Genetic variants in RNU4-2, which is transcribed into the U4 small nuclear RNA component of the major spliceosome, were recently shown to cause ReNU syndrome, a prevalent dominant neurodevelopmental disorder (NDD). These variants almost exclusively arise de novo and cluster within 18 nucleotides of RNU4-2. Here we describe a new recessive NDD associated with homozygous and compound heterozygous variants in RNU4-2. We identify 38 individuals with biallelic variants outside the 18-nucleotide ReNU syndrome region that cluster within other functionally important elements of U4: Stem II, the k-turn and the Sm protein binding site. We characterize the clinical phenotype in 31 individuals, demonstrating that the recessive disorder is clinically distinct from ReNU syndrome and is associated with distinctive white matter abnormalities, including enlarged perivascular spaces. Finally, we find reduced RNU4-2 transcript levels in individuals with the recessive disorder, suggesting a loss-of-function disease mechanism that is distinct from the mechanism underlying ReNU syndrome. Together, these findings expand the genotypic and phenotypic spectrum of RNU4-2-associated NDDs.

DNA replication is carried out by the replisome and is essential for maintaining genome integrity and cell proliferation. Pathogenic variants in genes encoding various replisome components cause microcephalic primordial dwarfism (MPD), characterized by growth retardation, microcephaly, and developmental abnormalities. Here, we report bi-allelic hypomorphic variants in WDHD1 as a cause of MPD with a broad spectrum of additional abnormalities, including acute liver failure, in 17 subjects from 14 families. WDHD1 encodes a replisome scaffolding protein (also known as AND-1 and Ctf4), which is essential for replisome assembly, replication fork stability, and sister chromatid cohesion. We found aberrant splicing of WDHD1 pre-mRNAs for all intronic variants tested and markedly reduced WDHD1 protein levels in subject-derived fibroblasts. Fibroblasts with bi-allelic WDHD1 variants showed globally reduced replication fork speed and impaired replication control, accompanied by spontaneous DNA damage and a G1-to-S transition defect. Using various cell biology approaches, we show that subject fibroblasts displayed reduced proliferation, abnormal nuclear morphology, including micronuclei, multilobed, and enlarged nuclei, as well as an increased number of metaphases with premature sister chromatid separation. Together, our findings establish WDHD1 as a protein required for normal organismal growth and development in humans and underscore its multiple functions in maintaining genome integrity.

Recent advances in Mendelian genomics reveal the importance of variant-level characterization of allelic disorders. Non-muscle actin isoforms, encoded by the genes ACTB and ACTG1, are the most abundant intracellular proteins, but historically, they are often regarded as merely being “housekeeping” molecules. Here, we illuminate the extraordinary clinical heterogeneity and complex pathobiology of genetic non-muscle actinopathies. To do this, we combine human genomics studies with molecular biology. Strikingly, variants in ACTB and ACTG1 isoforms generate at least eight distinct clinical disorders. A subset of disease-associated missense variants causes dysregulated actin polymerization-depolymerization and neuronal migration defects. In contrast, nonsense, frameshift, and missense variants enhancing protein degradation cause milder phenotypes or are benign. These results emphasize the essential functional aspects of the non-muscle actin isoforms. Critically, they additionally constitute a template for the personalized genetic variant-level-driven management of the pleiotropic allelic single-gene disorders.

While genome-wide association studies (GWAS) have linked common genetic variants to COVID-19 susceptibility and severity, rare high-impact variants may also contribute to phenotypic heterogeneity. Inborn errors of type I interferon immunity (IFN-I-IEIs), including X-linked TLR7 deficiency, account for 2

