Accurate proteome-wide missense variant effect prediction with AlphaMissense

A comprehensive map of missense trafficking variants in rhodopsin and their response to pharmacologic correction

Rhodopsin (RHO) missense variants are a leading cause of autosomal dominant retinitis pigmentosa (adRP), a progressive retinal degeneration with no currently approved therapies. Interpreting the pathogenicity of the growing number of identified RHO variants is a major clinical challenge, and understanding their disease mechanisms is essential for developing effective therapies. Here, we present a high-resolution map of RHO missense variant trafficking using two complementary deep mutational scanning (DMS) approaches based on a surface abundance immunoassay and a membrane proximity assay. We generated a comprehensive dataset encompassing all 6,612 possible single-residue missense variants, revealing a strong correlation between the two methods. Over 700 variants were identified with pathogenic trafficking scores, significantly expanding the number of RHO variants with functional evidence supporting pathogenicity. We demonstrate a high concordance between the trafficking scores and ClinVar pathogenicity classifications, highlighting this approach's utility in resolving variants of uncertain significance (VUS). The data also identified structurally clustered trafficking-deficient variants, predominantly within the N-terminal region and second extracellular loop, in and above the extracellular/intradiscal beta-plug region. Furthermore, we evaluated the efficacy of the non-retinoid pharmacological chaperone YC-001, observing significant rescue of trafficking defects in a majority of mistrafficking variants. This comprehensive functional map of RHO missense variants provides a valuable resource for pathogenicity assessment, genotype-phenotype correlations, and the development of targeted therapeutic strategies for RHO-adRP, paving the way for improved diagnosis and treatment for patients.

Predicting gene expression responses to environment in Arabidopsis thaliana using natural variation in DNA sequence

The evolution of gene expression responses is a critical component of adaptation to variable environments. Predicting how DNA sequence influences expression is challenging because the genotype to phenotype map is not well resolved for cis regulatory elements, transcription factor binding, regulatory interactions, and epigenetic features, not to mention how these factors respond to environment. We tested if flexible machine learning models could learn some of the underlying cis-regulatory genotype to phenotype map. We tested this approach using cold-responsive transcriptome profiles in 5 diverse Arabidopsis thaliana accessions. We first tested for evidence that cis regulation plays a role in environmental response, finding 14 and 15 motifs that were significantly enriched within the up- and down-stream regions of cold-responsive differentially regulated genes (DEGs). We next applied convolutional neural networks (CNNs), which learn de novo cis-regulatory motifs in DNA sequences to predict expression response to environment. We found that CNNs predicted differential expression with moderate accuracy, with evidence that predictions were hindered by biological complexity of regulation and the large potential regulatory code. Overall, approaches to predict DEGs between specific environments based only on proximate DNA sequences require further development, and additional information may be required.

Modeling SMAD2 Mutations in Induced Pluripotent Stem Cells Provides Insights Into Cardiovascular Disease Pathogenesis

BACKGROUND

SMAD2 is a coregulator that binds a variety of transcription factors in human development. Heterozygous SMAD2 loss-of-function and missense variants are identified in patients with congenital heart disease (CHD) or arterial aneurysms. Mechanisms that cause distinct cardiovascular phenotypes remain unknown. We aimed to define transcriptional and epigenetic effects of SMAD2 variants and their role in CHD. We also assessed the function of SMAD2 missense variants of uncertain significance.

METHODS AND RESULTS

Rare SMAD2 variants (minor allele frequency ≤10-5) were identified in exome sequencing of 11 336 participants with CHD. We constructed isogenic induced pluripotent stem cells with heterozygous or homozygous loss-of-function and missense SMAD2 variants identified in CHD probands. Wild-type and mutant induced pluripotent stem cells were analyzed using bulk RNA sequencing, chromatin accessibility (Assay for Transposase-Accessible Chromatin With Sequencing), and integrated with published SMAD2/3 chromatin immunoprecipitation data. Cardiomyocyte differentiation and contractility were evaluated. Thirty participants with CHD had heterozygous loss-of-function or missense SMAD2 variants. SMAD2 haploinsufficiency altered chromatin accessibility at promoters and dysregulated expression of 385 SMAD regulated genes, including 10 CHD-associated genes. Motifs enriched in differential Assay for Transposase-Accessible Chromatin peaks predicted that SMAD2 haploinsufficiency disrupts interactions with transcription factors NANOG (homeobox protein NANOG), ETS, TEAD3/4 (transcriptional enhanced associate domain 3/4), CREB1 (cAMP response element binding protein 1), and AP1 (activator protein 1). Compared with SMAD2-haploinsufficient cells, induced pluripotent stem cells with R114C or W274C variants exhibited distinct and shared chromatin accessibility and transcription factor binding changes.

