Genetics in Medicine

BackgroundPediatric neuromuscular diseases are genetically and clinically heterogeneous. A substantial proportion remain without a definitive genetic diagnosis despite available clinical molecular testing. RNA-sequencing (RNA-seq) can be used to complement genome or exome sequencing to elucidate or to identify the functional impact of variants of uncertain significance, but when manually analyzed is limited to candidate DNA variants or phenotype-driven gene lists. Open-source computational tools have been developed to systematically and unbiasedly analyze RNA-seq data for aberrant splicing, expression, or allelic imbalance. However, best use practice of these tools is yet to be established. MethodsTo assess the performance of selected tools, we collected RNA-seq from 97 previously diagnosed samples to establish a truth set for benchmarking. Pathogenic variants were categorized as: true positives with confirmed aberrant RNA events and true negatives with no transcriptomic effect. We assessed performance of eight commonly used tools for splicing, expression and allelic imbalance analysis. We then applied the optimal strategy to 74 undiagnosed RNA-seq samples to identify new candidate diagnoses. ResultsAcross 68 diagnosed probands with aberrant RNA events, tools correctly identified 28 diagnoses. Splicing analysis tools provided most of the findings, but allelic imbalance tools uniquely identified 4, underscoring their value. Conversely, the false positive rate was highest for the splice tools and lowest for expression analysis. Application of tools led to identification of candidate variants for only 9 out of 74 undiagnosed patients. ConclusionsInclusion of RNA-seq tools can expedite variant prioritization, characterization and interpretation in the diagnostic pipeline but remain complementary to manual analysis of loci where candidate variants were identified by DNA sequencing.

Introduction: BEACONS-NBS (Building Evidence and Collaboration for GenOmics in Nationwide Newborn Screening) is the first research study to integrate whole genome sequencing into newborn screening (NBS) across multiple U.S. states and territorial public health laboratory programs (PHLPs). We developed a list of conditions for screening. Methods: We designed inclusion criteria and assembled an initial condition list from published resources. The list was revised by clinical experts, molecular geneticists, genetic counselors, PHLPs, rare disease advocacy organizations, the BEACONS-NBS Community Advisory Board, and project leadership from the National Institutes of Health. For each condition, we provided a rationale for early detection, diagnostic signs or biomarkers, and treatments or surveillance strategies. Results: The BEACONS-NBS condition list includes 777 conditions associated with 743 genes, one copy number variant, and two aneuploidies and is larger than those used in other genomic NBS research studies in the U.S. and United Kingdom. Most conditions are inborn errors of immunity (37.2%), inherited metabolic disorders (18.7%), or endocrine conditions (18.1%). Nearly all conditions (93.3%) can be confirmed using a non-genetic test. Discussion: BEACONS-NBS has established a condition list for implementation across multiple state and territorial PHLPs, enabling the prospective evaluation of feasibility of population-wide genomic NBS.

BACKGROUNDGenetic variant curation, an important step in the implementation of Genomic Medicine, requires literature-guided comparison of variant prevalence in affected individuals versus healthy controls. This evidence is categorized as the PS4 evidence code by the AMP/ACMG variant interpretation guidelines and its manual extraction is a major bottleneck in clinical variant curation. This study aimed to evaluate whether reasoning-capable large language models (LLMs) can support guideline-constrained PS4 evidence extraction from literature. METHODSWe benchmarked five LLMs for publication-level variant detection and PS4-eligible proband counting under ACMG/AMP and ClinGen Variant Curation Expert Panel (VCEP) guidance using an expert-curated truth-set. We assembled an expert-curated truth-set of 281 publication-variant pairs from 275 peer-reviewed publications (58 genes and 128 variants). Five LLMs spanning frontier-scale, reasoning-optimized, and efficiency-oriented classes (Gemini 2.5 Pro, GPT-5, o3, o4-mini, and Claude Sonnet 4) were evaluated against this truth-set using identical inputs, a unified prompt template, and a schema-constrained JSON output format on two tasks: (1) determining whether a prespecified variant was correctly identified and (2) counting independent PS4-eligible probands under applicable guidance. Primary outcomes were Task 1 accuracy and Task 2 exact-count concordance (model PS4 count equals truth-set count). We also assessed prompt sensitivity, error modes, and output variability across models. RESULTSModels were able to detect the presence of a variant in a publication with high accuracy (93.6-97.9%). For PS4 case counting, exact-count concordance was highest for Gemini 2.5 Pro (91.1%) and GPT-5 (90.0%), followed by o3 (86.5%), o4-mini (79.4%), and Claude Sonnet 4 (73.0%). Most counting errors resulted from an inability of a model to correctly apply guidelines, including evaluating phenotype and family structure. Prompt refinements improved concordance for most models but reduced performance for Claude Sonnet 4, indicating model specific prompting may be warranted. CONCLUSIONSReasoning-capable LLMs can support automation of guideline-based PS4 evidence extraction, achieving high concordance with expert curation, but performance is model- and prompt-dependent and failures concentrate in applying guidelines. Our findings support a hybrid workflow for clinical use in which LLM outputs accelerate evidence extraction with expert escalation.

