Many pathogenic variants implicated in Mendelian diseases impair normal protein function, often through loss-of-function effects, while loss-of-function mutations in tumor suppressor genes commonly contribute to tumorigenesis. However, many disease-causing variants act through gain-of-function or other mechanisms that do not strictly disrupt the protein. Interpreting rare and novel variants remains a major challenge in clinical genomics, highlighting the need for computational tools informed by large, well-curated clinical datasets to reliably distinguish truly deleterious mutations from neutral variation. We developed MutAnt, a mutation meta‑annotator based on machine‑learning. It is trained on a large, clinically relevant dataset of variants using multiple variant properties, including synchronised predictions from other algorithms. MutAnt models demonstrate high F1 and ROC‑AUC scores (0.88-0.99) on hold‑out datasets and provide well‑calibrated probability scores that correlate with functional assays. MutAnt's deleteriousness predictions exhibited correlations with functional scores obtained from deep mutational scanning assays for tumor suppressor proteins BRCA1, PTEN, and p53 (ρ = 0.28-0.61), and with protein stability measurements from computational models. Moreover, MutAnt prediction scores of deleteriousness improved somatic variant calling from RNA sequencing data compared to standard approaches. MutAnt's high performance in distinguishing neutral and protein-disrupting mutations highlights its potential clinical utility in variant classification.

G-banded chromosome analysis, also known as G-banded karyotyping, remains a fundamental and irreplaceable diagnostic modality in clinical genetic testing. G-banded karyotypes provide whole genome visualization through chromosome banding patterns at the single-cell resolution for the diagnosis of chromosomal disorders. However, this method is labor-intensive and requires specialized expertise to manually analyze and karyotype metaphase spreads. In recent years, artificial intelligence (AI) algorithms have been utilized to automate karyotyping and assist with chromosome analysis. Despite this progress, there is a scarcity of studies evaluating the utility of artificial intelligence-assisted (AI-assisted) karyotyping analysis in cytogenetics diagnostic laboratories. This study highlights promising applications of AI-assisted karyotyping analysis in a cytogenetics diagnostic laboratory through a combination of a literature review, our data, and experience from a retrospective cohort study. This study also discusses important considerations of the use of AI-assisted karyotyping analysis in a cytogenetic diagnostic laboratory and outlines a two-stage framework for its implementation into clinical workflows. This approach aims to utilize the accuracy and efficiency of AI-assisted karyotyping analysis, potentially benefiting personalized patient care and contributing to advancements in the health system.

Parkinson's disease is a progressive neurodegenerative disorder characterized by symptoms such as bradykinesia, resting tremors, and muscle rigidity. Although several disease-causing genes of juvenile Parkinson's disease have been reported, the underlying mechanism remains unclear. Here, we identified SCAMP5 as a novel disease-causing gene of Parkinson's disease in a consanguineous family with juvenile Parkinson's disease. Functional studies in PC12 cell lines revealed that SCAMP5 deficiency increased the level of α-synuclein protein and α-synuclein oligomers, leading to increased cell apoptosis and decreased dopamine secretion. SCAMP5 knockdown in SH-SY5Y cells reduces α-synuclein secretion via exosome. Expression of human wild-type SCAMP5 rescued these effects, whereas the R91W mutant SCAMP5 did not. Scamp5a knockout zebrafish showed Parkinson's disease-like phenotypes, including bradykinesia, loss of dopamine neurons and decreased dopamine content in the brain. Transcriptome analysis unveiled upregulated JNK signaling in scamp5a knockout zebrafish, contributing to neuronal apoptosis. Importantly, human SCAMP5 prevented both dopamine neuron loss and bradykinesia in scamp5a knockout zebrafish, suggesting its therapeutic potential in Parkinson's disease. Overall, our findings identify SCAMP5 as a novel disease-causing gene of Parkinson's disease and highlight its neuroprotective role, opening new avenues for Parkinson's disease treatment.