Hand radiographs are routinely used by pediatric endocrinologists to assess bone age in children presenting with atypical growth, pubertal disturbances, or other disorders of physical maturation. However, their diagnostic utility beyond skeletal maturation, particularly for differential diagnosis of growth disorders, remains largely untapped. Here, we introduce novel and well-defined morphometric measurements in hand radiographs and present algorithms for their automatic extraction. We retrospectively analyzed pediatric hand radiographs from 9 conditions (n = 1283), representing a spectrum of growth-affecting endocrine, metabolic, and genetic disorders with variable skeletal involvement (such as SHOX deficiency, Noonan syndrome, and Ullrich-Turner syndrome) and skeletal dysplasias with pronounced hand dysmorphology (such as mucopolysaccharidosis and achondroplasia). From these images, we extracted 550 phenotypical features from metacarpal and phalangeal bones, which were further processed via age- and sex-independent z-scores based on unaffected (UA) controls (n = 1353) to generate a high-dimensional standardized phenotypic space. Using this feature space, a binary screening based solely on the largely available UA training data achieved 82% specificity on an independent UA test set and condition-dependent sensitivities 48.3%-98.9%, enabling robust detection of growth disorders without requiring affected training samples. As a secondary demonstration, a 10-class classifier achieved a balanced test-set accuracy of 75%, with particularly high accuracy for UA controls (94%). Our findings demonstrate that these new measurements provide a low-cost, interpretable, and reproducible method that can be integrated into routine bone age evaluation for growth disorder screening and clinical decision support.

Variants in BRSK2, encoding brain specific kinase-2, have recently been associated with an autosomal dominant neurodevelopmental disorder (NDD). We have assembled 52 cases with heterozygous BRSK2 variants and variable neurodevelopmental phenotypes with frequent neuropsychiatric and behavioral symptoms. The variant spectrum included 15 different truncating variants, seven (potential) splice variants, three structural variants, and 12 different missense variants. Of the missense variants, seven were in the kinase domain, and the others in the UBA and the KA1 domain or outside domains. Variants occurred de novo in 19 cases and were inherited in 18. We utilized Drosophila melanogaster as a model and assessed viability and performed climbing and bang sensitivity assays upon knockdown of the fly orthologue sff or upon overexpression of wildtype or mutant human BRSK2. Pan-neuronal knockdown of sff resulted in impaired locomotor behavior and seizure susceptibility. Ubiquitous or pan-neuronal overexpression of human wildtype BRSK2 in Drosophila resulted in lethality or locomotor impairment, respectively, indicating toxicity. Overexpressing mutant BRSK2 did not or incompletely affect viability and locomotor behavior for six of seven tested kinase domain missense variants and one KA1 domain variant, indicating a (partial) loss-of-function effect. Interestingly, overexpressing BRSK2 with the remaining missense variant from the kinase domain and the two most C-terminal missense variants resulted in possible gain of function. Our findings further delineate the clinical and molecular spectrum of BRSK2-associated NDD and provide further insights into the role of BRSK2/sff in nervous system function and dysfunction.

BACKGROUND:Multiparameter flow cytometry is a cornerstone of B cell non-Hodgkin lymphoma (B-NHL) diagnostics, but interpretation requires substantial expertise and is complicated by high-dimensional data, variable sample quality, limited data for rare entities, and evolving clinical classification systems. Current artificial intelligence approaches often require large training datasets and provide limited insight into the rationale behind individual diagnostic decisions. METHODS AND FINDINGS:We developed FlowXAI, a self-explaining artificial intelligence system designed to support B-NHL classification while explicitly reporting case-level diagnostic trustworthiness. FlowXAI combines unsupervised structural analysis with a clinically motivated, multi-level diagnostic framework reflecting routine diagnostic priorities. An unsupervised Tile Mining (TM) procedure performs pre-diagnostic sample-quality assessment by identifying structurally atypical samples. TM is applied to filter training data, enabling substantial reduction of training requirements while preserving unbiased evaluation on independent test samples. FlowXAI was evaluated using repeated cross-validation on 19,493 peripheral blood samples and further assessed on an independent external benchmark dataset generated at a separate diagnostic center using a different antibody panel. Across diagnostic levels, FlowXAI achieved performance comparable to a deep learning-based system despite requiring approximately two orders of magnitude fewer training samples. When predictions were classified as confident by the system's internal self-assessment, diagnostic performance exceeded that of the neural network baseline. Unsupervised structural analysis demonstrated clear separation between normal controls and selected lymphoma entities such as chronic lymphocytic leukemia-like lymphomas and hairy cell leukemia, while other entities were not clearly separable using the antibody panels studied. CONCLUSIONS:FlowXAI provides accurate, data-efficient, and transparent support for B-NHL immunophenotyping from nonstandardized flow cytometry data. By combining interpretable decision logic with explicit self-assessment, FlowXAI offers a clinically meaningful framework for diagnostic support and training, particularly in settings with limited expert availability or rare lymphoma subtypes. The main limitation is the retrospective evaluation using specific antibody panels, and FlowXAI requires prospective validation as a decision-support tool within integrated diagnostic workflows.