PurposeDiagnostic yield from exome and genome sequencing varies widely across studies. It remains unclear how much of this variation reflects patient-level factors (e.g., sex, clinical features, race/ethnicity, genetic ancestry) versus site-level practices such as sequencing modality or variant interpretation workflows. We aimed to quantify the contributions of these factors to diagnostic outcomes across five U.S. clinical sequencing sites. MethodsWe performed a cross-sectional analysis of 3,008 prenatal, neonatal, and pediatric cases from the NHGRI Clinical Sequencing Evidence-Generating Research (CSER) consortium (2017-2023). Clinical indications spanned neurodevelopmental, neurological, immunological, metabolic, craniofacial, skeletal, cardiac, prenatal, and oncologic presentations. Genetic ancestry was inferred from sequencing data, and variants were interpreted using ACMG/AMP guidelines to classify DNA-based diagnoses. Generalized linear mixed models were used to estimate associations between diagnostic yield and fixed effects (sex, prenatal status, isolated cancer, number of clinical indications, sequencing modality, race/ethnicity, and genetic ancestry), while modeling study site as a random effect to quantify between-site variation. ResultsThe overall diagnostic yield was 19.0%. Multiple clinical indications (OR=1.47, 95% CI 1.20-1.80, p<0.001) were associated with higher diagnostic yield, and male sex (OR=0.80, 95% CI 0.66-0.96, p=0.017) and prenatal status (OR=0.63, 95% CI 0.44-0.90, p=0.012) were associated with lower yield. Sequencing modality, race/ethnicity, genetic ancestry, and isolated cancer were not statistically significantly associated with diagnostic outcomes.. A model without fixed effects attributed [~]10% of variance in diagnostic yield to between-site differences. After adjusting for covariates, site-level variance decreased to 5.7%, indicating consistent variation across sites not explained by measured patient factors. ConclusionAcross five sites, patient-level clinical features influenced diagnostic yield, but substantial site-level variation remained even after adjustment. Differences in variant interpretation, or case-classification practices may contribute to this residual variability. Further efforts to increase consistency in exome- and genome-sequencing diagnostic workflows may help reduce inter-site differences.

BackgroundBreast cancer susceptibility gene testing (BCSG-testing) is expanding in relation to both eligibility for testing and number of genes included on testing panels. However, uncertainty remains regarding the most effective testing strategies for identifying clinically actionable germline pathogenic variants (gPVs) while balancing increased burden on breast and genetics clinical services. Patients and MethodsThe North Thames Mainstreaming of Breast Cancer Genetic Testing (NT-MBGT) programme piloted unselected breast cancer (BC) patient BCSG-testing via a clinician-light BRCA-DIRECT mainstreaming pathway. We present real-world evaluation of (i) gPV pick-up rates according to BC characteristics and (ii) operational feasibility, acceptability, and satisfaction with the BRCA-DIRECT expanded testing pathway. ResultsThe BRCA-DIRECT pathway successfully tested 3,517 newly-diagnosed BC patients within 14 National Health Service (NHS) breast oncology units, with high levels of patient and breast healthcare professional (HCP) satisfaction, and genetics HCPs reporting concomitant decrease in service referrals. The overall pick-up rate of gPVs was 4.7%. Current NHS eligibility criteria would have offered testing to 20.6% of patients and identified 49.2% of observed gPVs in high penetrance (HP)-BCSGs (BRCA1/BRCA2/PALB2) and 18.2% of gPVs in intermediate penetrance (IP)-BCSGs (CHEK2/ATM/RAD51C/RAD51D). Ultra-simple eligibility criteria could improve detection (sensitivity) to 74.6% and 61.4%, respectively, whilst increasing testing to 50.2% of BC cases. ConclusionsEvidence from the NT-MBGT programme demonstrates that expanding BCSG-testing via a clinician-light pathway is acceptable and feasible, without increasing the burden on limited breast and genetics workforce, and has high satisfaction. Simplified testing criteria could improve identification of gPVs in HP-BCSGs. The concomitant increased pick-up of gPVs in IP-BCSGs warrants further consideration. highlightsO_LIIn this real-world evaluation we observed the successful rollout of the BRCA-DIRECT streamlined, clinician-light mainstreaming pathway for a pilot of germline breast cancer susceptibility gene testing in 3517 unselected breast cancer patients from 14 regional breast oncology/surgical units. C_LIO_LIPatients undergoing testing via the pathway reported high levels of satisfaction and low decisional regret, with breast and genetics healthcare professionals highly recommending the pathway for mainstream testing. C_LIO_LIDifferences were observed between breast healthcare professionals preferring unselected breast cancer patient testing and genetics healthcare professionals preferring restriction to current national testing criteria due to broader concerns around equity of access to testing. C_LIO_LIWe identified that current national testing criteria would have missed identifying 50.8% of germline pathogenic variants in high-penetrance, clinically actionable genes, likely having implications for treatment and surgical decision-making in the breast cancer patients. C_LIO_LIWe evaluated the performance of two additional approaches for establishing testing eligibility criteria to understand how we could best balance maximising identification of germline pathogenic variants (sensitivity) whilst limiting (unnecessary) testing within the breast cancer patient population (specificity). C_LI