Genomic resources from Tibeto-Burman (TB)-speaking populations are underrepresented in human genome research, limiting the understanding of their evolutionary history and health-related genetic influences. We genotyped 95 individuals, including Baima and Amdo Tibetans from Jiuzhaigou on the eastern Qinghai-Xizang Plateau and Qiang from Mianyang Prefecture in Sichuan Province. These data were jointly analyzed with 1722 genomes from modern and ancient East Asian populations. Clustering patterns revealed by principal component analysis suggested that the Tibetan and Qiang populations formed three distinct genetic clines, which were supported by model-based ADMIXTURE and fineSTRUCTURE analyses, highlighting complex population histories and unique genetic clusters among the Qiang and Tibetan people. Shared genetic drift estimated via f/f-statistics revealed significant gene flow between the Qiang and Han groups, suggesting that interactions with geographically proximate Han populations likely drove genomic affinity. Comparisons among TB groups (Amdo, Baima, Ü-Tsang, and Qiang) revealed varying levels of genetic affinity with ancient populations, particularly those from the Qinghai-Xizang Plateau and Yellow River Basin. Identity-by-descent and runs of homozygosity analyses indicated the persistence of stable population structures over approximately 2700 years and revealed relative demographic similarities among culturally different Tibetan groups, characterized by smaller effective population sizes than Han groups. Twenty-five high-confidence regions under selection were identified in Tibetans through XP-EHH, PBS, and Fisher score statistics, whereas 28 regions were detected in Qiangs, most of which were first identified here. The Tibetan-specific selection signals included genes related to hypoxia adaptation (e.g., TNNI3K), whereas the Qiang populations presented selection related to skin pigmentation (e.g., SLC44A5) and alcohol metabolism. The results of functional enrichment analyses suggested that the shared and distinct adaptations among these populations involved cardiovascular, metabolic, and immune processes. Overall, our findings reveal the complex genetic structure, population history, and evolutionary adaptations of Tibetan and Qiang populations in northern Sichuan. The results emphasize the role of geographic and historical factors in shaping genetic diversity and adaptive traits, contributing to our understanding of human adaptation to high-altitude environments and UV radiation in East Asia.

In the last decade, substantial research efforts have started worldwide to foster the clinical translation of Polygenic Risk Scores (PRS). Understanding the views of key relevant groups becomes timely to critically inform the socio-ethical debate, impact future health policy, and support the development of guidelines for best practices in healthcare contexts. We performed 26 in-depth semi-structured interviews to investigate the perspectives of European researchers and healthcare providers from different specialties (clinical genetics, oncology, cardiology, psychiatry) on the ethical and social implications of PRS uses in healthcare contexts. Findings were conceptualized in four main themes: 1) appropriate clinical use, highlights that PRS should be considered complementary tools aimed at informing a clinical intervention, with notions of appropriateness differing according to clinical goals and condition-type; 2) clinical utility: what's the evidence? captures participants' orientations towards the capability of PRS to improve health outcomes compared to standard care, as well as the barriers, limitations, or emerging areas of utility; 3) balancing risk and responsibility: navigating ethical questions in patient care, addresses classical issues in clinical genetics, including communication and counselling, potential patient harms, relevance of PRS information to family members, and the use of PRS in pediatric settings; 4) searching for standards: clinical guidelines, gathers perspectives on the potential format and content of future clinical guidelines, relevant parties, and contexts of applicability. In conclusion, the present study outlines a framework to define the range of responsible uses in healthcare contexts; however, societal and public health considerations, including priority-setting in national healthcare systems, need to follow for a comprehensive, and contextual, evaluation of PRS.

Degraded samples pose a challenge in routine forensic practice. The commonly used short tandem repeat markers are not optimally suitable for the analysis of degraded samples because of their structural complexity and locus length. By contrast, single nucleotide polymorphisms (SNPs), characterised by their single base mutation feature, enable the design of short amplification fragments, conferring an advantage in detecting mutations in degraded samples. Hence, our team has developed a multiplex amplification system for individual identification of degraded samples, encompassing 507 autosomal SNP loci, five Y-InDel loci, and one amelogenin sex determination locus. The amplification fragment lengths in this multiplex system range from 81 to 116 bp. The forensic applicability of this panel was validated through sequencing analysis of 201 samples. Among these were 30 degraded samples (simulated degraded samples: heat degradation and ultrasonic fragmentation; formalin-fixed, paraffin-embedded samples). The results indicated that the genotyping accuracy of all loci included in this panel remained at 100% for samples with 0.1 ng of DNA input and severe degradation. Sensitivity experiments revealed that at a DNA input of only 31.25 pg, the locus detection rate reached 100%, with genotype accuracy of 92.5%. Based on population data analysis, the total discrimination power of this system reached 1-5.513 × 10. Furthermore, we included forensic case samples encompassing semen, saliva, menstrual blood, 10-year-old bloodstain cards, and 12-year-old bloodstain cards in the validation studies. The results demonstrated 100% genotyping accuracy across all sample types. Additional, validation data confirmed the system's species specificity (Homo sapiens-specific) and tolerance to inhibitors including humic acid, heme, ethylene diamine tetraacetic acid and indigo (up to 200 µM). In conclusion, this system can serve as a novel tool for the analysis of degraded samples in forensic work.