CONCLUSIONS

SMAD2 haploinsufficiency disrupts transcription factor binding and chromatin interactions critical for cardiovascular development. Differences between the molecular consequences of loss-of-function and missense variants likely contribute to phenotypic heterogeneity. These findings indicate opportunities for molecular analyses to improve reclassification of SMAD2 variants of uncertain clinical significance.

Evaluating predictors of kinase activity of STK11 variants identified in primary human non-small cell lung cancers

Critical evaluation of computational tools for predicting variant effects is important considering their increased use in disease diagnosis and driving molecular discoveries. In the sixth edition of the Critical Assessment of Genome Interpretation (CAGI) challenge, a dataset of 28 STK11 rare variants (27 missense, 1 single amino acid deletion), identified in primary non-small cell lung cancer biopsies, was experimentally assayed to characterize computational methods from four participating teams and five publicly available tools. Predictors demonstrated a high level of performance on key evaluation metrics, measuring correlation with the assay outputs and separating loss-of-function (LoF) variants from wildtype-like (WT-like) variants. The best participant model, 3Cnet, performed competitively with well-known tools. Unique to this challenge was that the functional data was generated with both biological and technical replicates, thus allowing the assessors to realistically establish maximum predictive performance based on experimental variability. Three out of the five publicly available tools and 3Cnet approached the performance of the assay replicates in separating LoF variants from WT-like variants. Surprisingly, REVEL, an often-used model, achieved a comparable correlation with the real-valued assay output as that seen for the experimental replicates. Performing variant interpretation by combining the new functional evidence with computational and population data evidence led to 16 new variants receiving a clinically actionable classification of likely pathogenic (LP) or likely benign (LB). Overall, the STK11 challenge highlights the utility of variant effect predictors in biomedical sciences and provides encouraging results for driving research in the field of computational genome interpretation.

Modified Rules for Classification of Variants Associated With Disorders of Somatic Mosaicism

Disorders of somatic mosaicism (DoSMs) are rare genetic disorders arising from postzygotic alteration leading to segmental/nonsegmental disease. Current professional guidelines for standardized variant interpretation focus on germline and cancer variants, making them suboptimal for DoSM variant interpretation. The Brain Malformations Variant Curation Expert Panel (BMVCEP) modified existing guidelines to account for brain-specific disorders of somatic mosaicism, but applicability to other DoSM presentations is limited. At Washington University in St. Louis School of Medicine, we have adopted the BMVCEP interpretation framework but suggested alterations that make it more suitable for generalized DoSM variant classification. These modifications include (1) expanding applicability beyond genes associated with brain malformations, (2) introduction of a criterion to interpret truncating variants at the C-terminus of gain of function genes, (3) establishment of a variant allele fraction (VAF) cutoff for applying de novo criteria, and (4) demonstration that in silico prediction tools are relevant to interpretation of gain of function missense variants. Furthermore, modifications to BMVCEP guidelines reduce the number of variants classified as uncertain. The variant classification considerations that we propose have the potential to improve the accuracy of somatic variant classification, better inform clinical care, and may benefit clinical laboratories also conducting DoSM testing.

Decoding the functional impact of the cancer genome through protein-protein interactions

Acquisition of genomic mutations enables cancer cells to gain fitness advantages under selective pressure and, ultimately, leads to oncogenic transformation. Interestingly, driver mutations, even within the same gene, can yield distinct phenotypes and clinical outcomes, necessitating a mutation-focused approach. Conversely, cellular functions are governed by molecular machines and signalling networks that are mostly controlled by protein-protein interactions (PPIs). The functional impact of individual genomic alterations could be transmitted through regulated nodes and hubs of PPIs. Oncogenic mutations may lead to modified residues of proteins, enabling interactions with other proteins that the wild-type protein does not typically interact with, or preventing interactions with proteins that the wild-type protein usually interacts with. This can result in the rewiring of molecular signalling cascades and the acquisition of an oncogenic phenotype. Here, we review the altered PPIs driven by oncogenic mutations, discuss technologies for monitoring PPIs and provide a functional analysis of mutation-directed PPIs. These driver mutation-enabled PPIs and mutation-perturbed PPIs present a new paradigm for the development of tumour-specific therapeutics. The intersection of cancer variants and altered PPI interfaces represents a new frontier for understanding oncogenic rewiring and developing tumour-selective therapeutic strategies.