BackgroundGenetic skeletal disorders (GSDs) comprise a heterogeneous group of rare, predominantly monogenic conditions that are increasingly diagnosed through high-throughput sequencing. While gene discovery has progressed rapidly, interpretation of pathogenic and uncertain variants remains a major bottleneck, in part because their functional consequences are determined at the protein structure level. However, a systematic assessment of structural knowledge across GSD-associated genes is currently lacking. Here, we present a comprehensive protein structure-centric analysis of 674 protein-coding genes implicated in GSDs. MethodsWe integrated experimental structures, AlphaFold2 (AF2) models, multimeric states, protein-protein interactions, and ClinVar variant annotations. ResultsWe quantify experimental structural availability and sequence coverage, revealing that 37% of GSD proteins lack any experimental structure and that, among proteins with structures, sequence coverage is often incomplete. We show that AF2 models provide high-confidence structural information for a substantial subset of proteins lacking experimental data, but that model reliability strongly correlates with existing structural coverage. Analysis of multimeric assemblies and co-occurring partners demonstrates that many GSD proteins function as obligate multimers, highlighting the importance of interface-level interpretation of variants. Finally, mapping clinically annotated missense variants onto representative protein structures illustrates how structural context can inform the interpretation of pathogenic and uncertain variants, particularly at interaction interfaces. ConclusionsTogether, this work provides a structure-aware reference framework for GSD genes, highlighting systematic gaps in current protein knowledge and demonstrating how integration of structural data can guide genomic variant interpretation. Our observations support a broader principle of structural equivalence, whereby distinct variants converge on shared structural perturbations that explain clustering patterns and enable mechanistic interpretation of nearby variants of uncertain significance.

PurposeGenetic diseases often present and are first diagnosed in the neonatal intensive care unit (NICU). Accurate identification of neonates with genetic diagnoses (GDs) in electronic health records (EHR) would enable a more complete understanding of their phenotypic spectrum, advancing care and personalized medicine. Prior research has used International Classification of Diseases (ICD) billing codes as proxies for GDs, though their accuracy for detecting confirmed GDs is uncertain. We evaluate the ICD codes for neonates with confirmed GDs and compare ICD billing code patterns between neonates with and without GD in two independent NICU cohorts. MethodsRetrospective analysis of patients admitted to the Boston Childrens Hospital (BCH) level IV NICU (1,344 neonates) and Nationwide Childrens Hospital (NCH)s neonatal network (33,315 neonates, mixed Level III/IV). For both cohorts, GDs captured by phecodes, aggregates of ICD codes, were compared with confirmed GDs. Two separate phenome-wide association studies (PheWAS) compared phecode patterns between neonates with GDs and those without, adjusting for sex, age at admission, gestational age, and NICU length of stay. ResultsGenetic phecodes were able to correctly identify 43.5% of neonates that received a GD in the BCH or NCH NICUs. Among 719 individuals with two or more genetic phecodes at BCH or NCH, 566 (78.72%) had a true GD. The BCH PheWAS analysis revealed a statistically significant positive association with atrioventricular septal defects and a negative association with bronchopulmonary dysplasia. The NCH pheWAS revealed 179 significantly associated phecodes, including many congenital anomalies. ConclusionThe use of ICD codes to identify NICU infants with GDs is neither sensitive nor accurate, though phecode analysis demonstrated stronger accuracy than sensitivity. Our data highlight clinical features of NICU infants more commonly seen in those that receive a GD (congenital heart defects) and those that are not (BPD). Our results can help to better predict and identify NICU neonates that receive a GD.