Microtia-anotia is a common congenital anomaly. In most cases, the genetic etiology remains unknown. The proper development of outer ear is closely related to cranial neural crest cells. Abnormal DNA recombination perturbing the function of long-range enhancers can lead to genomic disorder. Previously, we identified 4p16.1 duplications in microtia patients and revealed the enhancer function of an evolutionarily conserved region (ECR). Here we recruited additional patients and attempted to identify the minimal overlapping region and regulatory elements. We identified five individuals (F6-F10 probands) with 4p16.1 duplication. The duplications in F3 and F5 were refined to 192.6 kb and 96.1 kb. Precise junction breakpoints in F4 and F6-F10 were detected. The minimal overlapping region (chr4: 8,689,510-8712,827, hg19) contained conserved sequences in addition to ECR. Dual-luciferase assays detected enhancer activity in the TFAP2C binding and 1794 sequence. We present five additional cases of concha-type microtia with 4p16.1 duplication. The minimal overlapping region contains regulatory elements that function as in-cis tissue-specific modules, regulating downstream gene expression during development of cranial neural crest cell.

Previous studies on Y-chromosomal haplogroup diversity in Poland have been focused mainly on macro-haplogroups. Consequently, younger subclades have rarely been explored to elucidate the relatively recent history of the Polish population. Here we present the results of deep genotyping of 598 chromosome Y sequences from modern Poland and demonstrate that about 60% of Polish males can be assigned to subhaplogroups that are both relatively young and widely distributed among different Slavic populations, thus supporting the scenario in which Early Slavic mass migration and territorial expansion took place in the first millennium of the common era. While most of those young Slavic-associated subclades are part of haplogroup R1a, other haplogroups, including I2a, R1b and E1b, are also represented by specific subclades, which together may constitute an important clue when trying to identify the location of the Proto-Slavic homeland based on ancient DNA data. Additionally, we have identified two specifically Polish subclades (I-Y6343 and R-Z17913, from haplogroups I1a and R1b, respectively) that likely descend from Late Ancient or Early Medieval founders representing the local Pre-Slavic population of the Roman period.

This study aims to assess the genetic burden of fetal congenital diaphragmatic hernia (CDH) and identify prenatal, perinatal, and postnatal predictors to improve early diagnosis, monitoring, and intervention. This study included 130 CDH fetuses who underwent invasive prenatal diagnosis, with fetal prognosis evaluated using imaging parameters such as observed-to-expected lung-to-head ratio (o/e LHR), observed-to-expected total lung volume (o/e TLV), and percent predicted lung volume (PPLV). Clinical outcomes included neonatal outcomes, extracorporeal membrane oxygenation (ECMO) requirement, and post-neonatal prognosis. Logistic regression and receiver operating characteristic (ROC) curve analyses were used to evaluate prognostic indicators and construct predictive models. Chromosomal microarray analysis (CMA) and exome sequencing (ES) yielded diagnostic rates of 7.7% and 8.7%, respectively, identifying a wide spectrum of pathogenic variants and highlighting the genetic heterogeneity of CDH. Among imaging parameters, o/e LHR, o/e TLV, and PPLV were significantly associated with neonatal outcomes, ECMO requirement, and post-neonatal prognosis. Multivariable models incorporating these parameters achieved high predictive accuracy (AUCs > 0.85), with the neonatal outcomes model reaching an AUC of 0.929, sensitivity of 93.2%, and specificity of 78.6%. By integrating genetic, imaging and clinical outcome data, this study identified CMA and ES as key tools for detecting genetic burden in CDH fetuses, and confirmed o/e LHR, o/e TLV, PPLV, and liver herniation as reliable prognostic indicators. Multivariable models based on these parameters showed strong predictive performance. A combined genetic-imaging approach is recommended to support individualized risk assessment and guide perinatal management.