Evaluation of enzyme activity predictions for variants of unknown significance in Arylsulfatase A

Continued advances in variant effect prediction are necessary to demonstrate the ability of machine learning methods to accurately determine the clinical impact of variants of unknown significance (VUS). Towards this goal, the ARSA Critical Assessment of Genome Interpretation (CAGI) challenge was designed to characterize progress by utilizing 219 experimentally assayed missense VUS in the Arylsulfatase A (ARSA) gene to assess the performance of community-submitted predictions of variant functional effects. The challenge involved 15 teams, and evaluated additional predictions from established and recently released models. Notably, a model developed by participants of a genetics and coding bootcamp, trained with standard machine-learning tools in Python, demonstrated superior performance among submissions. Furthermore, the study observed that state-of-the-art deep learning methods provided small but statistically significant improvement in predictive performance compared to less elaborate techniques. These findings underscore the utility of variant effect prediction, and the potential for models trained with modest resources to accurately classify VUS in genetic and clinical research.

The mutational landscape and its longitudinal dynamics in relapsed and refractory classic Hodgkin lymphoma

In classic Hodgkin-lymphoma (cHL), only a few cases recur, and only a limited fraction of patients is primary-refractory to standard-polychemotherapy. Underlying genomic features of unfavorable clinical courses remain sparsely characterized. Here, we investigated the genomic characteristics of primary-refractory/relapsed cHL in contrast with responders. Therefore, ultra-deep next-generation panel-sequencing was performed on a total of 59 FFPE-samples (20 responders, 26 relapsed (rHL: 11 initial-diagnosis, 15 relapse) and 13 primary-refractory (prHL: 8 initial-diagnosis, 5 progression) from 44 cHL-patients applying a hybrid-capture approach. We compared samples associated with distinct disease courses concerning their oncogenic drivers, mutational signatures, and perturbed pathways. Compared to responders, mutations in genes such as PMS2, PDGFRB, KAT6A, EPHB1, and HGF were detected more frequently in prHL/rHL. Additionally, we observed that in rHL or prHL, BARD1-mutations occur, whereas ETV1, NF1, and MET-mutations were eliminated through clonal selection. A significant enrichment of non-synonymous variants was detected in prHL compared to responders and a significant selection process in favor of NOTCH-pathway mutations driving rHL or prHL was observed. However, our analysis revealed a negative selection process for non-synonymous variants affecting the hippo-pathway. This study delineates distinct mutational signatures between responders and rHL/prHL, whilst illustrating longitudinal dynamics in mutational profiles using paired samples. Further, several exploitable therapeutic vulnerabilities for rHL and prHL were identified.

A computational approach to predict the effects of missense mutations on protein amyloidogenicity: A case study in hereditary transthyretin cardiomyopathy

With many amyloidosis-associated missense mutations still unidentified and early diagnostic methods largely unavailable, there is an urgent need for a reliable computational approach to predict hereditary amyloidoses from gene sequencing data. Progress has been made in predicting amyloidosis-triggering sequences within intrinsically disordered regions. However, some diseases are caused by mutations in amyloidogenic regions within structured domains that must unfold for amyloid formation. Accurate prediction of amyloidogenic regions requires tools for detecting amyloidogenicity and assessing mutation effects on protein stability. We developed datasets of mutations linked to hereditary ATTR cardiomyopathy and others likely unrelated, evaluating TTR mutants with amyloidogenicity and stability predictors. Notably, the stability predictors consistently indicated that ATTR-related mutations tend to destabilize the TTR structure more than non-ATTR-associated mutations. Using these datasets and newly generated mutation features, we developed a machine learning model SDAM-TTR to predict mutations leading to ATTR cardiomyopathy.

Uncovering Phenotypic Expansion in AXIN2-Related Disorders through Precision Animal Modeling

Heterozygous pathogenic variants in AXIN2 are associated with oligodontia-colorectal cancer syndrome (ODCRCS), a disorder characterized by oligodontia, colorectal cancer, and in some cases, sparse hair and eyebrows. We have identified four individuals with one of two de novo , heterozygous variants (NM_004655.4:c.196G>A, p.(Glu66Lys) and c.199G>A, p.(Gly67Arg)) in AXIN2 whose presentations expand the phenotype of AXIN2-related disorders. In addition to ODCRCS features, these individuals have global developmental delay, microcephaly, and limb, ophthalmologic, and renal abnormalities. Structural modeling of these variants suggests that they disrupt AXIN2 binding to tankyrase, which regulates AXIN2 levels through PARsylation and subsequent proteasomal degradation. To test whether these variants produce a phenotype in vivo , we utilized an innovative prime editing N1 screen to phenotype heterozygous (p.E66K) mouse embryos, which were perinatal lethal with short palate and skeletal abnormalities, contrary to published viable Axin2 null mouse models. Modeling of the p.E66K variant in the Drosophila wing revealed gain-of-function activity compared to reference AXIN2. However, the variant showed loss-of-function activity in the fly eye compared to reference AXIN2, suggesting that the mechanism by which p.E66K affects AXIN2 function is cell context-dependent. Together, our studies in humans, mice, and flies demonstrate that specific variants in the tankyrase-binding domain of AXIN2 are pathogenic, leading to phenotypic expansion with context-dependent effects on AXIN2 function and WNT signaling. Moreover, the modeling strategies used to demonstrate variant pathogenicity may be beneficial for the resolution of other de novo heterozygous variants of uncertain significance associated with congenital anomalies in humans.

