PCA cannot by itself establish statistically significant genotype-dependent differences.

The decisive test came from a two-way factorial model fitted to each metabolite, with genotype, treatment, and their interaction as fixed factors. Because the design was slightly unbalanced—five rather than six stressed Col-4 plants—the team used Type II sums of squares and heteroscedasticity-consistent HC3 covariance estimates for inference, then applied Benjamini–Hochberg correction separately to the genotype, treatment, and interaction P values across all 54 metabolites. Two features, putatively annotated as phytol and a sinapic acid derivative, showed nominal genotype-by-treatment interaction effects below 0.05, but neither survived correction for multiple testing. The authors also acknowledge that the modest sample sizes may have limited statistical power to detect subtle interaction effects, so the null result should be read with appropriate caution rather than as proof of perfect equivalence.

Why does this matter beyond the Arabidopsis community? The 1001 Genomes Project and related surveys have revealed extensive worldwide genetic diversity in the species, and previous work has documented accession-specific proteomic signatures as well as differences in photosynthetic organization and pigment composition. Closely related laboratory strains are sometimes treated as interchangeable controls, yet the new data suggest that even two derivatives of the same Columbia lineage can show small, accession-associated differences in specific metabolic features—here involving an organic acid, a phenylpropanoid-related intermediate, and a chlorophyll-derived metabolite—under identical growth conditions. Those differences were nominal and unconfirmed by authentic standards, but they serve as a reminder that genetic background can leave measurable traces in molecular datasets even when plants look identical on the bench.

The practical takeaway for researchers is twofold. First, the broadly shared heat-associated metabolic patterns in Col-0 and Col-4 provide a concise comparative reference, suggesting that results obtained in one Columbia derivative are likely to translate reasonably well to the other for acute heat-stress metabolomics. Second, the study models a level of statistical transparency that metabolomics as a field increasingly demands: nominal thresholds for descriptive heatmaps, false-discovery-rate correction for inferential claims, QC injections excluded from biological statistics, and public deposition of raw data, metadata, and analysis code on Zenodo. As heat waves intensify and crop resilience becomes a global research priority, knowing exactly which wild type sits behind every data point—and how far its metabolic behavior can be generalized—is no longer a technical footnote. It is part of the result.

Subject of Research: Comparative untargeted metabolomics of the Arabidopsis thaliana accessions Col-0 and Col-4 under control and acute heat stress conditions

Article Title: Accession-associated metabolomic variation in wild type Arabidopsis thaliana under control and heat shock conditions

Article References: Páez-Franco, J. C., Sánchez-Vidal, H., Zarzoza-Mendoza, I. C., & Gutiérrez-Aguilar, M. (2026). Accession-associated metabolomic variation in wild type Arabidopsis thaliana under control and heat shock conditions. Discover Plants, 3(1), Article 448. https://doi.org/10.1007/s44372-026-00931-3

Image Credits: AI Generated

DOI: 10.1007/s44372-026-00931-3

Keywords: Arabidopsis thaliana, metabolomics, heat stress, Col-0, Col-4, GC-MS, natural variation, genetic background, principal component analysis, false discovery rate, plant physiology, stress response

The core problem the researchers address is known as support recovery. In compressive sensing, the support of a signal is the set of indices corresponding to its nonzero entries. Identifying that support correctly is often more important than the raw reconstruction itself, because downstream tasks, whether classification, prediction or anomaly detection, typically depend on knowing which features carry information. Classical ℓ1 minimization encourages sparsity by penalizing the sum of absolute values of the coefficients, but the penalty does not distinguish between a genuine signal component and a spurious one that happens to correlate with the measurements. When the measurement matrix has correlated columns, a situation that is common in practice, these methods can confidently select variables that have no true relationship to the underlying signal, with no built-in mechanism to flag or limit such errors.

KnockoffCS imports a remedy from a different corner of statistics: the knockoff filter, introduced by Rina Foygel Barber and Emmanuel Candès in 2015 and later extended through the model-X framework. The knockoff trick is conceptually striking. For every real variable in a problem, the method constructs a decoy, a synthetic variable engineered to mimic the correlation structure of the original as closely as possible while being provably unrelated to the response being predicted. The statistician then runs a competition: each genuine variable must prove that it explains the data better than its own knockoff does. Because the decoys are constructed to behave like null variables, the procedure yields a mathematically guaranteed bound on the false discovery rate, the expected proportion of selected variables that are truly irrelevant. This guarantee, rooted in the multiple-testing literature going back to Benjamini and Hochberg’s 1995 paper, is what ℓ1 methods have never been able to provide directly.

Integrating that machinery into compressive sensing is not straightforward, and the architecture of KnockoffCS reflects the difficulty. The framework separates the task into two stages. In the first, a knockoff-guided selection mechanism examines the compressed measurements and decides which signal components belong to the support, using the knockoff competition to control the false discovery rate at a user-chosen level. In the second, a dedicated estimation stage reconstructs the actual values of the signal, restricted to the components that survived selection. This division of labor matters because the objectives of selection and estimation pull in different directions: a selection procedure optimized to exclude false positives may be conservative, while an estimator given a verified support set can focus its capacity on accurate coefficient recovery rather than on simultaneously guessing which coefficients matter.

The theoretical analysis behind the paper is where the approach earns its keep. Zhang and Xiong show that enforcing false discovery rate control allows KnockoffCS to reliably identify the correct support and reconstruct the signal under conditions weaker than those typically demanded by ℓ1-based methods. Classical guarantees for LASSO-style recovery generally require restrictive assumptions on the measurement matrix, such as restricted isometry or irrepresentable conditions, which essentially demand that the columns of the measurement matrix be sufficiently uncorrelated. By routing the selection decision through knockoffs, the new framework relaxes these requirements while preserving high accuracy, meaning it can succeed in regimes where conventional compressed sensing would either fail silently or produce unquantifiable errors. The guarantee is not merely asymptotic hand-waving; it is a finite-sample statistical certificate attached to every selection the algorithm makes.

The empirical evidence is substantial. In synthetic simulations, where the true support is known and recovery can be scored exactly, KnockoffCS improved support recovery by up to 3.9 times in F1 score compared with existing baselines, while also reducing reconstruction error. That magnitude of improvement is unusual in a field where incremental gains of a few percentage points are the norm, and it reflects the fact that the baselines were never designed to control false discoveries in the first place. When the ground truth is a sparse signal buried in correlated measurements, a method that systematically limits false positives has an obvious structural advantage over one that does not.