Deep learning-based facial phenotyping represents a major paradigm shift in the diagnosis of rare and ultra-rare genetic disorders. By capturing disease-specific craniofacial 'gestalts' that are often subtle, overlapping, but overlooked in routine clinical practice, these technologies surpass the traditional limits of dysmorphology assessment. Despite this, data scarcity and stringent privacy policies constraint centralized model training and its clinical translation. Swarm learning, a decentralized paradigm that combines edge-based training with blockchain-mediated parameter synchronization, offers a potential solution by enabling collaborative model development without sharing raw patient data. However, it remains uncertain whether a decentralized model can achieve performance comparable to the centralized model, particularly in the context of rare disease diagnosis. Using two complementary datasets, GMDB, and AIDY, we evaluate the feasibility, performance, and clinical relevance of swarm learning against centralized and institution-specific local models. Model performance is assessed across in-distribution, cross-institutional, and ultra-rare disorder scenarios (not part of model training), with additional analyses of calibration and epistemic uncertainty. Our results show that swarm-trained models consistently match the accuracy of centralized training and outperformed local models. Importantly, swarm learning preserves sensitivity to low-prevalence and ultra-rare syndromes despite extreme data scarcity across sites, while exhibiting more conservative and reliable uncertainty calibration. Although swarm learning has previously been applied to well-characterized diseases, this study represents the first application of a swarm model in real-world settings involving a large and diverse set of disease classifications. Taken together, swarm learning emerges as a scalable, equitable, and trustworthy framework that shortens the diagnostic odyssey and advances precision medicine for rare disease diagnosis in routine clinical practice ### Competing Interest Statement The authors have declared no competing interest. ### Funding Statement This study did not receive any funding ### Author Declarations I confirm all relevant ethical guidelines have been followed, and any necessary IRB and/or ethics committee approvals have been obtained. Yes The details of the IRB/oversight body that provided approval or exemption for the research described are given below: This research was approved by the Institutional Review Board of GestaltMatcher Data Base (GMDB) and Artificial Intelligence for Dysmorphology (AIDY) I confirm that all necessary patient/participant consent has been obtained and the appropriate institutional forms have been archived, and that any patient/participant/sample identifiers included were not known to anyone (e.g., hospital staff, patients or participants themselves) outside the research group so cannot be used to identify individuals. Yes I understand that all clinical trials and any other prospective interventional studies must be registered with an ICMJE-approved registry, such as ClinicalTrials.gov. I confirm that any such study reported in the manuscript has been registered and the trial registration ID is provided (note: if posting a prospective study registered retrospectively, please provide a statement in the trial ID field explaining why the study was not registered in advance). Yes I have followed all appropriate research reporting guidelines, such as any relevant EQUATOR Network research reporting checklist(s) and other pertinent material, if applicable. Yes The data used in this study are not available for public access due to hospital confidentiality agreements, in compliance with data protection regulations.

Most patients with a rare movement disorder (MD) do not receive a molecular diagnosis, and the underlying genetic variants and mediating genes remain elusive. Here, we evaluate the diagnostic accuracy of conventional and next-generation sequencing-based genetic testing strategies in a cohort of 2,811 individuals with ataxia, spastic paraplegia and dystonia. Exome sequencing establishes genetic diagnoses in 19.3% of cases, and specificity of phenotypic features and age at testing are positive predictors. Genome analysis 'beyond the exome' increases the diagnostic yield by 7.5%, mostly due to the improved detection of structural variants and repeat expansions. Unsolved cases are included in the Solve-RD cohort and subjected to gene-burden analysis, providing evidence for loss-of-function variants in X-chromosomal CD99L2 causing spastic ataxia. Cellular studies show that the transmembrane protein CD99L2 occurs mainly in a ubiquitinated form and serves as an activating interactor of the calcium-dependent protease CAPN1. Ablation of cytoplasmic or extracellular domains of CD99L2 leads to its intracellular mislocalization and abrogation of its interplay with CAPN1. Transcriptome analysis in CD99L2 patient-derived fibroblasts reveals synaptic function-specific disturbances. Impaired CAPN1 activation and dysregulation of downstream neuronal pathways constitute the likely molecular cause for neurodegeneration.