Chemical Tomography of Cancer Organoids and Cyto-Proteo-Genomic Development Stages Through Chemical Communication Signals

Organoids mimic human organ function, offering insights into development and disease. However, non-destructive, real-time monitoring is lacking, as traditional methods are often costly, destructive, and low-throughput. In this article, a non-destructive chemical tomographic strategy is presented for decoding cyto-proteo-genomics of organoid using volatile signaling molecules, hereby, Volatile Organic Compounds (VOCs), to indicate metabolic activity and development of organoids. Combining a hierarchical design of graphene-based sensor arrays with AI-driven analysis, this method maps VOC spatiotemporal distribution and generate detailed digital profiles of organoid morphology and proteo-genomic features. Lens- and label-free, it avoids phototoxicity, distortion, and environmental disruption. Results from testing organoids with the reported chemical tomography approach demonstrate effective differentiation between cyto-proteo-genomic profiles of normal and diseased states, particularly during dynamic transitions such as epithelial-mesenchymal transition (EMT). Additionally, the reported approach identifies key VOC-related biochemical pathways, metabolic markers, and pathways associated with cancerous transformations such as aromatic acid degradation and lipid metabolism. This real-time, non-destructive approach captures subtle genetic and structural variations with high sensitivity and specificity, providing a robust platform for multi-omics integration and advancing cancer biomarker discovery.

VarMeter2: An enhanced structure-based method for predicting pathogenic missense variants through Mahalanobis distance

Various computational methods have been developed to predict the pathogenicity of missense variants, which is crucial for diagnosing rare diseases. Recently, we introduced VarMeter, a diagnostic tool for predicting variant pathogenicity based on normalized solvent-accessible surface area (nSASA) and mutation energy calculated from AlphaFold 3D models, and validated it on arylsulfatase L. To evaluate the broader applicability of VarMeter and enhance its predictive accuracy, here we analyzed 296 pathogenic and 240 benign variants extracted from the ClinVar database. By comparing structural features including nSASA, mutation energy, and predicted local distance difference test (pLDDT) score, we identified distinct characteristics between pathogenic and benign variants. These features were used to develop VarMeter2, which classifies variants based on Mahalanobis distance. VarMeter2 achieved a prediction accuracy of 82 % for the ClinVar dataset, a marked improvement over the original VarMeter (74 %), and 84 % for published missense variants of N-sulphoglucosamine sulphohydrolase (SGSH), an enzyme associated with Sanfillippo syndrome A. Application of VarMeter 2 to SGSH variants in our clinical database identified a novel SGSH variant, Q365P, as pathogenic. The recombinant Q365P protein lacked enzymatic activity as compared with wild-type SGSH. Furthermore, it was largely retained in the endoplasmic reticulum and failed to reach the Golgi, probably due to misfolding. Protein stability assays confirmed reduced stability of the variant, further explaining its loss of function. Consistently, the patient homozygous for this variant was diagnosed with Sanfilippo syndrome A. These results underscore the predictive power and versatility of VarMeter2 in assessing the pathogenicity of missense variants.

Biallelic variants of SEMA3F are associated with nonsyndromic hearing loss

It is crucial to manage hearing loss and its associated public health impacts. In this study, we aimed to understand the role of Sema3f in the development and maintenance of the auditory system. Inner ear-specific Sema3f knockout mice exhibited hearing loss at 8 weeks with an elevated threshold for auditory brainstem response and an absent threshold for distortion product optoacoustic emission tests. Additionally, an increased number of outer hair cells and abnormal patterns of spiral ganglion neuron projections in the outer hair cell regions were observed. Through the analyses of sequencing data from 558 families with hearing loss, we identified biallelic variants of SEMA3F, which encodes semaphorin-3F, in one of the families. In the family, the proband showed profound progressive nonsyndromic hearing loss with congenital onset. In vitro analysis revealed that the identified missense variants decreased the furin-mediated processing of SEMA3F and abolished the cellular abilities of SEMA3F, which collapsed the filamentous actin cytoskeleton in human umbilical vein-derived endothelial cells. Our data suggest that SEMA3F is essential for normal hearing and is associated with nonsyndromic hearing loss in humans.

