Klee, Eric W.

Abstract Background Pathogenic variants in the DMD gene are associated with Duchenne and Becker muscular dystrophy (DMD/BMD). Professional medical societies currently recommend DMD carrier screening for all patients who are pregnant or considering pregnancy. Such testing may identify intragenic duplications of DMD (i.e., involving one or more exons but fully contained with the gene) in patients without a personal or family history of DMD/BMD. These incidentally detected DMD intragenic duplications, however, pose significant challenges for variant interpretation. Pathogenicity of an intragenic duplication depends on its structural configuration. Tandem duplications may cause reading frame disruption and gene loss-of-function; conversely, non-tandem duplications may not disrupt gene function. Current carrier screening assays cannot determine the structural configuration of an intragenic duplication. The current guidelines (e.g., the ClinGen PVS1 decision tree) presume intragenic duplications are in tandem unless proven otherwise. If this presumption of tandem configuration is inaccurate, it may lead to overestimation of variant pathogenicity, potentially resulting in unnecessary invasive procedures (e.g., amniocentesis for prenatal diagnosis) and/or having other negative impacts with regards to prenatal decision-making. This study aims to evaluate the validity of this presumption for incidentally detected DMD intragenic duplications. Methods We performed high-coverage whole-genome long-read sequencing on the PromethION instrument (Oxford Nanopore Technologies) for 15 cases with previously identified DMD intragenic duplications. Among these cases, four had personal or family history of DMD/BMD, while 11 had no known history (i.e., incidentally detected). The structural configurations of the duplications were determined by an in-house bioinformatics tool and manual inspection of the alignment files. Results In all 4 cases with a known personal or family history of DMD/BMD, the DMD intragenic duplications were confirmed to be in tandem, aligning with expectations of their pathogenicity. Interestingly, among the 11 cases with incidentally detected DMD intragenic duplications, seven (64%) were found to involve interspersed duplications. In five of these seven interspersed duplication cases, the involved DMD exons were co-duplicated with distinct segments of the Xp arm located 0.5–30 Mb away, forming complex structural variants. In another case, the duplicated DMD exons have inserted into the Xq arm, about 90 Mb downstream of DMD. The final case involved the duplicated DMD exons inserted into a repetitive region most consistent with 9q12 (i.e., the 9qh region). Conclusions Our study reveals that while DMD intragenic duplications in patients with a personal or family history of DMD/BMD were consistently in tandem, a substantial 64% of incidentally detected DMD intragenic duplications were not in tandem. These findings have important implications for both clinicians and laboratorians. They challenge the presumption of tandem configuration for DMD intragenic duplications in individuals where the variants were identified incidentally, e.g., through carrier screening in those without a family history of DMD/BMD. In these cases, we recommend classifying the DMD intragenic duplications as variants of uncertain significance (VUS), unless there is evidence of tandem configuration.

ABSTRACT Post‐translational modifications, such as glycosylation and phosphorylation, play a critical role in protein trafficking, interactions, and stability. Disruptions in these pathways can lead to glycosylphosphatidylinositol (GPI) deficiencies, which present with a spectrum of clinical features, including congenital anomalies, dysmorphic features, developmental delay, hypotonia, and epilepsy. Biallelic variants in PIGV , a key mannosyltransferase in GPI biosynthesis, cause Hyperphosphatasia with Impaired Intellectual Development Syndrome 1 (HPMRS1), a rare disorder characterized by hyperphosphatasia, seizures, developmental delay, hypotonia, abnormal MRI findings, and distinct facial dysmorphisms. Fewer than 30 cases have been reported to date. We conducted a case study and literature review; with clinical data obtained from medical records. A 2‐year‐old female presented with developmental delay, myoclonic epilepsy, hypotonia, and mild facial dysmorphism. Laboratory results showed elevated alkaline phosphatase (899 U/L) and low pyridoxal 5′‐phosphate levels in cerebrospinal fluid (CSF). Brain MRI revealed diffuse cerebral volume loss and ventriculomegaly. Trio genome sequencing identified two PIGV variants: c.1415T>C‐p.(Leu472Pro) and c.524T>C‐p.(Leu175Pro). Functional studies using PIGV ‐knockout HEK293 cells transfected with wild‐type or mutant PIGV constructs demonstrated reduced GPI‐anchored protein (GPI‐AP) expression under weaker promoters, indicating impaired enzymatic activity. We report two novel PIGV variants associated with HPMRS1, emphasizing the value of functional assays in variant interpretation. Our findings also highlight the diagnostic relevance of alkaline phosphatase measurement in patients with refractory seizures.