Background: Artificial intelligence (AI)-driven variant prioritization has demonstrated substantial utility in expediting genetic diagnosis by ranking the most likely causative variants. While a variety of tools have been developed, few address the unique clinical and technical constraints in prenatal genetic diagnosis. Methods: We introduce Berrylyzer, a novel, end-to-end variant prioritization system applied to prenatal diagnosis.Inspired by clinician's reasoning process during variant interpretation, Berrylyzer applies a modular, stepwise scoring architecture that jointly integrates phenotypic and genomic evidence and delivers a ranked list of candidate variants, achieving high computational efficiency without compromising analytical rigor. Moreover, Berrylyzer natively supports both structured ontologies and free-text clinical narratives, enabling flexible integration into diverse clinical environments. Its performance was rigorously evaluated across two independent, real-world prenatal cohorts and benchmarked against three state-of-the-art methods: Xrare, Exomiser, and PhenIX. Results: Across the two datasets, Berrylyzer ranked 56.41% and 58.12% of diagnostic variants first, and achieved recall rates of 94.02% and 97.42% within top 20, respectively. Berrylyzer outperformed Xrare (85.19% and 87.08%), Exomiser (84.90% and 85.98%), and PhenIX (82.05% and 88.93%). Stratified analysis consistently demonstrated superior performance across diverse disease categories, inheritance patterns, and analytical strategies. Notably, Berrylyzer exhibited robustness regardless of phenotype forms, yielding comparable top 20 recall rates for free-text descriptions and standardized terminologies. Conclusion: Berrylyzer represents an accurate, interpretable, and computationally lightweight variant prioritization system for prenatal genetic diagnosis. The superior performance across heterogeneous diagnostic contexts enables it as a practical solution for seamless integration into clinical pipelines, thereby advancing precision medicine in prenatal settings.

PURPOSE To evaluate cancer risk, age-specific penetrance, and mortality associated with heterozygous pathogenic or likely pathogenic (P/LP) germline PALB2 variants identified through genomic ascertainment and to assess modification by family history of cancer. PATIENTS AND METHODS We conducted a case-control study in two large population-based adult cohorts: the UK Biobank (n=469,580) and Geisinger MyCode (n=167,050). Individuals with heterozygous PALB2 P/LP variants were identified via exome sequencing and compared with non-carriers. Cancer diagnoses and vital status were obtained from linked registry and electronic health record data. We used multivariable logistic regression to estimate odds ratios (ORs) for cancer outcomes and Cox proportional hazards models to estimate hazard ratios (HRs) for all-cause mortality. Age-specific cumulative incidence (penetrance) was estimated using Kaplan-Meier methods. Models were adjusted for birth year, sex (when applicable), smoking status, and body mass index; stratified analyses assessed modification by family history of cancer. RESULTS PALB2 P/LP variant prevalence was 1:571 in UK Biobank and 1:940 in MyCode, with the higher prevalence in the UK cohort driven by the PALB2 p.Trp1038Ter founder variant. Compared with non-carriers, heterozygotes had significantly increased odds of any cancer, female breast cancer, pancreatic cancer, and cancers of ill-defined or secondary sites in both cohorts (P < 0.01). Adjusted hazard ratios for any cancer and female breast cancer ranged from 1.7 to 3.6. All-cause mortality was increased among PALB2-heterozygotes (HR 1.61-1.67), and survival after cancer diagnosis was reduced. Family history further modified cancer risk. CONCLUSION Genomic ascertainment of PALB2-heterozygotes identifies elevated risk for multiple cancers and increased mortality, although risks were lower than estimates from familial ascertainment. These findings inform risk management for individuals identified through genomic screening.

Background: Large language models (LLMs) have been evaluated as tools to assist rare disease diagnosis, yet evidence on their accuracy remains fragmented. We conducted a systematic review and meta-analysis to synthesize the available evidence on the diagnostic performance of LLMs, identify sources of heterogeneity, and evaluate the current evidence base for clinical translation. Methods: We searched PubMed, Embase, Web of Science, Cochrane Library, arXiv, and medRxiv (January 2020-February 2026). Full-text articles and preprints were considered for inclusion. Eligible studies applied LLM-based systems to generate differential diagnoses for rare diseases and provided Recall@1 (R@1; proportion with the correct diagnosis ranked first). We pooled R@1 using Freeman-Tukey double arcsine transformation with DerSimonian-Laird random-effects models. Pre-specified subgroup analyses examined LLM knowledge augmentation strategy and input modality. Because both retained high residual heterogeneity, we conducted a post-hoc exploratory analysis of evaluation benchmark disease composition, mapping diseases from major benchmarks to Orphanet prevalence classifications. Risk of bias was assessed using a modified QUADAS-3 instrument. Findings: We identified 902 records, of which 564 were screened and 15 studies were eligible. These 15 studies contributed 19 system-dataset entries to the meta-analysis (total N=39,529 cases). The pooled R@1 was 43.3% (95% CI 35.1-51.6; I2=99.6%). Augmented LLM systems (agent-based reasoning, retrieval, or fine-tuning; k=8) achieved R@1 of 52.5% (42.0-62.9) versus 35.4% (30.6-40.4) for standalone LLMs (k=11; p=0.004). Post-hoc exploratory analysis indicated that evaluation benchmark disease composition was associated with differences in diagnostic performance: R@1 was lower on the Phenopacket Store dataset, which contained a higher proportion of ultra-rare diseases (52.8%; k=2), than on RareBench (29.3%; k=6) at 21.7% (18.2-25.5) versus 52.0% (40.7-63.2; p<0.001). All 19 system-dataset entries were assessed to be at high risk of bias, most commonly due to potential data leakage and limited reproducibility. No study provided prospective clinical validation. Interpretation: Diagnostic performance of LLM-based systems for rare diseases varied substantially across evaluation benchmarks. Post-hoc exploratory analysis indicated that performance was associated with benchmark disease composition. Performance was higher in benchmarks containing fewer ultra-rare diseases and in systems incorporating external knowledge at inference time. However, all included studies were at high risk of bias, and none reported prospective clinical validation. These findings highlight the need for prevalence-stratified evaluation benchmarks and independent prospective studies before clinical deployment. Funding: This work was supported in part by the National Institutes of Health Common Fund, grant 15-HG-0130 from the National Human Genome Research Institute, U01NS134349 from the National Institute of Neurological Disorders and Stroke, R00LM014429 from the National Library of Medicine, and the Potocsnak Center for Undiagnosed and Rare Disorders.