ZFP-CanPred: Predicting the effect of mutations in zinc-finger proteins in cancers using protein language models

Zinc-finger proteins (ZNFs) constitute the largest family of transcription factors and play crucial roles in various cellular processes. Missense mutations in ZNFs significantly alter protein-DNA interactions, potentially leading to the development of various types of cancers. This study presents ZFP-CanPred, a novel deep learning-based model for predicting cancer-associated driver mutations in ZNFs. The representations derived from protein language models (PLMs) from the structural neighbourhood of mutated sites were utilized to train ZFP-CanPred for differentiating between cancer-causing and neutral mutations. ZFP-CanPred, achieved a superior performance with an accuracy of 0.72, F1-score of 0.79, and area under the Receiver Operating Characteristics (ROC) Curve (AUC) of 0.74, on an independent test set. In a comparative analysis against 11 existing prediction tools using a curated dataset of 331 mutations, ZFP-CanPred demonstrated the highest AU-ROC of 0.74, outperforming both generic and cancer-specific methods. The model's balanced performance across specificity and sensitivity addresses a significant limitation of current methodologies. The source code and other related files are available on GitHub at https://github.com/amitphogat/ZFP-CanPred.git. We envisage that the present study contributes to understand the oncogenic processes and developing targeted therapeutic strategies.

Integrating Artificial Intelligence and Bioinformatics Methods to Identify Disruptive STAT1 Variants Impacting Protein Stability and Function

Background: The Signal Transducer and Activator of Transcription 1 (STAT1) gene is an essential component of the JAK-STAT signaling pathway. This pathway plays a pivotal role in the regulation of different cellular processes, including immune responses, cell growth, and apoptosis. Mutations in the STAT1 gene contribute to a variety of immune system dysfunctions. Objectives: We aim to identify disease-susceptible single-nucleotide polymorphisms (SNPs) in STAT1 gene and predict structural changes associated with the mutations that disrupt normal protein-protein interactions using different computational algorithms. Methods: Several in silico tools, such as SIFT, Polyphen v2, PROVEAN, SNAP2, PhD-SNP, SNPs&GO, Pmut, and PANTHER, were used to determine the deleterious nsSNPs of the STAT1. Further, we evaluated the potentially deleterious SNPs for their effect on protein stability using I-Mutant, MUpro, and DDMUT. Additionally, we predicted the functional and structural effects of the nsSNPs using MutPred. We used Alpha-Missense to predict missense variant pathogenicity. Moreover, we predicted the 3D structure of STAT1 using an artificial intelligence system, alphafold, and the visualization of the 3D structures of the wild-type amino acids and the mutant residues was performed using ChimeraX 1.9 software. Furthermore, we analyzed the structural and conformational variations that have resulted from SNPs using Project Hope, while changes in the biological interactions between wild type, mutant amino acids, and neighborhood residues was studied using DDMUT. Conservational analysis and surface accessibility prediction of STAT1 was performed using ConSurf. We predicted the protein-protein interaction using STRING database. Results: In the current study, we identified six deleterious nsSNPs (R602W, I648T, V642D, L600P, I578N, and W504C) and their effect on protein structure, function, and stability. Conclusions: These findings highlight the potential of approaches to pinpoint pathogenic SNPs, providing a time- and cost-effective alternative to experimental approaches. To the best of our knowledge, this is the first comprehensive study in which we analyze STAT1 gene variants using both bioinformatics and artificial-intelligence-based model tools.

Artificial intelligence in clinical genetics

Artificial intelligence (AI) has been growing more powerful and accessible, and will increasingly impact many areas, including virtually all aspects of medicine and biomedical research. This review focuses on previous, current, and especially emerging applications of AI in clinical genetics. Topics covered include a brief explanation of different general categories of AI, including machine learning, deep learning, and generative AI. After introductory explanations and examples, the review discusses AI in clinical genetics in three main categories: clinical diagnostics; management and therapeutics; clinical support. The review concludes with short, medium, and long-term predictions about the ways that AI may affect the field of clinical genetics. Overall, while the precise speed at which AI will continue to change clinical genetics is unclear, as are the overall ramifications for patients, families, clinicians, researchers, and others, it is likely that AI will result in dramatic evolution in clinical genetics. It will be important for all those involved in clinical genetics to prepare accordingly in order to minimize the risks and maximize benefits related to the use of AI in the field.