Real-world performance was tested on an unusually broad portfolio of eight public datasets spanning medical imaging, astrophysics, music analysis, materials science, web search and particle physics. The collection includes facial and oral temperature data from PhysioNet, NASA’s Mars asteroid observations, the Million Song Dataset, a hybrid nanofluid density prediction set, the HIGGS particle physics benchmark from the UCI repository, NASA’s Kepler exoplanet data, the MAGIC Gamma Telescope measurements, and the Yahoo! Learning to Rank Challenge corpus. Across these datasets, KnockoffCS surpassed existing compressive sensing methods in 71.4 percent of the evaluated model-dataset combinations and ranked in the top two among the compared methods, with gains relative to the compressed baseline varying by dataset and predictor. The authors are candid that the advantage is not universal: on certain datasets, compressed signal baselines remained competitive, a reminder that statistical guarantees come with trade-offs in power and that no single method dominates every regime.

The significance of the work extends beyond the specific numbers. Compressive sensing increasingly operates inside machine learning pipelines, where compressed features feed into predictors whose reliability matters, whether in medical diagnostics, exoplanet detection or high-energy physics. In such settings, an uncontrolled rate of false discoveries in the support can propagate silently into downstream decisions. By embedding a false discovery rate guarantee directly into the sensing pipeline, KnockoffCS offers what earlier work on FDR-controlled variable selection, including split knockoffs, deep knockoffs and e-value-based derandomized knockoffs, had achieved for regression but not for compressed measurement. The framework effectively turns a signal-processing tool into a statistical machine learning method with interpretable error control, which is exactly the direction the field has been pushing as compressed representations meet high-stakes prediction.

There are, of course, caveats worth keeping in view. Knockoff construction itself depends on being able to generate decoys with the right correlation structure, and the quality of that construction shapes both the guarantee and the power of the procedure. The two-stage design also introduces its own hyperparameters, and the paper’s own results show dataset-dependent variability, so practitioners should expect to validate the method on their specific measurement geometry rather than assume uniform superiority. The authors have released code through a public repository to support reproduction, and the datasets used are all publicly available, which lowers the barrier for independent verification.

Still, the conceptual payoff is clear and likely to resonate. For nearly two decades, compressive sensing has traded measurement cost for reconstruction accuracy under assumptions that practitioners could rarely verify. KnockoffCS reframes the problem: instead of asking whether the measurement matrix satisfies a stringent geometric condition, it asks whether each selected component can beat its own statistically calibrated decoy, and it quantifies exactly how often that competition can be wrong. The result is a compressive sensing framework that tells users not only what the signal probably looks like, but how much confidence they are entitled to place in each piece of the answer. As compressed measurements continue to spread through scientific and industrial machine learning, that kind of built-in honesty may prove to be the most valuable signal of all.

Subject of Research: False discovery rate-controlled support recovery in compressive sensing using statistical machine learning

Article Title: Knockoff-Guided Compressive Sensing: A Statistical Machine Learning Framework for Support-Assured Signal Recovery

Article References: Zhang, X., & Xiong, H. (2026). Knockoff-Guided Compressive Sensing: A Statistical Machine Learning Framework for Support-Assured Signal Recovery. Machine Learning, 115(10), Article 236. https://doi.org/10.1007/s10994-026-07167-y

Image Credits: AI Generated

DOI: 10.1007/s10994-026-07167-y

Keywords: compressive sensing, knockoffs, false discovery rate, support recovery, LASSO, sparse signal reconstruction, statistical machine learning, variable selection, signal processing, HIGGS dataset, machine learning, statistical guarantees

The research, conducted by Berkay Çağdaş of the Department of Nuclear Medicine at Afyonkarahisar State Hospital and Ege University’s Institute of Nuclear Sciences, and Elif Kardelen Çağdaş of the hospital’s Department of Pathology and Ankara University’s Graduate School of Biotechnology, set out to determine whether regional differences in immune and stromal composition inside breast tumors correspond to measurable differences in how the tumor consumes glucose on PET imaging. The premise is biologically plausible: tumors infiltrated by lymphocytes often behave differently from immune-cold tumors, and fibroblast-like stromal cells can alter metabolism, perfusion, and tissue architecture in ways that might plausibly change fluorodeoxyglucose uptake.

The study enrolled 61 women with invasive breast carcinoma who underwent baseline 18F-FDG PET/CT before treatment. Rather than treating each tumor as a single homogeneous blob, the team took a deliberately regional approach. A pathologist, blinded to all PET findings, annotated ten separate intratumoral regions on whole-slide histopathology images, each at least one square millimeter in area. QuPath, an open-access machine-learning platform for digital pathology, then quantified two key features within each region: the density of stromal tumor-infiltrating lymphocytes, the immune cells that gather in the connective tissue surrounding cancer cells, and the density of fibroblast-like stromal cells, spindle-shaped cells identified morphologically on routine hematoxylin and eosin stains.

From those ten regions, the investigators retained the ones with the numerically highest and lowest lymphocyte density for paired analysis, an approach designed to capture the observed range of immune-stromal composition within each tumor. The authors are careful to note that these paired regions represent the extremes of what was measured, not a comprehensive spatial map of every tumor. On the imaging side, the team extracted a rich panel of 30 PET features from each patient’s scan. These included the familiar intensity measures such as maximum, mean, and peak standardized uptake values, volumetric measures including metabolic tumor volume and total lesion glycolysis, texture descriptors derived from gray-level co-occurrence, run-length, and size-zone matrices, and dedicated heterogeneity indices, including the HI2 and HI3 indices and a slope calculated across metabolic volume thresholds from 30 to 50 percent of maximum uptake.

The statistical framework was deliberately conservative. The team ran Spearman correlations between nine immune-stromal endpoints and all 30 PET features, generating 270 individual tests, and then applied Benjamini-Hochberg correction to control the false discovery rate, a standard safeguard against being fooled by chance when many comparisons are made simultaneously. The results were sobering but informative. Thirty-eight of the 270 correlations reached nominal significance at the conventional p-value threshold of 0.05, yet not a single one survived the false discovery rate correction, whether the correction was applied within each endpoint family or across the entire study. The smallest corrected q-value within an endpoint was 0.077, and study-wide it was 0.33, both well above any threshold that would conventionally be called significant.