Objective The decline of the perinatal demise rate is slowing and demises are often unexplained. Significant research has been done regarding diagnostic yield and genetic causes of demise, but little is known about how Geneticist involvement impacts outcomes. The goal of the study was to evaluate post-mortem genetic testing practices and effects of the geneticists involvement. Methods Retrospective data from 111 perinatal demise cases was examined, including rates of prenatal genetic counseling, post-delivery genetics consult, genetic testing, and autopsy investigation. Results In this cohort 54% received genetic testing and 25% received a genetics consult. When compared to those without, cases with genetic specialist involvement were associated with significant increases in testing uptake (p=0.007), diagnostic yield (p<0.001), and patient education (p<0.001). Second trimester stillbirths and those with fewer ultrasound (US) abnormalities were less likely to receive genetic testing (both p values <0.001) and consults (p<0.001, p=0.020). Conclusion Though it was not possible to avoid ascertainment bias, this data demonstrates that geneticist involvement correlates with a higher rate of testing, greater diagnostic yield, and more thorough counseling. These findings underscore the importance of integrating genetics providers into perinatal postmortem healthcare teams.

Background: Saudi Arabia bears a disproportionate burden of autosomal recessive genetic disorders, driven by consanguineous marriage rates of 50 to 58% and elevated carrier frequencies for conditions such as sickle cell disease (carrier rate up to 25%), betathalassemia (12%), and spinal muscular atrophy (6%). The existing premarital screening program screens for only two conditions. We developed SafeGene, a computational platform that expands predictive genetic screening to 50+ conditions using region specific population genetics. Methods: SafeGene integrates five risk calculation engines: (1) Mendelian inheritance models for AR, AD, XR, and XD conditions; (2) Hardy Weinberg equilibrium based carrier probability estimation using Saudi, Gulf, and global databases; (3) a six level consanguinity coefficient calculator (F = 0 to 1/8) with risk amplification multipliers; (4) multifactorial polygenic risk models for 12 complex diseases; and (5) maternal age dependent trisomy risk curves. Built using React.js, Node.js/Express, and MongoDB with bilingual Arabic/English support. Results: The platform encompasses 50 genetic conditions across 12 categories. Validation against published Saudi data demonstrated concordance with observed disease frequencies. Economic modeling projects that expanding screening could prevent 2,800 to 4,200 affected births annually, yielding savings of SAR 1.2 to 2.8 billion ($320 to 746 million USD) per year. Conclusions: SafeGene represents a scalable, evidence-based digital health solution for comprehensive genetic screening addressing the unique population genetics of consanguineous Gulf societies. The platform is protected under pending patent applications in South Africa (CIPC) and Saudi Arabia (SAIP).