Within-family genome-wide association studies (GWAS) can separate direct genetic effects from non-direct genetic biases introduced by analyses based on unrelated individuals, yet evidence regarding metabolic phenotypes remains sparse. Here, we aim to uncover non-direct genetic effects for metabolic traits and the role of diet in the non-direct genetic mechanism. We conducted family-based GWAS studies on six metabolic traits using data from full siblings (N = 777) and parent-offspring trios (N = 386). We calculated and compared within-family and population-based polygenic score (PGS) associations to identify non-direct genetic effects. Additionally, we assessed the parental indirect genetic effects of diet on offspring's metabolic traits. Within-sibship GWAS analyses were also conducted to evaluate the impact of non-direct genetic effects at the individual variant level. On average, the magnitudes of within-family PGS associations for metabolic traits showed a 35.2% reduction compared to population-based estimates, suggesting the presence of non-direct genetic effects. This discrepancy diminished after accounting for dietary score, indicating that diet is a major source of non-direct genetic effects. Additionally, parental indirect genetic effects of diet were revealed in parent-offspring models. For instance, PGS of parental fat consumption was positively related to the child's blood glucose levels (β: 0.44, 95% CI 0.21-0.67). After excluding non-direct genetic effects, within-sibship GWAS models are more effective at identifying functional genes associated with metabolic traits. Our study showed significant contributions of non-direct genetic effects on metabolic traits and also identified diet as a major source of non-direct genetic effects. These findings underlined the importance of family-based GWAS data in disentangling the genetic effects and gene-environment correlations underlying metabolic traits.

Genome Language Models (GLMs) represent a transformative convergence of artificial intelligence (AI) and genomics, offering unprecedented capabilities for biological discovery, healthcare innovation, and therapeutic design applications. However, these powerful tools create novel regulatory challenges that existing frameworks-whether AI governance or genomic privacy protections-cannot adequately address alone. This paper examines the critical regulatory gaps emerging at this intersection, highlighting tensions between AI principles that favor broad data access and genomic governance that demands stringent privacy protections and informed consent. We analyze how GLMs challenge conventional regulatory approaches as they pertain to applications in disease risk prediction, international research collaboration, and open-source model distribution. We propose a multilayered governance framework that combines policy innovations such as regulatory sandboxes and certification frameworks with technical solutions for privacy preservation and model interpretability. By developing adaptive governance strategies that bridge AI and genomic regulation, we can enable responsible GLM innovation while safeguarding individual rights, promoting equity, and addressing emerging biosecurity concerns in this rapidly evolving field.

Recessive variants in TWNK cause syndromes arising from mitochondrial DNA (mtDNA) depletion. Hearing loss is the most prevalent manifestation in individuals with these disorders. However, the clinical and pathophysiological features have not been fully elucidated. In this study, we collected five cases of hearing loss carrying bi-allelic TWNK variants from three unrelated Chinese families and identified two cases with isolated auditory neuropathy (AN) and three cases segregating with Perrault syndrome, characterized by AN, global developmental delay, and ovarian dysgenesis in females. All patients with cochlear implantation (CI) show poor speech discrimination outcomes, suggesting that the defect involves post-synaptic sites. In the mouse inner ear, Twinkle was immunolocalized to inner phalangeal cells and spiral ganglion neurons. Additionally, the broad expression pattern of Twinkle was observed in the auditory cortex, which to some extent explains the poor rehabilitation outcomes following CI. At the cellular level, Twinkle is localized at the mtDNA membrane, and the p.(Arg609AlaTer6) variant prevents the protein from reaching the mtDNA while the p.(Arg65Trp) variant exhibits a similar localization to the wild type, indicating a second mechanism of action. RT-PCR results indicated that the canonical transcript was abundant in the inner ear, while the shorter transcript was more abundant in the brain. Our findings revealed that bi-allelic TWNK variants lead to AN, which can be either syndromic or non-syndromic, with the molecular pathogenesis involving defects in mtDNA replication at post-synaptic sites. Patients with TWNK-associated conditions are not ideal candidates for CI and gene therapy may offer a solution for hearingrehabilitation.

A large proportion of individuals with celiac disease (CeD) remain undiagnosed, often presenting at an older age of onset or with non-classical symptoms compared to diagnosed cases. Such heterogeneity might be related to genetic factors. The aim was to utilize a CeD-screened adult population to compare the genetic variants in known and newly diagnosed cases. In the fourth wave of the population-based Trøndelag Health Study (HUNT4) 826 CeD and 51,516 non-CeD individuals were included. Medical registries identified 361 previously diagnosed cases, while screening identified 465 new cases. A validated polygenic risk score (PRS) was used to assess the genetic risk of CeD among the two case groups versus non-CeD individuals. Additional genetic variants not included in the PRS were also analyzed. The PRS distinguished cases from non-cases with high accuracy (AUROC: 85% for known cases, 83% for new cases). The genetic variation explained by the PRS was similar for known and new cases (17.1% versus 14.5%). The odds ratio for being in the highest genetic risk group (top 10%) was 22.7 (95% CI 14.1-36.4) for known cases and 18.6 (95% CI 12.4-27.9) for new cases versus the median group (40%-60%). Differences in effect size among specific genome-wide variants were observed but were not significantly associated with CeD. A validated PRS showed significant genetic difference between CeD cases and the general population, with similar association in both known and newly diagnosed cases. This suggests that genetic architectures of the two groups are comparable, implying that other non-genetic factors may drive CeD in adults.