The strongest raw association was, by the authors’ own framing, weak. The density of stromal tumor-infiltrating lymphocytes in the TIL-rich region correlated with the HI2 heterogeneity index at a Spearman coefficient of 0.379 in absolute value, with a 95 percent confidence interval spanning 0.126 to 0.586 and a nominal p-value of 0.0026. Beyond that single headline number, a coherent pattern emerged in the direction of the nominal findings. Lymphocyte density correlated inversely with metabolic tumor volume, with coefficients ranging from negative 0.29 to negative 0.35, and positively with the HI2 heterogeneity measure. The ratio of lymphocytes to fibroblast-like stromal cells correlated with SUV-based parameters, the largest being a coefficient of 0.299 with mean standardized uptake value. Fibroblast density on its own, and the metrics describing regional differences between the paired areas, produced weaker and less consistent signals.

One of the study’s most technically rigorous contributions is its validation of the pathology measurement itself. Because fibroblast-like stromal cells were defined morphologically on routine H&E stains, the authors cross-checked their machine-learning quantification against alpha-smooth muscle actin immunohistochemistry, a stain that highlights activated fibroblasts and myofibroblasts, in the same 122 paired regions. The agreement was remarkably tight: a correlation of 0.992 and a Lin concordance correlation coefficient of 0.996. When the entire correlation analysis was repeated using the immunohistochemistry-based stromal density instead of the H&E-based one, the profile of results was reproduced almost exactly, with correlations between the two analyses ranging from 0.94 to 0.996. This means the null finding is not easily explained by measurement error in the pathology pipeline.

The team also stress-tested the associations that did reach nominal significance. When extreme lymphocyte densities or the smallest annotated regions were excluded from the analysis, between 9 and 24 of the original 38 nominal associations persisted, indicating that some of the signals were fragile and dependent on particular data points. After statistical adjustment for clinical covariates including T stage, Nottingham histologic score, Ki-67 proliferation index, and molecular subtype, only 13 of the 38 associations remained nominally significant. The picture that emerges is one of faint, directionally interesting signals that dissolve under the weight of proper multiple-comparison correction, exactly the pattern one expects when exploring a genuinely uncertain hypothesis with a modest sample size.

Why does this matter? PET radiomics has generated enormous enthusiasm in oncology, with hundreds of studies proposing imaging texture features as noninvasive surrogates for everything from gene expression to immune infiltration. The promise is seductive: if a scan could reveal the immune landscape of a tumor without a biopsy, clinicians could select patients for immunotherapy, track treatment response, and monitor clonal evolution noninvasively. But the field has been repeatedly criticized for underpowered studies, lax statistical correction, and features that fail to replicate. This study is a refreshing counterexample. By pre-specifying a conservative statistical framework, validating its pathology measurements against an independent stain, and openly reporting that nothing survived correction, the authors have modeled the kind of intellectual honesty that imaging science needs more of.

The authors themselves are unambiguous about the limits of their work. This is a hypothesis-generating, exploratory, single-institution study of 61 patients, and the paired-region design captures only the extremes of immune-stromal composition rather than a full spatial map. They state plainly that the findings require independent validation in larger, multi-institutional cohorts before any biological or clinical interpretation can be drawn. That caution is warranted, but so is the value of the attempt. The study demonstrates a feasible, reproducible workflow for integrating region-level digital pathology with PET radiomics in breast cancer, and it establishes a realistic benchmark for how weak the true links between tissue microenvironment and metabolic imaging may actually be. If future larger studies confirm even the modest directional trends seen here, immune-rich tumors may one day be recognizable on PET as smaller, more heterogeneous glucose-avid lesions. For now, the honest answer to the question of whether a PET scan can see a tumor’s immune landscape is: perhaps faintly, and not yet reliably.

Subject of Research: Associations between regional immune-stromal tumor features and 18F-FDG PET metabolic heterogeneity in breast cancer

Article Title: Exploratory associations between regional immune-stromal features and 18F-FDG PET-derived metabolic heterogeneity in breast cancer

Article References: Çağdaş, B., & Çağdaş, E. K. (2026). Exploratory associations between regional immune-stromal features and 18F-FDG PET-derived metabolic heterogeneity in breast cancer. BMC Medical Imaging. https://doi.org/10.1186/s12880-026-02789-z

Image Credits: AI Generated

DOI: 10.1186/s12880-026-02789-z

Keywords: breast cancer, PET imaging, radiomics, tumor microenvironment, tumor-infiltrating lymphocytes, digital pathology, metabolic heterogeneity, 18F-FDG, QuPath, fibroblasts, false discovery rate, machine learning

The research team, led by investigators at the Noguchi Memorial Institute for Medical Research at the University of Ghana together with the Ghana Field Epidemiology and Laboratory Training Programme, recruited travelers aged 18 years and above who entered the country through one of ten major land borders between February and July 2022. Participants were enrolled consecutively after providing written informed consent, in accordance with ethical approvals from the NMIMR Institutional Review Board and the Ghana Health Service Ethical Review Committee. During enrollment, each participant completed a semi-structured questionnaire capturing demographic characteristics, vaccination status, recent travel history, and any symptoms suggestive of COVID-19. Blood samples were then collected and shipped to laboratories associated with the Noguchi Memorial Institute and the Infectious Diseases Epidemiology Laboratory for serological analysis.

The laboratory centerpiece of the study was the WANTAI enzyme-linked immunosorbent assay, a commercial ELISA designed to detect antibodies that block the interaction between the viral spike protein and the human ACE2 receptor. This assay targets the spike-receptor binding domain, the region of the SARS-CoV-2 spike glycoprotein that the virus uses to latch onto host cells. Antibodies that bind this domain and prevent receptor engagement are known as neutralizing antibodies, and their presence in serum is widely regarded as a functional correlate of protection against infection. Unlike assays that detect antibodies against the nucleocapsid protein, which distinguish natural infection from vaccination, a spike-RBD neutralizing assay captures immunity induced by both vaccination and prior infection, making it well suited to estimating the total immune experience of a mobile population such as cross-border travelers.

The headline finding was striking. Overall seroprevalence of SARS-CoV-2 neutralizing antibodies among the sampled travelers stood at 92.3 percent, with a 95 percent confidence interval of 91.4 to 93.0 percent. In other words, nearly every traveler screened had immunological evidence of exposure to the virus or of vaccination against it. Seroprevalence was marginally higher before the formal reopening of the land borders, at 95.7 percent (95 percent CI: 94.2 to 97.1), than after reopening, when it measured 91.6 percent (95 percent CI: 91.0 to 92.5). The authors interpret this pattern as evidence that by early to mid-2022, transmission of SARS-CoV-2 in Ghana and neighboring countries had already been so widespread that the closed land borders were no longer holding back a susceptible population; immunity was essentially saturated on both sides of the boundary.