BackgroundMitochondrial diseases are the most common inherited metabolic disorders, characterized by pronounced clinical and genetic heterogeneity that complicates molecular diagnosis. Although DNA-based sequencing approaches have become standard in genetic testing, up to half of patients remain without a definitive diagnosis. RNA sequencing (RNA-seq) provides a complementary layer of evidence by revealing functional consequences of genetic variation, thereby improving diagnostic yield. MethodsWe performed RNA-seq on skin fibroblasts from 140 pediatric patients with suspected mitochondrial disease who remained genetically undiagnosed after whole exome sequencing (WES). Aberrant RNA expression and splicing were identified using the detection of RNA outliers pipeline (DROP). Based on WES findings, patients were stratified into a candidate group (n=28), in which RNA-seq evaluated the pathogenicity of WES-identified variants of uncertain significance and an unsolved group (n=112), in which RNA-seq was used to pinpoint candidate genes. In six cases where RNA-seq identified the aberrant RNA-event but WES did not detect the causative variants, whole genome sequencing (WGS) was performed. ResultsIntegrative RNA-seq, WES, and WGS analysis resulted in a genetic diagnosis in 25% of patients overall (20/28 [71%] in the candidate group; 15/112 [13%] in the unsolved group). Aberrant splicing explained most candidate-group diagnoses, including variants misclassified by in silico predictors such as SpliceAI. Fourteen percent of protein-truncating variants predicted to undergo nonsense-mediated decay (NMD) escaped degradation, highlighting the functional limits of current predictions. The variants identified in the unsolved cohort included synonymous, missense, deep intronic, near-splice-site variants, and large deletions. The most frequent amongst them was a recurrent synonymous East Asian founder mutation in ECHS1, accounting for seven cases. Interestingly, across 231 pathogenic variants associated with aberrant RNA phenotypes compiled from this study and prior reports, half were non-coding and half were coding variants. ConclusionRNA-seq substantially enhances molecular diagnosis in mitochondrial disease by exposing cryptic splicing, regulatory, and NMD-escape events invisible to DNA sequencing alone. These data advocate transcriptome analysis as an essential component of comprehensive genomic diagnostics in neuro-metabolic disease. Significance StatementMitochondrial diseases remain among the most challenging inherited metabolic disorders to diagnose, with nearly half of patients unresolved despite advanced DNA sequencing. By integrating transcriptome profiling into the diagnostic workflow, this study demonstrates that RNA sequencing can reveal pathogenic mechanisms invisible to exome or genome analysis, including cryptic splicing, regulatory variants, and transcripts that escape nonsense-mediated decay. The findings establish RNA-seq as a decisive bridge between genotype and phenotype, uncovering functional consequences of genetic variation and redefining molecular diagnostics for mitochondrial and other neuro-metabolic diseases.

BackgroundNewborn screening (NBS) has progressively expanded through technological innovations, from tandem mass spectrometry enabling expanded NBS (eNBS) to the prospect of genomic NBS (gNBS). While these developments promise earlier diagnosis and richer information, they also raise concerns regarding actionability, uncertainty, equity and psychosocial impact. As technological feasibility alone does not ensure public confidence, parental perspectives are central to evaluating future expansions. Using acceptability concept as an anticipatory lens, this study assessed parental views on NBS expansion in France, examining its determinants, distinguishing test modalities, and exploring whether genomics raises specific concerns. MethodsA nationwide cross-sectional survey (September 2022-February 2023) included 1,640 parents recruited postpartum in maternity wards and through an online quota panel. Acceptability of eNBS and gNBS, intermediate evaluative components, and sociodemographic characteristics were assessed. Analyses combined descriptive statistics, multivariable regression, and thematic analysis of free-text comments. ResultsSupport was very high for eNBS (93%) and remained high for gNBS (89%), with genetics mainly shifting responses from complete to partial acceptability. Affective attitude and perceived effectiveness were the strongest predictors of both outcomes, while ethical concerns distinguished assured from conditional support. Most parents prioritised minimising uncertain results, whereas a smaller subgroup accepted greater ambiguity. Foreign-born and single parents reported lower levels of complete acceptability, while health-sector workers and parents with rare-disease experience were more supportive. No independent association with the age of the youngest child was observed. ConclusionParental acceptability of eNBS and gNBS is high but nuanced, shaped primarily by anticipated health benefits, emotional orientation and tolerance for uncertainty, with trust and social distance modulating support. As genomic expansion progresses, implementation will require proportionate, culturally adapted information and clear governance, and should be informed by real-world evidence from pilots such as PERIGENOMED. Trial registrationClinicalTrials.gov, NCT06111456. Last verified: October 2023.

Introduction: Accurate estimation of disease prevalence is crucial for public health and therapeutic development, but traditional methods are often inaccurate. Genetic prevalence, which estimates the proportion of a population with a causal genotype, using allele frequencies from population data, offers an important alternative. Methods: We partnered with 18 Rare As One patient organizations to estimate genetic prevalence for 22 autosomal recessive conditions using population data from two releases of the Genome Aggregation Database (gnomAD). To standardize and democratize these analyses, we developed the Genetic Prevalence Estimator (GeniE), a publicly available tool, for accessible calculations. Results: Conservative carrier frequencies in gnomAD v4.1 ranged from 1/164 to 1/11,888. The median change in genetic prevalence frequency between v2.1 to v4.1 was 0.806. Partnership with patient advocacy groups provided critical real-world context that refined the interpretation of these estimates. Discussion: These findings highlight that genetic prevalence is not a static figure but a dynamic, evolving measure with important caveats that need to be considered. Our study underscores the necessity of re-evaluations as databases expand. By integrating patient-partnered insights with the GeniE platform, we empower the genomics community to maintain transparent, up-to-date, and actionable data for rare disease advocacy and drug development.