ABSTRACT Aims Common genetic variants have been linked to an increased risk of non-ischaemic cardiomyopathy (NICM). The primary objective of this study was to evaluate the significance of these variants in a general population. Methods and results Thirteen NICM risk alleles, previously validated by meta-analyses of case-control genetic association studies, were used to determine genetic risk (GR) in a large, unselected cohort (n = 1897) of Olmsted County residents aged >45 years. Clinical and echocardiographic data, with biomarkers, were correlated with this NICM-GR over a median of 16 years of follow-up (interquartile range [IQR] 15–17). There were no significant differences between the study population quartiles, based on NICM-GR, in terms of age (median 61.6 [IQR 53.6–70.4]), sex (48% male), and cardiovascular risk factors. A significant correlation was detected by specific linear-regression models between NICM-GR and left ventricular (LV) ejection-fraction (beta [standard error]: −2.68 [0.68], p < 0.001), LV end-systolic-diameter (0.22 [0.06], p < 0.001), and LV end-diastolic-diameter (0.14 [0.06], p = 0.01). During follow-up, 46 subjects developed NICM. Using a 3:1 matched-cohort of non-NICM versus NICM patients, linear-regression analysis found a significant correlation between NICM-GR and NICM development (odds ratio 2.27, 95% confidence interval 1.06–4.85, p = 0.03). There was no significant association between these variants and heart failure biomarkers (aldosterone, natriuretic peptides, ST2, galectin-3). Conclusions The NICM-GR correlates with LV function and dimension, and subsequent development of NICM in the general population. The lack of correlation of these variants with cardiovascular risk factors and the renin–angiotensin–aldosterone axis or natriuretic peptides suggest different biological mechanisms involved. These findings support the need to explore newer pathophysiologic pathways in the development of NICM.

Abstract Clonal hematopoiesis (CH) of indeterminate potential (CHIP), driven by somatic mutations in leukemia-associated genes, confers increased risk of hematologic malignancies, cardiovascular disease, and all-cause mortality. In blood of healthy individuals, small CH clones can expand over time to reach 2% variant allele frequency (VAF), the current threshold for CHIP. Nevertheless, reliable detection of low-VAF CHIP mutations is challenging, often relying on deep targeted sequencing. Here, we present UNISOM, a streamlined workflow for enhancing CHIP detection from whole-genome and whole-exome sequencing data that are underpowered, especially for low VAFs. UNISOM utilizes a meta-caller for variant detection, in couple with machine learning models which classify variants into CHIP, germline, and artifact. In whole-exome sequencing data, UNISOM recovered nearly 80% of the CHIP mutations identified via deep targeted sequencing in the same cohort. Applied to whole-genome sequencing data from Mayo Clinic Biobank, it recapitulated the patterns previously established in much larger cohorts, including the most frequently mutated CHIP genes and predominant mutation types and signatures, as well as strong associations of CHIP with age and smoking status. Notably, 30% of the identified CHIP mutations had < 5% VAFs, demonstrating its high sensitivity toward small mutant clones. This workflow is applicable to CHIP screening in population genomic studies. The UNISOM pipeline is freely available at https://github.com/shulanmayo/UNISOM and https://ngdc.cncb.ac.cn/biocode/tool/7816.