To move beyond raw percentages, the investigators fitted separate logistic regression models for the pre-border-opening and post-border-opening periods. Logistic regression is a statistical technique that models the probability of a binary outcome, in this case seropositivity, as a function of explanatory variables such as border of entry and vaccination status, yielding adjusted odds ratios that quantify the independent contribution of each factor. Because the study tested many coefficients simultaneously, the team applied the Benjamini-Hochberg procedure to control the false discovery rate across all non-reference adjusted coefficient tests within each model. This multiple-testing correction converts conventional p-values into q-values, protecting against spurious associations that inevitably arise when dozens of comparisons are made on a single dataset. Only associations surviving this correction, with q values below 0.05, were treated as statistically robust.

The geography of immunity emerged as a meaningful variable. After false discovery rate correction, travelers entering through the Oseikojokrom and Akanu borders had significantly lower odds of seropositivity before reopening compared with those entering through Aflao, one of Ghana’s busiest eastern border crossings. After reopening, the set of borders with lower odds relative to Aflao expanded to include Oseikojokrom, Akanu, Elubo, Paga, and Sampa. These differences likely reflect heterogeneity in local transmission intensity, vaccination coverage, and the volume and composition of cross-border traffic at each point of entry. Aflao, situated on the densely trafficked frontier with Togo, appears to have functioned as an immunological high-water mark against which the quieter western and northern crossings could be compared.

Vaccination status proved to be a consistent and powerful predictor across both time periods. Before the borders reopened, vaccinated travelers had 2.5 times higher odds of testing positive for neutralizing antibodies than unvaccinated travelers, with an adjusted odds ratio of 2.5 (95 percent CI: 1.7 to 3.5, q less than 0.001). After reopening, the association persisted though it attenuated somewhat, with vaccinated individuals showing 1.8 times higher odds of seropositivity (aOR = 1.8, 95 percent CI: 1.3 to 2.6, q = 0.005). The attenuation is plausibly explained by the growing contribution of natural infection to population immunity as the Omicron era progressed, which would have raised seropositivity among the unvaccinated and thereby narrowed the gap attributable to vaccination alone.

For the authors, the central policy implication is one of reassurance rather than alarm. The persistently high seroprevalence observed both before and after border reopening suggests that Ghana posed little to no risk of cross-border transmission of SARS-CoV-2 during the study window. Indeed, the pre-opening assessment provided empirical evidence that informed recommendations to reopen the land borders, a decision with substantial economic weight given the role of overland trade in the region and the ambitions of the African Continental Free Trade Area, which depends on the free movement of goods and people across these very crossings. Prolonged border closures carry costs of their own, from disrupted livelihoods to stalled commerce, and serosurveillance offers a way to weigh those costs against a realistic estimate of epidemiological risk.

The study also underscores the value of points of entry as natural surveillance platforms. Land borders are notoriously difficult settings for infectious disease monitoring, since travelers move in high volumes and formal health screening captures only symptomatic individuals. By embedding serological sampling directly into the border environment, with support from the Ghana Immigration Service and staff at the points of entry, the research team converted a routine administrative boundary into a population-level immunological observatory. The approach could be adapted for future pathogens, providing early warning of shifting immunity landscapes and helping authorities calibrate travel restrictions that are proportionate to actual risk rather than precautionary defaults.

Caveats remain, as they do in any cross-sectional serosurvey. Because the WANTAI ELISA detects binding to the spike-receptor binding domain, it cannot by itself disentangle antibodies generated by vaccination from those generated by infection, and a positive result indicates prior immune stimulation rather than current protection against emerging variants. The sampling of travelers also means the findings speak most directly to mobile, border-crossing populations, which may differ from the general population in age, vaccination uptake, and exposure history. Nevertheless, the sheer magnitude of the seroprevalence figures, consistently above 90 percent across both periods, leaves little ambiguity about the broader picture. By 2022, the virus had left its immunological signature across virtually the entire traveling population of Ghana’s land frontiers, and the borders could be reopened with confidence that they were not gateways for a naive population’s first encounter with SARS-CoV-2. The work, funded through the Global Fund’s COVID-19 Response Mechanism via the National AIDS Control Programme of the Ghana Health Service, stands as a template for evidence-based border policy in future epidemic emergencies.

Subject of Research: Seroprevalence of SARS-CoV-2 neutralizing antibodies among travelers entering Ghana through major land borders before and after border reopening in 2022

Article Title: Seroprevalence of SARS-CoV-2 neutralizing antibodies among travelers entering Ghana through major land borders, pre- and post-opening of borders, 2022

Article References: Donkor, I. O., Lomotey, E. S., Odumang, D. A., Mohammed, A. G., Asante, I. A., Kploanyi, E. E., Noora, C. L., Addo-Lartey, A., Affram, Y., Frimpong, J. A., Odikro, M. A., & Kenu, E. (2026). Seroprevalence of SARS-CoV-2 neutralizing antibodies among travelers entering Ghana through major land borders, pre- and post-opening of borders, 2022. BMC Infectious Diseases. https://doi.org/10.1186/s12879-026-14483-0

Image Credits: AI Generated

DOI: 10.1186/s12879-026-14483-0

Keywords: SARS-CoV-2, seroprevalence, neutralizing antibodies, Ghana, land borders, WANTAI ELISA, COVID-19 vaccination, border surveillance, logistic regression, false discovery rate, public health, BMC Infectious Diseases

A new cross-sectional study published in BMC Geriatrics by Ji-Hye Jung, Woo-young Shin, and Jung-Ha Kim of Chung-Ang University College of Medicine set out to answer a deceptively simple question: what actually shapes family caregivers’ attitudes toward deprescribing? The researchers surveyed 973 family caregivers of residents living in South Korean long-term care facilities, using an online questionnaire and a validated Korean-language instrument called the revised Patients’ Attitudes Towards Deprescribing questionnaire, or K-rPATD. Their findings, published as an open-access research article, offer a data-driven map of when families are most receptive to the conversation about cutting back medications.

The survey population itself tells a story about who makes these decisions. Among the 973 caregivers, 56.0 percent were male and 88.0 percent had a college education or higher — a notably well-educated cohort. Fully 73.0 percent reported that their relatives carried dual-track prescriptions, meaning medications were being managed through more than one prescribing pathway. The residents themselves took an average of 5.6 medications, with a standard deviation of 3.5, and 16.3 percent had experienced at least one adverse drug event, the clinical term for a harmful reaction to a medication.