Despite advances in the genetic diagnosis of hearing loss, there remains room for improvement. One way to improve the genetic diagnostic rate is the proper assessment of synonymous variants that are often bioinformatically filtered out. We used GSDME as a model to demonstrate the importance of assessing synonymous variants. Variants in the gene GSDME (also known as DFNA5) are associated with autosomal dominant nonsyndromic hearing loss. The hearing loss is typically progressive and downsloping. All reported causative variants of GSDME-related hearing loss involve the skipping of exon 8, which results in the expression of a constitutively active, but truncated protein that induces apoptosis of cochlear hair cells. A retrospective search of previously tested patients identified 3 novel pathogenic synonymous GSDME variants. The functional impact of these variants was confirmed in vitro via a minigene splicing assay. We also observed variant-dependent differences in the levels of aberrant splicing, leading us to hypothesize that partial loss of splicing will result in a less severe hearing loss phenotype as compared to complete loss of splicing. Audiometric analysis found an association between complete loss of splicing and greater initial and/or more quickly progressing hearing loss as compared to partial loss of splicing. Over the course of the study, we also found limited correlation between in silico prediction and in vitro observed effects of a variant on splicing, indicating the need to cautiously apply in silico prediction tools in the context of genetic diagnosis.

Inflammatory bowel diseases (IBDs) are chronic inflammatory disorders influenced by environmental factors and characterised by a dysregulated immune response. DNA methylation (DNAm) a key epigenetic mechanism plays a role in the etiology of complex diseases like IBD. Epigenetic clocks which estimate biological aging through DNAm patterns have also been linked to various health states, including IBD. Previously, we profiled DNA methylation in peripheral blood from adult IBD patients and controls using the Illumina 450K microarray (n = 184). We now expand this dataset with 8-year clinical follow-up data, including disease progression and treatment response. Additionally, we generate second and third-generation epigenetic clock measures in this cohort to investigate if IBD patients exhibit epigenetic age acceleration compared to healthy controls. We identified one CpG site (cg03583111) significantly differentially methylated in IBD patients with long-term clinical recurrence (after the first year of study) compared to non-recurrence (no treatment escalation after 8 years). We assessed DNAm aging signatures in IBD patients versus controls, finding evidence of significant epigenetic age acceleration, as measured by three epigenetic clocks (GrimAge, GrimAge2, and DunedinPACE), in IBD patients compared to controls. These associations were replicated in two independent IBD cohorts: adult (GSE87648, n = 377) and paediatric (GSE112611, n = 238). Moreover, we observed higher age acceleration (GrimAge, U = 669, p = 0.003) and a faster pace of aging (DunedinPACE, t = 3.233, p = 0.002) in patients with active UC compared to inactive disease, but not for CD. These findings suggest that blood-based DNAm signatures could serve as biomarkers for detecting, monitoring, and classifying IBD.

Genome-wide genetic correlation studies have demonstrated widespread shared genetic architecture between complex traits, yet the impact of vertical pleiotropy on these genetic correlation estimates remains unclear. To address this, we propose the Horizontal and Vertical Pleiotropy (HVP) model, designed to disentangle horizontal from vertical pleiotropy effects. This approach provides unbiased genetic correlation estimates specifically attributed to horizontal pleiotropy. Through simulations, we verify that the HVP model corrects biases introduced by vertical pleiotropy-particularly the causal influence of exposure on outcomes-across various scenarios, improving the accuracy of heritability and genetic correlation estimates. Vertical pleiotropy biases genetic variances and covariances, influencing essential estimates such as SNP-based heritability and genetic correlation in traditional methods. By addressing these biases, the HVP model enhances accuracy in parameter estimation. Real data analysis shows that horizontal pleiotropy significantly contributes to genetic correlations between metabolic syndrome (MetS) and traits such as type 2 diabetes, C-reactive protein (CRP), sleep apnoea, and cholelithiasis, whereas vertical pleiotropy is more relevant between body mass index (BMI) and MetS, and MetS and cardiovascular diseases. These findings suggest that action on modifiable factors like lowering BMI may effectively reduce MetS risk, while CRP-though not causative-serves as a useful marker in risk prediction through horizontal pleiotropic genes. These results confirm the HVP model's relevance and utility in revealing the complex genetic architecture underlying traits such as metabolic syndrome, highlighting its potential to inform precision healthcare.