Abstract Deep learning has advanced digital pathology and multi-omics data analysis. However, there is still a critical need to integrate these modalities to better understand complex diseases. Deep learning models, such as Prov-Gigapath, a state-of-the-art pathology foundation model, have been developed for analyzing whole-slide images (WSI). Our previously published deep-learning framework, OmicsFootPrint (OFP), transforms high-dimensional multi-omics data into intuitive genomically-arranged circular images and utilizes transfer learning to overcome the challenges of high-dimensional, low-sample-size data. In this study, we aimed to combine WSI and multi-omics data to develop a framework capable of distinguishing between invasive ductal carcinoma (IDC) and invasive lobular carcinoma (ILC), which is clinically significant as it may help explain differential responses to therapies, such as tamoxifen versus letrozole. We enhanced the performance of our OFP model by dividing each multi-omics circular image (1024x1024x3 pixels) into 16 equal-sized patches (256x256x3 pixels). Each patch was trained using a separate EfficientNetV2 architecture with pre-trained ImageNet weights. The probability outcomes from these 16 models were subsequently combined through a meta-learner model using the AutoGluon framework. Concurrently, WSI images corresponding to the same patients were independently trained, maintaining consistent train/validation/test splits. The Prov-Gigapath model, which was trained using a dataset from the Providence Health System, was employed to extract patch-level embeddings from the WSI data. These embeddings were fine-tuned using a LongNet-based masked autoencoder to predict outcomes. Finally, predictions from the OFP model and the WSI-based Prov-Gigapath model were integrated through a meta-learner, preserving the original train/validation/test split. Model performances (AUC) were evaluated using the held-out test set. The gold standard used for the analysis was histological identification of IDC versus ILC, against which the performance of omics data alone, WSI data alone, and the multi-modal integration approach were compared. We evaluated our integrated model as a proof-of-concept using The Cancer Genome Atlas (TCGA) breast cancer cohort, comprising 92 ILC and 391 IDC samples. Multi-omics data, including gene expression, microRNA, copy number variation, and RPPA, were analyzed using the OFP framework, while WSI data were processed using the Prov-Gigapath method. The results demonstrated that the multi-modal integration approach significantly outperformed individual modalities, achieving an AUC of 0.98 compared to 0.87 for OFP (original), 0.95 for OFP (16-patches), and 0.92 for WSI. Integrating WSI and multi-omics data improves predictive accuracy, significantly improving the classification of breast cancer subtypes over histology alone. Citation Format: Naresh Prodduturi, Xiaojia Tang, Hamid R. Tizhoosh, Kevin J. Thompson, Richard Weinshilboum, Karthik V. Giridhar, Judy C. Boughey, Eric W. Klee, Liewei Wang, Matthew P. Goetz, Vera Suman, Krishna R. Kalari. A multi-modal learning framework to integrate digital pathology image and multi-omics data in breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2429.

Background and Aims: Metabolic dysfunction–associated steatotic liver disease affects about a third of adults worldwide and is projected soon to be the leading cause of liver cirrhosis. It occurs when fat accumulates in hepatocytes and can progress to metabolic dysfunction–associated steatohepatitis, liver cirrhosis, and HCC. Metabolic dysfunction–associated steatotic liver disease pathogenesis is believed to involve a combination of genetic and environmental risk factors. Single nucleotide polymorphisms have been implicated, but non-syndromic monogenic causes are lacking. Approach and Results: We identified a novel genetic variant in a familial case of metabolic dysfunction–associated steatohepatitis and performed deep variant functional analysis, including protein modeling, dynamics, and cell-based assays to assess molecular mechanisms of dysfunction and altered cellular signaling. We analyzed exome sequencing data of 3904 individuals with steatotic liver disease (SLD) to identify additional cases and establish the link between specific gene variants and SLD diagnosis. We discovered and functionally validated the NM_000245.4:c.3505A>T; p.(Ile1169Phe) variant in the MET (mesenchymal-epithelial transition) kinase domain as a monogenic cause of SLD. Subsequently, we detected additional ultra-rare, previously uninterpreted, and likely deleterious variants in MET from screening sequencing data. Among individuals with confirmed SLD based on electronic record review, 1.1% (45/3904) had rare predicted deleterious MET variants. Eight of 45 (17.7%) individuals had predicted deleterious variants in the MET kinase domain confirmed to be functionally like the familial case variant. Conclusions: We report the first germline nonmalignant rare MET -driven disease, a monogenic form of SLD.

DNA methylation is an essential epigenetic modification that plays a crucial role in regulating gene expression and maintaining genomic stability. With the advancement in sequencing technology, methylation studies have provided valuable insights into the diagnosis of rare diseases through the various identification of episignatures, epivariation, epioutliers, and allele-specific methylation. However, current methylation studies are not without limitations. This mini-review explores the current understanding of DNA methylation in rare diseases, highlighting the key mechanisms and diagnostic potential, and emphasizing the need for advanced methodologies and integrative approaches to enhance the understanding of disease progression and design more personable treatment for patients, given the nature of rare diseases.