The K-rPATD breaks attitudes into subscales, and the pattern of scores is revealing. The involvement subscale — how much caregivers want to be part of deprescribing decisions — scored highest at a mean of 3.63 on the instrument’s scale, followed closely by the burden subscale at 3.61, which captures how much caregivers perceive medications as a burden. Concerns about stopping scored 3.33, while the reverse-scored appropriateness subscale, which reflects how necessary caregivers believe current medications to be, had the lowest mean at 2.52. In plain terms: families overwhelmingly want to be involved, they feel the weight of the medication regimen, but they also tend to believe the drugs their relatives take are appropriate and necessary.

Here is where the study’s methodological rigor becomes the real story. The researchers measured 16 potential predictors, represented by 18 regression terms, and entered them simultaneously into multiple linear regression models predicting four subscale scores and two global attitude items. That produced 108 regression coefficients in total. In a large dataset, many coefficients will look statistically significant by chance alone — a phenomenon statisticians call the multiple comparisons problem. To guard against it, the team computed both unadjusted p values and p values adjusted using the Benjamini-Hochberg false discovery rate procedure, a technique that controls the expected proportion of false positives among declared findings. They also used heteroscedasticity-consistent standard errors to keep the estimates robust even when the variability of responses differed across predictor levels.

After this statistical gauntlet, only eight associations survived. Adverse drug events were linked with higher scores on the burden and concerns-about-stopping subscales and lower scores on the appropriateness subscale — an intuitive cluster, since a family that has watched a loved one suffer a medication side effect is more likely to see the drugs as burdensome, more worried about what they do, and less convinced they are all needed. Medication cost was associated with higher burden and concerns-about-stopping scores, suggesting that the financial weight of prescriptions shapes attitudes in parallel with the clinical weight. The number of medications a resident took predicted higher burden scores, and the number of chronic conditions predicted both higher concerns-about-stopping scores and higher involvement scores.

What did not survive the correction is just as telling. Caregivers’ own sociodemographic characteristics — their age, sex, education, and similar attributes — did not emerge as reliable drivers of deprescribing attitudes once the false discovery rate correction was applied. The nominally significant associations that appeared before correction are, the authors emphasize, exploratory only. The message is that attitudes track the resident’s medication-related circumstances, not the caregiver’s background. A family’s willingness to discuss deprescribing is forged by lived experience with the drug regimen: the side effects they have witnessed, the bills they pay, the pill count, and the chronic disease burden their relative carries.

The authors are careful about what this cross-sectional design can and cannot show. Because attitudes and circumstances were measured at the same moment, the study cannot establish causation — it cannot prove, for example, that experiencing an adverse drug event causes caregivers to view medications as more burdensome, only that the two travel together. The findings are described as hypothesis-generating, a scientific term of art meaning they point toward questions that future longitudinal or interventional studies should test. Still, the practical implication is concrete: clinicians who want to raise deprescribing with families now have an evidence-based way to identify the caregivers with whom the conversation is most likely to resonate — those whose relatives have experienced adverse drug events, face high medication costs, take many medications, or manage multiple chronic conditions.

The study was approved by the Institutional Review Board of Chung-Ang University and conducted in accordance with the Declaration of Helsinki, with informed consent from all participants. It was funded by a grant from the Patient-Centered Clinical Research Coordinating Center, funded by South Korea’s Ministry of Health and Welfare, with no funder involvement in design, analysis, or writing. The authors report no competing interests, and they disclosed using AI-assisted tools during manuscript preparation while taking full responsibility for the final content. The complete regression tables, sensitivity analyses, and a completed STROBE reporting checklist are available as supplementary materials, reflecting a transparency standard that makes the statistical machinery — all 108 coefficients, adjusted and unadjusted — open to scrutiny.

For a field where deprescribing conversations often stall at the family’s door, this study reframes the problem. Rather than assuming some caregivers are simply more receptive by temperament or education, the data suggest receptivity is situational, rooted in the concrete realities of the medication regimen itself. That insight could reshape how clinicians approach families in long-term care: instead of a generic pitch for fewer pills, the conversation might begin where the family’s own experience already lives — with the side effect that frightened them, the cost that stings, or the mounting list of prescriptions for a body fighting several chronic diseases at once. In the arithmetic of aging, the study suggests, the path to fewer medications runs through the family’s lived experience of the ones already in the pill organizer.

Subject of Research: Family caregivers' attitudes toward deprescribing medications for older adults in long-term care facilities in South Korea

Article Title: Factors associated with family caregivers’ attitudes toward deprescribing in long-term care facilities: a cross-sectional study in South Korea

Article References: Jung, J.-H., Shin, W.-Y., & Kim, J.-H. (2026). Factors associated with family caregivers’ attitudes toward deprescribing in long-term care facilities: a cross-sectional study in South Korea. BMC Geriatrics. https://doi.org/10.1186/s12877-026-08333-6

Image Credits: AI Generated

DOI: 10.1186/s12877-026-08333-6

Keywords: deprescribing, polypharmacy, family caregivers, long-term care facilities, K-rPATD, adverse drug events, medication cost, geriatrics, South Korea, cross-sectional study, false discovery rate, medication burden

The central conceptual move is a sharp separation between what is observed and what is inferred. For each unit i, the method defines a parameter-conditional tail functional, the probability that a replicated count drawn from the fitted model would exceed a threshold, given that unit’s covariates. Averaging this functional over the posterior distribution of the model parameters yields the posterior predictive exceedance probability, the quantity used for ranking and calibration. A large realized count, the paper stresses, is informative but is not itself the inferential target: a person with moderate observed utilization may still carry high replicated-count risk once health status, insurance characteristics, and model uncertainty are accounted for. Conversely, an isolated spike in the data may reflect contamination rather than genuine tail risk.

The working model is a negative binomial regression, a standard choice for overdispersed counts because it relaxes the restrictive variance assumption of the Poisson model while retaining an interpretable conditional mean. The robustness, however, comes from an additional layer. Each observation receives a positive latent scale that locally inflates or deflates its mean, drawn from a two-component mixture: a point mass at one, which preserves the ordinary negative binomial baseline, and a Beta-prime slab with substantial mass near zero and a polynomially decaying right tail. This construction, adapted from recent robust Bayesian count modeling, allows excess zeros and isolated large counts to be absorbed by the local scale rather than forcing them to distort the regression coefficients or the fitted tail probabilities. In effect, unusual observations are quarantined instead of being allowed to hijack the analysis.