Despite the multidimensional value of implementing genomic medicine, in terms of diagnostic yield, cost-effectiveness, and optimisation of care trajectories, its deployment in many African countries, including Algeria, remains constrained by major structural and interpretive challenges, compounded by the persistent underrepresentation of African populations in genomic databases with direct consequences for variant interpretation and clinical decision-making. We implemented a fully in-house whole-exome sequencing (WES) workflow structured through a clinically driven sequential framework in 14 unrelated patients with unexplained neurodevelopmental disorders, in a context of high consanguinity and enriched recessive inheritance. A definitive molecular diagnosis was established in 8 cases, with pathogenic or likely pathogenic variants identified in MECP2, PTPN11, FOXG1, ARV1, GNAO1, ATM, ROBO3, and CHD3. Five cases yielded variants of uncertain significance and one clinically relevant incidental finding was identified. Beyond its diagnostic contribution, this study reveals persistent interpretive limitations: a disproportionate VUS burden, complex incidental finding management, and reduced accessibility to classification criteria, reflecting database underrepresentation, the predominance of private variants, and the limits of current frameworks in consanguineous settings. These findings underscore the necessity of population-specific reference datasets, iterative phenotyping, adapted ethical frameworks, and strategies addressing territorial disparities in access. This work demonstrates that WES implementation requires a structured multidisciplinary ecosystem integrating clinical, bioinformatic, and ethical dimensions, and provides a transferable model for the sustainable integration of genomic medicine in under-resourced settings, while highlighting the global scientific value of incorporating underrepresented populations into genomic research.

BackgroundKCNT1-related disorders are a rare, severe neurogenetic disorder associated with early-onset, treatment-resistant seizures and significant developmental comorbidities. Currently there are no treatment-modifying therapeutics for this condition, and the condition necessitates complex, lifelong care that places a profound financial strain on affected families and healthcare systems. However, data quantifying this economic burden is sparse. ObjectiveTo evaluate the annual cost burden of KCNT1-related disorders in the United States using both caregiver-reported expenditures and electronic medical record (EMR) data, providing a comprehensive analysis of direct, indirect, and out-of-pocket expenses. MethodsA retrospective cohort analysis was conducted using two complementary data sources. In 2025, 34 U.S-based. caregivers from the KCNT1 Epilepsy Foundation registry completed a survey capturing insurance status, medical and non-medical expenses, and indirect costs. Separately, EMR data from 49 U.S.-based patients with KCNT1 variants were extracted from the Citizen Health database. Clinical services were mapped to CPT and HCPCS codes, and costs were calculated using Medicare fee schedules and other publicly available datasets. ResultsCaregiver-reported data revealed that all respondents possessed some form of insurance coverage, primarily through private insurance purchased independently or through their employer, or Medicaid. Nearly half of respondents (18/34) experienced financial hardship, citing high out-of-pocket expenses, medical debt, and loss of income due to caregiving responsibilities, and twelve percent of respondents delayed treatment due to financial strain (n=4). The estimated mean total annual medical cost per family--including direct, indirect, non-medical, and non-covered expenses--ranged from $355,474 to $797,727, based on upper and lower bounds of response categories from 10 respondents. EMR analysis, which only reported on direct medical costs, revealed that average first-year direct medical costs reached $154,389 per patient based on the records from 49 patients. This cost was primarily driven by hospitalizations, medications, and therapeutic procedures. Based on EMR data, direct medical costs declined once the patients reached two years of age and stabilized in subsequent years. Hospitalizations remained the most substantial cost contributor regardless of the age of the patient. ConclusionKCNT1-related disorders imposes a substantial economic burden on families and healthcare systems, particularly in the first year after diagnosis. This study highlights the need for rapid diagnostic procedures, targeted therapies, improved insurance coverage, and legislative support for families managing rare, high-burden conditions. Findings provide essential cost data to support drug development, healthcare planning, and rare disease policy reform. SignificanceThis is the first U.S.-based study to quantify both medical and non-medical costs associated with KCNT1-related disorders using combined caregiver and EMR data. The results highlight the urgency of disease-modifying treatments and equitable access to care, informing clinical trials and advocacy for systemic healthcare support.