BCL11A intellectual developmental disorder: defining the clinical spectrum and genotype-phenotype correlations

An increasing number of individuals with intellectual developmental disorder (IDD) and heterozygous variants in BCL11A are identified, yet our knowledge of manifestations and mutational spectrum is lacking. To address this, we performed detailed analysis of 42 individuals with BCL11A-related IDD (BCL11A-IDD, a.k.a. Dias-Logan syndrome) ascertained through an international collaborative network, and reviewed 35 additional previously reported patients. Analysis of 77 affected individuals identified 60 unique disease-causing variants (30 frameshift, 7 missense, 6 splice-site, 17 stop-gain) and 8 unique BCL11A microdeletions. We define the most prevalent features of BCL11A-IDD: IDD, postnatal-onset microcephaly, hypotonia, behavioral abnormalities, autism spectrum disorder, and persistence of fetal hemoglobin (HbF), and identify autonomic dysregulation as new feature. BCL11A-IDD is distinguished from 2p16 microdeletion syndrome, which has a higher incidence of congenital anomalies. Our results underscore BCL11A as an important transcription factor in human hindbrain development, identifying a previously underrecognized phenotype of a small brainstem with a reduced pons/medulla ratio. Genotype-phenotype correlation revealed an isoform-dependent trend in severity of truncating variants: those affecting all isoforms are associated with higher frequency of hypotonia, and those affecting the long (BCL11A-L) and extra-long (-XL) isoforms, sparing the short (-S), are associated with higher frequency of postnatal microcephaly. With the largest international cohort to date, this study highlights persistence of fetal hemoglobin as a consistent biomarker and hindbrain abnormalities as a common feature. It contributes significantly to our understanding of BCL11A-IDD through an extensive unbiased multi-center assessment, providing valuable insights for diagnosis, management and counselling, and into BCL11A's role in brain development.

Can GPTs Accelerate the Development of Intelligent Diagnosis and Treatment in Traditional Chinese Medicine? A Survey and Empirical Analysis

Intelligent traditional Chinese medicine (TCM) is a key pathway toward the modernization and globalization of TCM in the era of artificial intelligence. Due to its unique terminology and diagnostic framework, TCM's intelligentization process has long faced a range of challenges, from the digitization and formalization of knowledge bases to the differentiation of syndromes and personalized treatment. Recently, the advent of large language models (LLMs) like GPTs has marked a transformative milestone in semantic understanding tasks, attracting widespread attention from the medical, academic, and industrial communities. Nonetheless, LLMs often suffer from accuracy and logical reasoning limitations within specific fields and may manifest hallucinations in the generative outputs. Through a comprehensive review of existing literature and empirical analyses, this study delves into the potential and challenges of adapting LLMs to TCM. Promising perspectives on future developments at this innovative intersection are discussed.

Advancing evolutionary medicine with complete primate genomes and advanced biotechnologies

Evolutionary medicine, which integrates evolutionary biology and medicine, significantly enhances our understanding of human traits and disease susceptibility. However, previous studies in this field have often focused on single-nucleotide variants due to technological limitations in characterizing complex genomic regions, hindering the comprehensive analyses of their evolutionary origins and clinical significance. In this review, we summarize recent advancements in complete telomere-to-telomere (T2T), primate genomes and other primate resources, and illustrate how these resources facilitate the research of complex regions. We focus on several biomedically relevant regions to examine the relationship between primate genome evolution and human diseases. We also highlight the potentials of high-throughput functional genomic technologies for assessing candidate loci. Finally, we discuss future directions for primate research within the context of evolutionary medicine.

An augmented transformer model trained on protein family specific variant data leads to improved prediction of variants of uncertain significance

Variants of uncertain significance (VUS) represent variants that lack sufficient evidence to be confidently associated with a disease, thus posing a challenge in the interpretation of genetic testing results. Here we report an improved method for predicting the VUS of Arylsulfatase A (ARSA) gene as part of the Critical Assessment of Genome Interpretation challenge (CAGI6). Our method uses a transfer learning approach that leverages a pre-trained protein language model to predict the impact of mutations on the activity of the ARSA enzyme, whose deficiency is known to cause a rare genetic disorder, metachromatic leukodystrophy. Our innovative framework combines zero-shot log odds scores and embeddings from the ESM, an evolutionary scale model as features for training a supervised model on gene variants functionally related to the ARSA gene. The zero-shot log odds score feature captures the generic properties of the proteins learned due to its pre-training on millions of sequences in the UniProt data, while the ESM embeddings for the proteins in the ARSA family capture features specific to the family. We also tested our approach on another enzyme, N-acetyl-glucosaminidase (NAGLU), that belongs to the same superfamily as ARSA. Our results demonstrate that the performance of our family models (augmented ESM models) is either comparable or better than the ESM models. The ARSA model compares favorably with the majority of state-of-the-art predictors on area under precision and recall curve (AUPRC) performance metric. However, the NAGLU model outperforms all pathogenicity predictors evaluated in this study on AUPRC metric. The improved AUPRC has relevance in a diagnostic setting where variant prioritization generally entails identifying a small number of pathogenic variants from a larger number of benign variants. Our results also indicate that genes that have sparse or no experimental variant impact data, the family variant data can serve as a proxy training data for making accurate predictions. Attention analysis of active sites and binding sites in ARSA and NAGLU proteins shed light on probable mechanisms of pathogenicity for positions that are highly attended.