The theory is deliberately framed for the realistic case where the model is wrong. Rather than assuming the data follow the negative binomial family exactly, the contraction results are stated around the Kullback–Leibler projection of the true law onto the working family — a pseudo-true parameter that represents the best available approximation. Following the classical misspecification perspective of Berk and of Kleijn and van der Vaart, the posterior is shown to concentrate near this projection, and a Lipschitz transfer argument shows that concentration in parameters carries over to uniform concentration of the exceedance functionals themselves. Pinsker’s inequality then bounds the error between true tail probabilities and the working-model targets by the Kullback–Leibler discrepancy, giving a quantitative account of how much misspecification can distort screening decisions.

Discovery, as opposed to ranking, is handled by a separate layer built on e-values. Posterior thresholding of the evidence probability — the posterior probability that a unit’s exceedance probability exceeds a chosen level gamma — is a natural local Bayes rule under a loss function penalizing false positives and false negatives, but it does not by itself control the false discovery rate across many simultaneous tests. The paper therefore converts posterior odds into Bayes factors and submits them to the e-BH procedure of Wang and Ramdas, which controls the false discovery rate under arbitrary dependence among the test statistics. The validity of this step is stated with unusual care: under the Bayes-marginal null induced by the same hierarchical model, posterior odds are Bayes factors and hence valid e-values with null expectation at most one. Under a mere working-model interpretation, the same quantities are evidence scores unless calibrated, and the author includes an approximate-validity result in which the FDR bound is inflated by a calibration error term. This honesty avoids the common pitfall of treating a parametric Bayesian fit as automatically producing model-free frequentist guarantees.

Computation is provided at three levels of fidelity. A reference Markov chain Monte Carlo sampler uses Pólya–Gamma augmentation to update the regression coefficients in a single Gaussian block, with slice sampling and adaptive random-walk Metropolis steps for the latent scales, dispersion parameter, and slab hyperparameters; convergence is monitored with rank-normalized split R-hat statistics and effective sample size thresholds. For large datasets and repeated simulations, a bounded-influence iteratively reweighted least squares Laplace approximation downweights extreme Pearson residuals and high-leverage observations, producing a Gaussian approximate posterior from which the exceedance summaries can be drawn directly. A mean-field variational option offers fast screening. A theorem bounds the error of any such approximation: the discrepancy between exact and approximate posterior exceedance probabilities is controlled by the total variation distance between the two posteriors, which in turn is bounded by a square root of the Kullback–Leibler divergence between them.

The simulation evidence demonstrates a clear and quantified robustness–efficiency tradeoff. Under clean data, the ordinary negative binomial fit is slightly more efficient, with marginally higher true positive rates and smaller calibration error. Under contamination — engineered as a mixture of near-zero structural zeros and heavy upper-tail multiplicative contamination at rates of ten and twenty percent — the picture reverses dramatically. At ten percent contamination, the robust procedure cuts the false discovery proportion from 0.029 to 0.002 and halves the calibration error metrics. At twenty percent contamination and a sample size of one thousand, the naive method’s false discovery proportion climbs to roughly 0.099 while the robust method holds near 0.004, with expected calibration error of about 0.090 versus 0.040. A more aggressive variant using Tukey’s biweight weights proves most conservative of all, illustrating that the degree of robustness tunes a power–conservatism dial rather than dominating uniformly. Reliability plots confirm that the robust fit tracks the clean oracle tail probabilities across the range, while the naive fit inflates the upper bins under contamination.

The real-data application turns to the RAND Health Insurance Experiment, a classic dataset of 20,190 observations on outpatient physician visits, adjusted for insurance plan, income, physical limitation, chronic disease burden, and self-rated health. Screening for individuals whose replicated visit count exceeds eight visits, with a discovery level of 0.3 and e-BH applied at a false discovery rate of 0.1, both fits select subgroups whose observed exceedance rates — 0.339 for the naive fit and 0.315 for the robust fit — tower over the full-sample rate of 0.071. The robust fit selects a somewhat larger set while shrinking tail probabilities more aggressively over most of the sample, yet still assigns high risk to individuals with severe health profiles. A training–holdout split with 2,000 posterior predictive replicates showed the observed zero rate falling at the edge of its predictive interval, while the mean count and exceedance rate exceeded theirs — a candid indication that the working model remains conservative in the upper tail and not fully distributionally adequate.

Perhaps the most practically important feature of the paper is its insistence on reporting three distinct outputs separately: calibrated tail-risk rankings, posterior evidence scores, and multiplicity-controlled discovery sets. Some individuals selected by the procedure did not actually exceed eight visits in the observed record, which the author presents as expected behavior rather than a failure — the method targets replicated-count risk conditional on covariates and model uncertainty, not the realized indicator alone. The selected individuals are characterized by physical limitation and high chronic-disease burden, exactly the covariate profiles one would associate with latent high utilization. The paper also warns against a subtle computational trap: e-value caps that are too small can mechanically prevent the e-BH procedure from making any rejections, so decision steps must use uncapped Bayes factors or caps compatible with the rejection thresholds.

Taken together, the study offers a transparent template for combining Bayesian tail modeling, robust count regression, and e-value-based multiple testing in settings where the data are messy, overdispersed, and possibly contaminated. The author suggests that future work may extend the framework to richer dependence structures, more flexible count families, and sequential e-value methods for online monitoring — a natural next step for applications ranging from epidemiological surveillance to industrial quality control, where the question is not merely how many events occur, but who is genuinely at risk of crossing the line.

Subject of Research: Robust Bayesian exceedance screening and false discovery rate control for overdispersed count data

Article Title: Robust Bayesian exceedance screening for count data

Article References: Sadeghkhani, A. (2026). Robust Bayesian exceedance screening for count data. International Journal of Data Science and Analytics, 22(1), Article 323. https://doi.org/10.1007/s41060-026-01312-5

Image Credits: AI Generated

DOI: 10.1007/s41060-026-01312-5

Keywords: Bayesian inference, exceedance probability, e-values, negative binomial regression, robust statistics, false discovery rate, count data, Kullback–Leibler projection, RAND Health Insurance Experiment, multiple testing, posterior contraction, health utilization

The study, led by Qiyu Wang and Yiluan Li of the University of Science and Technology of China together with Yunfei Peng and Zheng-Zheng Tang of the University of Wisconsin-Madison, systematically benchmarked nine existing microbiome mediation methods across a broad sweep of simulation settings and a real data application. Their central finding is stark: across the simulation settings examined, many of the methods exhibited inflated error rates, and the inflation grew worse as compositional effects intensified. In other words, the very property that makes microbiome data distinctive is the property that undermines the statistical guarantees these methods are supposed to provide.