Introduction: Identifying patients at risk for ischemic events after percutaneous coronary intervention (PCI) relies on traditional analysis of limited clinical and imaging variables. Machine learning (ML) has shown promise in effectively predicting cardiovascular risk in population studies. While existing ML models mainly predict mortality and incorporate clinical variables, there is a lack of tools that have utilized genetic data and that predict ischemic events. Aims: This study aims to develop and validate a ML model incorporating genotyping and clinical data to enhance prediction of ischemic outcomes for PCI patients utilizing large prospectively derived diverse datasets. Methods: Patients from the TAILOR-PCI trial (n=5302) were utilized for model development. 50% of the sample was utilized for Boruta feature selection and 50% for training and testing using cross validation. Features included demographics, medical history, medications, PCI characteristics, and genetic data (specifically, CYP2C19 *2, *3, *17 alleles). The primary endpoint was a composite of cardiovascular death, myocardial infarction, stroke, stent thrombosis, and severe recurrent ischemia at 12 months. Multiple ML classification algorithms, including Support Vector Machine (SVM) polynomial, Random Forest, Light gradient boost, XG Boost, among others, were benchmarked for their prediction performance on rare events. The top performing classifiers were externally validated on an independent dataset from the PRECISION PCI study (n=3,745). Results: Mean participant age of the training set was 64.2 ± 11.0 years, with 75.4% being male. During follow-up of 12 months, among 4,572 patients in the entire cohort 343 (7.5%) met the primary outcome. The SVM polynomial model demonstrated the highest area under the curve (AUC) of 0.67 for predicting the primary outcome with test dataset. The sensitivity, specificity, precision, and recall were 0.87, 0.28, 0.07, and 0.87 respectively (Figure). Peripheral arterial disease, body mass index, and age were among the top variables by feature importance. Conclusion: ML models incorporating both clinical and genetic data are feasible and highly promising in predicting major adverse cardiac events that may help guide use of anti-platelet drug therapy. The AUC values are reasonable given imbalances and misclassifications in datasets, and further model optimization with prospective utilization of the model will be paramount.

Background: Cardiomyopathies (CM) are a leading cause of heart failure. Identification of genetic risk preemptively may allow detection of the earliest phenotypic manifestations through implementation of interval-based cardiac screening. Previous population-based genetic screening studies have limited follow-up and inadequate phenotyping. Hypothesis: Disease penetrance of CM variants is significantly higher than previously reported. Aims: To determine the prevalence, penetrance, and disease expression of incidental CM-related genetic variants in an unselected, richly phenotyped, adult population. Methods: We analyzed a random sample of 10,000 patients from a quaternary medical center-based biobank cohort for reportable variants in 22-cardiomyopathy genes. Prioritization of genetic variants was performed using Semi-Automated Variant Interpretation, a novel software that has been previously developed and validated at our center. Prioritized variants were then manually curated by 2 independent variant interpretation specialists. Clinical phenotypes and outcomes of participants with likely pathogenic/pathogenic (LP/P) variants were determined. Results: LP/P cardiomyopathy variants were present in 1% of our cohort (n=104) within the following genes: TNNT2 (n=1), DES (n=2), LMNA (n=2), MYL2 (n=2), TNNI3 (n=2), DSG (n=4), FLNC (n=4), DSC2 (n=6), SCN5A (n=7), PKP2 (n=10), DSP (n=12), MYH7 (n=14), TTN (n=17), MYBPC3 (n=21). Only 1 participant carried two different LP/P variants. A definitive disease penetrance defined as heart failure or cardiomyopathy was found in 34% of genotype positive (G+) individuals. Possible disease penetrance defined as presence of structural heart disease identified by echocardiography was found in an additional 16% of participants (concentric left ventricular hypertrophy, diastolic dysfunction, biatrial and left atrial enlargement). Combined definitive and possible disease penetrance was found in 50% of G+ individuals. Arrhythmias and/or cardiac conduction disease (atrial fibrillation, atrioventricular block, ventricular arrhythmia) were present in 31% of G+ individuals. Conclusion: Cardiomyopathy associated LP/P variants are present in a small subset of a large quaternary medical center population as previously reported but unlike prior studies, disease penetrance manifesting as cardiac structural abnormalities and heart failure is high in G+ individuals.