A quantitative intracellular peptide-binding assay reveals recognition determinants and context dependence of short linear motifs

Transient protein-protein interactions play key roles in controlling dynamic cellular responses. Many examples involve globular protein domains that bind to peptide sequences known as short linear motifs (SLiMs), which are enriched in intrinsically disordered regions of proteins. Here we describe a novel functional assay for measuring SLiM binding, called systematic intracellular motif-binding analysis (SIMBA). In this method, binding of a foreign globular domain to its cognate SLiM peptide allows yeast cells to proliferate by blocking a growth arrest signal. A high-throughput application of the SIMBA method involving competitive growth and deep sequencing provides rapid quantification of the relative binding strength for thousands of SLiM sequence variants and a comprehensive interrogation of SLiM sequence features that control their recognition and potency. We show that multiple distinct classes of SLiM-binding domains can be analyzed by this method and that the relative binding strength of peptides in vivo correlates with their biochemical affinities measured in vitro. Deep mutational scanning provides high-resolution definitions of motif recognition determinants and reveals how sequence variations at noncore positions can modulate binding strength. Furthermore, mutational scanning of multiple parent peptides that bind human tankyrase ARC or YAP WW domains identifies distinct binding modes and uncovers context effects in which the preferred residues at one position depend on residues elsewhere. The findings establish SIMBA as a fast and incisive approach for interrogating SLiM recognition via massively parallel quantification of protein-peptide binding strength in vivo.

QAFI: a novel method for quantitative estimation of missense variant impact using protein-specific predictors and ensemble learning

Next-generation sequencing (NGS) has revolutionized genetic diagnostics, yet its application in precision medicine remains incomplete, despite significant advances in computational tools for variant annotation. Many variants remain unannotated, and existing tools often fail to accurately predict the range of impacts that variants have on protein function. This limitation restricts their utility in relevant applications such as predicting disease severity and onset age. In response to these challenges, a new generation of computational models is emerging, aimed at producing quantitative predictions of genetic variant impacts. However, the field is still in its early stages, and several issues need to be addressed, including improved performance and better interpretability. This study introduces QAFI, a novel methodology that integrates protein-specific regression models within an ensemble learning framework, utilizing conservation-based and structure-related features derived from AlphaFold models. Our findings indicate that QAFI significantly enhances the accuracy of quantitative predictions across various proteins. The approach has been rigorously validated through its application in the CAGI6 contest, focusing on ARSA protein variants, and further tested on a comprehensive set of clinically labeled variants, demonstrating its generalizability and robust predictive power. The straightforward nature of our models may also contribute to better interpretability of the results.

Critical assessment of missense variant effect predictors on disease-relevant variant data

Regular, systematic, and independent assessments of computational tools that are used to predict the pathogenicity of missense variants are necessary to evaluate their clinical and research utility and guide future improvements. The Critical Assessment of Genome Interpretation (CAGI) conducts the ongoing Annotate-All-Missense (Missense Marathon) challenge, in which missense variant effect predictors (also called variant impact predictors) are evaluated on missense variants added to disease-relevant databases following the prediction submission deadline. Here we assess predictors submitted to the CAGI 6 Annotate-All-Missense challenge, predictors commonly used in clinical genetics, and recently developed deep learning methods. We examine performance across a range of settings relevant for clinical and research applications, focusing on different subsets of the evaluation data as well as high-specificity and high-sensitivity regimes. Our evaluations reveal notable advances in current methods relative to older, well-cited tools in the field. While meta-predictors tend to outperform their constituent individual predictors, several newer individual predictors perform comparably to commonly used meta-predictors. Predictor performance varies between high-specificity and high-sensitivity regimes, highlighting that different methods may be optimal for different use cases. We also characterize two potential sources of bias. Predictors that incorporate allele frequency as a predictive feature tend to have reduced performance when distinguishing pathogenic variants from very rare benign variants, and predictors trained on pathogenicity labels from curated variant databases often inherit gene-level label imbalances. Our findings help illuminate the clinical and research utility of modern missense variant effect predictors and identify potential areas for future development.