To understand why, one has to appreciate a technical subtlety that plagues nearly all microbiome sequencing. Standard sequencing workflows do not directly measure how many cells of each taxon are present in a sample. Instead, they yield relative abundances, which describe each taxon’s share of the total sequenced material. Absolute abundances, the actual quantities that usually matter for mechanistic interpretation, are not directly observed. Because every relative abundance depends on all the others through the constraint that shares must sum to one, changing the true amount of one taxon mechanically alters the measured proportions of every other taxon in the sample, even when nothing biological has happened to them. This is the compositional effect, and it is well known to cause trouble in differential abundance testing.

What the new study establishes is that the same compositional distortion systematically corrupts mediation analysis as well. Mediation tests evaluate whether the effect of an exposure on an outcome travels through a candidate taxon, typically by modeling the exposure’s effect on the microbe and the microbe’s effect on the outcome and then testing the product of these pathways. When the taxon measurements are relative rather than absolute, spurious correlations can be induced between taxa and outcomes simply because of shifts elsewhere in the community. The result is statistical signals that look like genuine mediation but are in fact artifacts of the compositional structure of the data. Despite documented compositionality-induced false positives in other microbiome analyses, systematic benchmarking of microbiome mediation methods had remained scarce, leaving practitioners without a clear picture of which tools, if any, controlled error reliably.

A second innovation of the study lies in how the simulations were constructed. Synthetic data used to evaluate bioinformatic methods can easily be unrealistic, and a benchmark is only as convincing as the data it rests on. The team anchored their simulations to an experimentally derived absolute abundance template dataset, generating synthetic absolute abundance profiles that better preserve the empirical structure of real microbiome communities than commonly used simulators. This matters because the severity of compositional effects depends on realistic properties such as the distribution of taxon proportions and the correlation structure among taxa, so a simulator that smooths these features away can dramatically understate the problems that arise in practice. By grounding the simulations in real experimental measurements, the researchers aimed to make their benchmark reflect the conditions analysts actually face.

The results of that benchmark were sobering. Across the simulation settings, the nine methods varied widely in how well they controlled false discoveries, with a substantial number showing inflation of error rates beyond their nominal levels, particularly as compositional effects grew stronger. The pattern carried over into real data. In their application to a real microbiome dataset, the researchers observed that the methods which had exhibited severe inflation in simulations also returned substantially larger sets of candidate mediators, a hallmark of methods that are too permissive and flag taxa indiscriminately. For a field where mediator lists often drive downstream experiments and therapeutic hypotheses, such inflated mediator sets represent more than a statistical nuisance; they can send entire research programs chasing signals that do not exist.

The remedy proposed by the team is a new method called CAMRA, which is designed to infer and test mediation effects at the absolute abundance level while working from the relative abundance data that standard sequencing pipelines produce. Rather than treating the observed relative abundances as if they were the quantities of mechanistic interest, CAMRA attempts to account for the compositional relationship between what is measured and what is biologically meaningful, and it performs its mediation testing on that inferred absolute scale. The goal is to align the statistical target of the analysis with the scale on which microbial mechanisms operate, thereby removing the source of the spurious associations that plague relative abundance level testing.

According to the study, CAMRA delivered on both fronts that matter for practitioners. In simulations, the method improved calibration of the false discovery rate, meaning that when it reported a controlled error level, the actual rate of false discoveries conformed to that promise. It also improved power for taxon level mediator discovery, detecting genuine mediators more effectively than the approaches that either drowned in false positives or, in guarding against them, sacrificed sensitivity. Equally important for real world adoption, CAMRA maintained favorable runtime, so the gains in statistical reliability do not come at a computational cost that would make the method impractical for the high dimensional taxon tables typical of microbiome studies.

The implications reach well beyond the immediate technical community. Identifying microbial taxa that mediate the effects of exposures on health outcomes can elucidate causal pathways and suggest therapeutic targets, which is precisely the kind of translational promise that has drawn enormous investment into microbiome research. If the mediation analyses underlying such claims are systematically inflated, therapeutic leads may be built on statistical mirages, and failed validation studies down the line may be misattributed to biology rather than to methodology. Well-calibrated inference, the authors argue, is a prerequisite for reliable microbial mediator discovery, and their benchmark provides the field with its first systematic map of where current methods stand.

For working microbiome scientists, the practical guidance emerging from this study is to treat mediator lists produced by relative abundance level tests with heightened caution, especially in communities where compositional effects are strong, and to consider absolute abundance targeted approaches such as CAMRA when the goal is taxon level mediator discovery. The study also underscores the value of rigorous, simulation grounded benchmarking as a discipline in its own right: by anchoring synthetic data to experimental templates and evaluating error control explicitly, methodologists can surface failure modes that individual applications would never reveal. As microbiome mediation analysis continues to move from exploratory statistics toward the justification of therapeutic interventions, the difference between a well calibrated test and an inflated one may ultimately determine which microbial targets turn out to be real.

Subject of Research: Statistical error control and compositional bias in microbiome mediation analysis

Article Title: Error control for microbiome mediation analysis: benchmarking and remedy

Article References: Wang, Q., Li, Y., Peng, Y., & Tang, Z.-Z. (2026). Error control for microbiome mediation analysis: benchmarking and remedy. Microbiome. https://doi.org/10.1186/s40168-026-02527-1

Image Credits: AI Generated

DOI: 10.1186/s40168-026-02527-1

Keywords: microbiome, mediation analysis, relative abundance, absolute abundance, compositional effects, false discovery rate, benchmarking, CAMRA, statistical methods, microbial taxa, causal inference, bioinformatics

The core innovation of KNOT lies in its use of knockoffs, a concept borrowed from modern high-dimensional statistics. A knockoff is a deliberately constructed decoy variable that mimics the correlation structure of the original genetic variants but has no genuine relationship with the phenotype being studied. By presenting a neural network with both real variants and their knockoff counterparts, researchers can ask a pointed question: does the model treat a real variant as more important than its decoy? If a variant consistently outranks its knockoff across many comparisons, that variant earns a place on the candidate list. Crucially, the proportion of decoys that slip through provides a direct estimate of how many false positives are contaminating the results. This is the essence of false discovery rate control, a guarantee that conventional deep-learning pipelines for genetic association simply do not offer. KNOT wraps its neural network inside this knockoff framework, producing variable selection that is both flexible enough to capture complex genetic patterns and disciplined enough to quantify its own error rate.