FaceMesh2HPO is a framework for classifying facial phenotypic descriptors aligned with the Human Phenotype Ontology (HPO) to support clinical diagnosis. Using annotations from 124 clinicians across 10 disorders (107 HPO terms) combined with non-syndromic controls, we generated 3D facial meshes (478 landmarks) from 2D images and trained a hierarchical PointNet-based pipeline with cascading classification and feature elimination. The best models, incorporating 3D meshes, facial outline, and demographic metadata, achieved AUROCs between 0.55 and 0.89, with higher performance at parent nodes than leaf terms. External validation showed variable generalizability across disorders. Results demonstrate that hierarchical modeling of 3D facial geometry enables interpretable, ontology-linked phenotype classification, though performance on rare leaf terms remains limited. Improved data diversity and feature selection strategies are needed to enhance robustness and clinical utility.

AI-assisted facial phenotyping supports rare genetic disorder prioritization by retrieving visually similar diagnosed cases from facial image reference databases such as the GestaltMatcher Database (GMDB). Existing GestaltMatcher-based retrieval frameworks compare each test image with individual gallery images in a facial phenotype embedding space. However, this pointwise formulation does not fully exploit available evidence, because patients may have multiple images and disorders may be represented by multiple diagnosed gallery patients. We propose an inference-time multi-level evidence aggregation framework that improves facial phenotype retrieval without modifying the underlying GestaltMatcher-Arc encoder. The framework combines embedding-level patient aggregation of multiple images from the same individual, patient-weighted disorder centroids, and hybrid individual-centroid scoring to integrate test-patient observations, disorder-level gallery evidence, and local nearest-neighbor evidence. We evaluated the approach on GMDB v1.1.4 across disorders represented during training (GMDB-Freq), unseen disorders (GMDB-Rare), and multi-image patient subsets, using a unified gallery containing both GMDB-Freq and GMDB-Rare disorders. Multi-level evidence aggregation improved mean per-disorder top-N retrieval accuracy across all evaluation subsets. Top-1 accuracy increased from 38.52 These findings show that inference-time aggregation can improve next-generation facial phenotype retrieval without retraining the encoder, supporting a shift from isolated single-image matching toward multi-level aggregation of patient and disorder evidence for rare-disorder prioritization.

Background and Objectives Next-generation phenotyping (NGP) tools, such as GestaltMatcher, have revolutionized the diagnosis of rare genetic disorders through computational facial analysis. While NGP has been widely integrated into differential diagnosis workflows, its application in variant reclassification within the ACMG framework remains underexplored.Methods We applied GestaltMatcher to a 4-year-old patient with an undiagnosed neurodevelopmental disorder, suspected Mowat-Wilson syndrome (MWS), and a de novo ZEB2 variant. In addition to facial image analysis, we used the PEDIA framework, integrating Human Phenotype Ontology (HPO) terms and simulated exome data to refine variant prioritization. Bayesian likelihood modeling was used to establish Gestalt score thresholds for PP4 evidence levels (supporting, moderate, strong, and very strong). Brain MRI analysis was also performed to assess structural abnormalities characteristic of MWS.Results GestaltMatcher ranked MWS as the top differential diagnosis, and PEDIA integration further confirmed ZEB2 as the most likely disease-causing gene. Three of the patient's 4 facial images met the PP4 moderate threshold, while one met PP4 supporting. MRI analysis revealed subtle corpus callosum thinning, consistent with MWS. In addition, an exploratory case of an infant with molecularly confirmed MWS demonstrated the capability of GestaltMatcher to prioritize the diagnosis solely based on infant facial features.Discussion This study highlights the potential of NGP-driven facial phenotyping and multimodal integration in dysmorphology. The results support the broader application of AI-assisted phenotyping to improve diagnostic accuracy, particularly in neurodevelopmental disorders with distinct facial features.