BackgroundMost rare coding variants in monogenic disease genes remain classified as Variants of Uncertain Significance (VUS), limiting their use in clinical care. Many variant classifications have been submitted to ClinVar, often with rich free-text summaries of the evidence underlying each classification. These narratives are not standardized and are difficult to mine systematically, making it challenging to identify variants that might be reclassified as new evidence becomes available. MethodsWe developed a two-stage language-model pipeline that (i) detects whether functional, population, or computational evidence is described in ClinVar and ClinGen variant summaries, and (ii) classifies whether it is evidence of pathogenicity or benignity. We first constructed Variant Evidence Text Annotations (VETA), a dataset of 44,522 ACMG/AMP keyword-description pairs derived from 18,678 ClinVar and ClinGen variant summaries using an LLM-based consensus annotation procedure. We then fine-tuned BioBERT-large models for each evidence type and stage, and validated performance using independent ClinGen expert-curated summaries as well as orthogonal variant-level evidence, including functional screening, computational scores, and population estimates of disease impact. ResultsAcross evidence types, our models accurately identify whether functional, population, and computational evidence is present and whether it leans toward a pathogenic or benign impact. We find high agreement with ClinGen expert annotations and highly significant separation of validation scores between model-predicted benign and pathogenic groups (functional assays p = 8.13 x 10-30, variant allele frequencies p = 4.11 x 10-22, computational predictions p < 8.88 x 10-16). We applied the full workflow to approximately 6,000 ClinVar VUS variants whose submission summaries lacked explicit functional or population evidence. By aggregating external functional, population, computational, and diagnostic evidence using the ACMG/AMP SVI point-based framework, we found that about 17% of these VUS meet quantitative thresholds for a likely benign or likely pathogenic classification, including 492 VUS in genes reviewed by ClinGen Variant Curation Expert Panels. ConclusionsTransforming unstructured variant summaries into a structured, evidence-type matrix enables scalable detection of evidence gaps, allowing for the systematic integration of new data sources, and prioritization of VUS that are most likely to be reclassified. This language model-enabled pipeline provides a generalizable digital approach to identify clinical evidence gaps as functional screens, biobank resources, and computational predictors continue to evolve.

Fanconi anemia (FA) is a rare genetic disorder of impaired DNA repair characterized by progressive bone marrow failure, congenital malformations, and cancer predisposition. Early identification of individuals with FA is critical for timely clinical management, yet phenotype-driven approaches to FA identification are hindered by inconsistencies in existing phenotypic profiles. We compared the Human Phenotype Ontology (HPO) annotations for FA in OMIM (215 terms across 22 complementation group entries) and Orphanet (106 terms in a single entry, ORPHA:84), quantifying overlap and anatomical system coverage. To address identified gaps, we developed a comprehensive custom HPO profile by extracting phenotypic terms from the entire Fanconi Cancer Foundation (FCF) Clinical Care Guidelines using OntoGPT, an LLM-based ontology extraction tool, followed by manual curation to ensure accuracy and clinical relevance. OMIM and Orphanet shared only 36 HPO terms (12.6% of their combined 285 unique terms), demonstrating substantial discordance. Our custom profile comprises 264 unique HPO terms, of which 161 (61.0%) are novel and not present in either existing source. The novel terms expand coverage particularly in musculoskeletal (39 terms, 23.8%), genitourinary (26 terms, 15.9%), limb (26 terms, 15.9%), head or neck (20 terms 12.2%), and digestive system (17 terms, 10.4%) phenotypes. Community-curated phenotypic profiles derived from clinical practice guidelines can substantially augment existing disease annotations. Our FA profile, the most comprehensive HPO-based phenotypic characterization of FA to date, is publicly available and provides a foundation for improved clinical decision support and EHR-based computable phenotyping that can accelerate diagnosis for individuals with FA. Furthermore, the LLM-assisted approach offers generalizable methods to improve the diagnostic odyssey for all rare diseases.

Purpose To systematically review population preferences for expanded carrier screening programmes to inform service delivery and health policy. Methods PubMed, CINAHL, and Scopus were searched from 1995 to 2025 on carrier screening for autosomal or X-linked recessive genetic conditions across adult general populations. Included studies elicited preferences on attributes regarding the design or delivery of carrier screening programs. We extracted preferences for each attribute, mapped qualitative findings to these preferences, assessed risk of bias and performed meta-analysis on the willingness-to-pay for screening using Bayesian multilevel modelling. All findings are reported in 2024 USD. Results Thirty one studies, including 16 quantitative, 11 qualitative, and 4 mixed-methods studies were included. Participants expressed preferences for which conditions to include in ECS, joint vs individual screening, the value of information provided before screening, in-person over online counselling, type of healthcare provider, and preconception testing. Willingness-to-pay was right-skewed with 9% of participants not willing to pay any amount, a median of $107 and an interquartile range between $41 and $226. Most studies demonstrated a high risk of bias. Conclusions We report preferences of the general population regarding expanded carrier screening programmes, including suggested amounts for copayment if subsidised by the health system.