Stabilizing the network’s stochastic behavior is a second pillar of the method. Neural networks initialized with different random seeds can produce different importance rankings, which is unacceptable when the goal is to nominate specific variants for biological follow-up. KNOT addresses this instability through contrastive learning, a training strategy that teaches the model to distinguish between related and unrelated samples. In family-based association studies, where parents and siblings share predictable fractions of their genomes, modeling sample relatedness is not optional; it is the difference between a signal that reflects genuine inherited risk and one that merely reflects family structure. By learning representations that account for these kinship relationships, KNOT reduces the variance of its selections and produces rankings that remain consistent across repeated runs, a property the authors identify as essential for trustworthy variable selection in genetic data.

Attention mechanisms supply the third technical ingredient. Within any genomic region, nearby variants travel together through generations as inherited blocks, a phenomenon known as linkage disequilibrium. An attention-based architecture allows the network to weigh the contributions of all variants in a block simultaneously, learning which positions within a correlated cluster carry the strongest association with disease. This is a fundamentally different approach from the marginal testing of classical GWAS, which treats each variant in isolation and can attribute a signal to the wrong member of a correlated group. By explicitly modeling the correlation structure rather than ignoring it, KNOT aims to sharpen the localization of risk variants, bringing the nominated positions closer to the variants that actually matter biologically.

Epistasis, the interaction between variants at different genomic locations, represents one of the most tantalizing and least explored dimensions of genetic risk. Two variants may be harmless individually yet consequential in combination, a pattern that single-marker tests are structurally blind to. Exhaustive testing of all pairwise interactions across millions of variants is computationally prohibitive and statistically punishing, because the sheer number of tests inflates the burden of multiple-comparison correction. KNOT incorporates a permutation test to detect these epistatic interactions, using label permutation to establish a null distribution against which observed interaction effects can be judged. The permutation approach preserves the correlation structure of the data while destroying any genuine relationship between genotype and phenotype, providing a principled benchmark for deciding when an interaction is real.

The choice of a family-based study design is itself significant. Family-based association studies compare the alleles transmitted from parents to affected children against those not transmitted, a design that is inherently robust to confounding from population stratification, the spurious associations that arise when cases and controls differ in ancestry. This robustness comes at a cost: family designs typically have fewer effective samples and more complex correlation patterns than population-based case-control cohorts, which is precisely why deep-learning methods that ignore relatedness have struggled in this setting. KNOT was engineered from the outset for this environment, with its contrastive learning component explicitly modeling the relatedness that family data impose. The method therefore occupies a deliberately chosen niche, applying modern representation learning where its statistical hazards are most acute and its potential payoff is greatest.

To demonstrate the method’s practical value, Fu and Yang applied KNOT to two autism spectrum disorder family cohorts. Autism spectrum disorder is a compelling test case for any new genetic discovery method. Its heritability is high, yet much of that heritability remains unexplained by the variants identified to date, a puzzle often described as the missing heritability problem. Epistatic interactions are frequently invoked as one plausible reservoir of this unexplained risk, making a method that can jointly identify risk variants and their interactions especially relevant. In these applications, KNOT outperformed conventional methods, identifying more known and putative risk loci as well as additional epistatic interactions associated with autism. The recovery of known loci serves as a sanity check that the method is finding real signals, while the putative novel loci and interactions represent hypotheses for the field to pursue.

The broader significance of this work extends beyond autism genetics. The challenge that KNOT addresses, namely how to extract interpretable and statistically controlled discoveries from a flexible neural network applied to correlated high-dimensional data, is a defining problem of the current era of machine learning in the life sciences. Knockoff-based frameworks offer a general template: the neural network provides the representational power to model complex dependencies, while the knockoff construction provides the inferential scaffolding to keep the resulting claims honest. If this template generalizes, it could reshape how deep learning is deployed across association studies of other complex traits, from cardiovascular disease to psychiatric conditions, wherever correlated predictors and limited sample sizes conspire to produce fragile discoveries.

There are, of course, caveats that temper the enthusiasm. Knockoff construction depends on an accurate model of the correlation structure among variables, and genomic linkage patterns vary across populations; imperfect knockoffs can degrade the precision of the false discovery rate guarantee. Neural network training remains computationally demanding, and the permutation testing for epistasis adds further cost. The authors report their work from the Department of Biostatistics at City University of Hong Kong, supported by the Research Grants Council of Hong Kong Early Career Scheme, and the method’s performance will ultimately be judged by how well its discoveries replicate in independent cohorts, the ultimate arbiter of any genetic association claim. The study, published open access on 24 September 2026 in Genome Biology under a Creative Commons Attribution 4.0 license, invites exactly that scrutiny.

What makes KNOT a noteworthy development is not any single component but the synthesis. Contrastive learning tames the randomness of neural training, attention mechanisms respect the correlation architecture of the genome, knockoffs impose a quantifiable error budget on the selections, and permutation testing opens a window onto variant interactions that conventional pipelines cannot see. Applied to two autism family cohorts, this combination surfaced more known and putative risk loci and more epistatic interactions than established approaches. For a field that has spent years wrestling with the trade-off between the flexibility of deep learning and the rigor of classical statistics, KNOT offers a persuasive demonstration that the two need not be enemies. The variants and interactions it nominates now join the long queue of candidates awaiting biological validation, but the statistical machinery that nominated them marks a meaningful step toward deep-learning-driven genetics that researchers can actually trust.

Subject of Research: A knockoff-augmented deep learning method for identifying genetic risk variants and epistatic interactions in family-based genome-wide association studies of autism

Article Title: KNOT: a knockoff-augmented neural network for identifying risk variants and epistatic interactions in family-based association studies

Article References: Fu, Y., & Yang, Y. (2026). KNOT: a knockoff-augmented neural network for identifying risk variants and epistatic interactions in family-based association studies. Genome Biology. https://doi.org/10.1186/s13059-026-04285-7

Image Credits: AI Generated

DOI: 10.1186/s13059-026-04285-7

Keywords: GWAS, deep learning, neural networks, knockoff statistics, epistasis, autism, family-based design, false discovery rate, linkage disequilibrium, contrastive learning, attention mechanisms, Genome Biology