PACS1-related disorder (PACS1-RD), also known as Schuurs-Hoeijmakers syndrome, is a rare autosomal dominant neurodevelopmental disorder predominantly caused by the recurrent de novo c.607 C > T p.(Arg203Trp) gain-of-function variant. Although core clinical features have been delineated, systematic data on developmental milestones, growth parameters, and clinical variability remain limited. We assembled a series of 24 previously unreported, unrelated individuals with PACS1-RD and compared their clinical and molecular features with 84 individuals from the literature. Genome-wide DNA methylation profiling was performed on peripheral blood DNA using bisulfite sequencing, interrogating ~860,000 CpG sites. Our study expands the phenotypic spectrum of PACS1-RD by reporting median age at independent walking and first spoken words (both 24 months), cross-sectional growth parameters, and previously undescribed clinical features, including congenital kidney malformations (25%) and feeding difficulties (75%). Compared with the literature, our series showed a higher prevalence of cryptorchidism (77.8%), congenital heart defect (45.8%), and hypotonia (75%). Methylation analysis identified a specific episignature for PACS1-RD, consistently observed in individuals carrying either the canonical p.(Arg203Trp) or the non-recurrent p.(Arg203Gln) variant. This episignature further enabled PACS1-RD diagnosis in one unsolved individual initially suspected of Kabuki syndrome. These findings refine the clinical delineation of PACS1-RD and establish an episignature that will support diagnosis in unresolved neurodevelopmental disorders and guide pathogenicity assessment of non-recurrent PACS1 variants.

PURPOSE:Biallelic variants in the minor spliceosomal gene RNU4ATAC were successively identified in Taybi-Linder/Microcephalic osteodysplastic primordial dwarfism type I, Roifman, and Lowry-Wood syndromes, which are characterized by variable microcephaly, short stature, neurodevelopmental impairment, skeletal dysplasia, and immunodeficiency. Two-thirds of the reported individuals present with Taybi-Linder syndrome, the first-described and most severe form. METHODS:We collected clinical and molecular data from individuals with biallelic RNU4ATAC variants through various French and European networks and clinics to refine the phenotypic spectrum of RNU4ATAC-opathies. RESULTS:We enrolled 69 participants and identified 18 new pathogenic variants. We report a significant proportion of attenuated or atypical presentations, novel rare symptoms, and, unexpectedly, a broad spectrum of autoimmune or inflammatory manifestations, affecting nearly half of the participants. Integrating our data with the 109 published cases, we propose a novel classification based on the main manifestations, immunodeficiency, and microcephalic primordial dwarfism. Using computer-assisted facial analysis, we also demonstrated the existence of a specific dysmorphic pattern in RNU4ATAC-opathies that is distinct among some sub-syndromes. CONCLUSION:We present a large cohort of individuals with RNU4ATAC-opathies and expand the phenotypic spectrum to paucisymptomatic forms, indicating that these diseases are likely to remain underdiagnosed.

Abstract Evaluating hand and wrist radiographs is essential in pediatric endocrinology and clinical genetics, particularly for the assessment of suspected skeletal anomalies. In this study, we present Auto-Bone-Caliper , an automated system for the segmentation and length measurement of metacarpal and phalangeal (M&P) bones, trained and evaluated on public datasets comprising both normal and dysmorphic cases. We first introduce InstanceSAM, a two-stage framework that detects and segments all 19 M&P bones in pediatric hand radiographs, achieving Dice scores of 98.7% for normal bones and 95.0% for dysmorphic bones. We further develop and evaluate three methods for bone-length estimation, identifying a k -means–based approach as the most accurate, with relative errors of 2.2% for normal bones and 4.5% for dysmorphic bones. Our automated pipeline, Auto-Bone-Caliper , integrates InstanceSAM with the k -means–based length-estimation method. To enable scale-independent downstream analyses, we derive relative bone-length measures from the automated measurements. Using these relative measures, we statistically compare measurements obtained using Auto-Bone-Caliper on an independent dataset with a healthy reference catalog of normal bone morphologies, observing a high level of agreement (Wasserstein-1 distance = 0.012). Finally, we demonstrate a potential clinical use case of Auto-Bone-Caliper by obtaining relative metacarpophalangeal pattern profiles for three genetic conditions, namely Turner syndrome, achondroplasia, and pseudohypoparathyroidism. Our results highlight the potential of the Auto-Bone-Caliper to streamline and standardize M&P length measurement, providing an objective and reproducible tool suitable for clinical application.

The genetic liability to a complex phenotype can be assessed using polygenic risk scores (PRSs) and is calculated as the sum of genotypes weighted by effect-size estimates derived from summary statistics of genome-wide association study (GWAS) data. Due to different allele frequencies (AFs) and linkage disequilibrium (LD) patterns across populations, PRSs developed in one population drop drastically in predictive performance when transferred to another. One of the major factors contributing to AF and LD heterogeneity is genetic drift, which acts strongly during population bottlenecks and is influenced by the dominance of certain alleles. In particular, because causal variants on empirical data are typically not known, the presence of population-specific LD patterns will strongly affect the transferability of PRS models. In this work, we therefore conducted demographic simulations to investigate the influence of the dominance coefficient on the transferability of PRSs among European, African, and Asian populations. By modifying the length and size of the bottleneck leading to the split of Eurasian and African populations, we gain a deeper understanding of the underlying dynamics. Finally, we illustrate that in our simulations, PRS models that are adapted to the underlying dominance coefficient can substantially increase the prediction performance in out-of-target populations.

Zusammenfassung Die Blickdiagnose gehört zu den ältesten und faszinierendsten Fähigkeiten in der Medizin. Sie erlaubt es, allein durch den geübten Blick eine Diagnose zu stellen oder zumindest eine Verdachtsdiagnose zu formulieren. Bei seltenen genetischen Erkrankungen ist diese Art der Diagnose jedoch meist Dysmorphologen – Experten, die speziell auf das Erkennen subtiler und auf das Vorliegen eines Syndroms hindeutender Gesichtsmerkmale geschult sind, – vorbehalten. Viele Kinder- und Jugendärzte besitzen eine ausgeprägte Intuition dafür, ob ein Kind eine auffällige Fazies aufweist. Diese wird in den klinischen Aufzeichnungen im englischsprachigen Raum manchmal mit der Abkürzung „funny looking child“ (FLC) vermerkt – ein Ausdruck, der die Schwierigkeit beschreibt, eine visuell erkannte Auffälligkeit eindeutig zu benennen und zuzuordnen. Mit der Hilfe von künstlicher Intelligenz (KI), insbesondere durch das System GestaltMatcher (sog. Next Generation Phenotyping, NGP), und anschließender genetischer Diagnostik (Next Generation Sequencing, NGS) können Kinder- und Jugendärzte bei auffälliger Fazies des Kindes schneller als bisher zur Diagnose selbst sehr seltener genetischer Erkrankungen kommen. Dies verkürzt nicht nur die Zeit der Unsicherheit bis zur Diagnosestellung, sondern kann ggf. auch einen erheblichen therapeutischen und psychologischen Nutzen für die betroffenen Kinder und ihre Familien darstellen.

The existence of transgenerational effects of radiation exposure on the human germline remains controversial. Evidence for transgenerational biomarkers are of particular interest for populations, who have been exposed to higher than average levels of ionizing radiation (IR). This study investigated signatures of parental exposure to IR in offspring of former German radar operators and Chernobyl cleanup workers, focusing on clustered de novo mutations (cDNMs), defined as multiple de novo mutations (DNMs) within 20 bp. We recruited 110 offspring of former German radar operators, who were likely to have been exposed to IR (Radar cohort, exposure = 0-353 mGy), and reanalyzed sequencing data of 130 offspring of Chernobyl cleanup workers (CRU, exposure = 0-4080 mGy) from Yeager, et al. In addition, we analyzed whole genome trio data of 1275 offspring from unexposed families (Inova cohort). We observed on average 2.65 cDNMs (0.61 adjusted for the positive predictive value (PPV)) per offspring in the CRU cohort, 1.48 (0.34 PPV) in the Radar cohort and 0.88 (0.20 PPV) in the Inova cohort. Although under the condition that the proportion of true mutations is low in this analysis, this represented a significant increase ([Formula: see text]) of cDNMs counts, that scaled with paternal exposure to IR ([Formula: see text]). Our findings corroborate that cDNMs are a potential transgenerational biomarker of paternal IR exposure.