Rare genetic associations with human lifespan in UK Biobank are enriched for oncogenic genes

Human lifespan is shaped by genetic and environmental factors. To enable precision health, understanding how genetic variants influence mortality is essential. We conducted a survival analysis in European ancestry participants of the UK Biobank, using age-at-death (N=35,551) and last-known-age (N=358,282). The associations identified were predominantly driven by cancer. We found lifespan-associated loci (APOE, ZSCAN23) for common variants and six genes where burden of loss-of-function variants were linked to reduced lifespan (TET2, ATM, BRCA2, CKMT1B, BRCA1, ASXL1). Additionally, eight genes with pathogenic missense variants were associated with reduced lifespan (DNMT3A, SF3B1, TET2, PTEN, SOX21, TP53, SRSF2, RLIM). Many of these genes are involved in oncogenic pathways and clonal hematopoiesis. Our findings highlight the importance of understanding genetic factors driving the most prevalent causes of mortality at a population level, highlighting the potential of early genetic testing to identify germline and somatic variants increasing one's susceptibility to cancer and/or early death.

Landscapes of missense variant impact for human superoxide dismutase 1

Amyotrophic lateral sclerosis (ALS) is a progressive motor neuron disease for which important subtypes are caused by variation in the Superoxide Dismutase 1 gene SOD1. Diagnosis based on SOD1 sequencing can not only be definitive but also indicate specific therapies available for SOD1-associated ALS (SOD1-ALS). Unfortunately, SOD1-ALS diagnosis is limited by the fact that a substantial fraction (currently 26%) of ClinVar SOD1 missense variants are classified as "variants of uncertain significance" (VUS). Although functional assays can provide strong evidence for clinical variant interpretation, SOD1 assay validation is challenging, given the current incomplete and controversial understanding of SOD1-ALS disease mechanism. Using saturation mutagenesis and multiplexed cell-based assays, we measured the functional impact of over two thousand SOD1 amino acid substitutions on both enzymatic function and protein abundance. The resulting 'missense variant effect maps' not only reflect prior biochemical knowledge of SOD1 but also provide sequence-structure-function insights. Importantly, our variant abundance assay can discriminate pathogenic missense variation and provides new evidence for 41% of missense variants that had been previously reported as VUS, offering the potential to identify additional patients who would benefit from therapy approved for SOD1-ALS.

The times they are AI-changing: AI-powered advances in the application of extracellular vesicles to liquid biopsy in breast cancer

Artificial intelligence (AI) is revolutionizing scientific research by facilitating a paradigm shift in data analysis and discovery. This transformation is characterized by a fundamental change in scientific methods and concepts due to AI's ability to process vast datasets with unprecedented speed and accuracy. In breast cancer research, AI aids in early detection, prognosis, and personalized treatment strategies. Liquid biopsy, a noninvasive tool for detecting circulating tumor traits, could ideally benefit from AI's analytical capabilities, enhancing the detection of minimal residual disease and improving treatment monitoring. Extracellular vesicles (EVs), which are key elements in cell communication and cancer progression, could be analyzed with AI to identify disease-specific biomarkers. AI combined with EV analysis promises an enhancement in diagnosis precision, aiding in early detection and treatment monitoring. Studies show that AI can differentiate cancer types and predict drug efficacy, exemplifying its potential in personalized medicine. Overall, the integration of AI in biomedical research and clinical practice promises significant changes and advancements in diagnostics, personalized medicine-based approaches, and our understanding of complex diseases like cancer.

Discordance between a deep learning model and clinical-grade variant pathogenicity classification in a rare disease cohort

Genetic testing is essential for diagnosing and managing clinical conditions, particularly rare Mendelian diseases. Although efforts to identify rare phenotype-associated variants have focused on protein-truncating variants, interpreting missense variants remains challenging. Deep learning algorithms excel in various biomedical tasks1,2, yet distinguishing pathogenic from benign missense variants remains elusive3-5. Our investigation of AlphaMissense (AM)5, a deep learning tool for predicting the potential functional impact of missense variants and assessing gene essentiality, reveals limitations in identifying pathogenic missense variants over 45 rare diseases, including very early onset inflammatory bowel disease. For the expert-curated pathogenic variants identified in our cohort, AM's precision was 32.9%, and recall was 57.6%. Notably, AM struggles to evaluate pathogenicity in intrinsically disordered regions (IDRs), resulting in unreliable gene-level essentiality scores for genes containing IDRs. This observation underscores ongoing challenges in clinical genetics, highlighting the need for continued refinement of computational methods in variant pathogenicity prediction.

Machine Learning Methods for Classifying Multiple Sclerosis and Alzheimer's Disease Using Genomic Data

Complex diseases pose challenges in prediction due to their multifactorial and polygenic nature. This study employed machine learning (ML) to analyze genomic data from the UK Biobank, aiming to predict the genomic predisposition to complex diseases like multiple sclerosis (MS) and Alzheimer's disease (AD). We tested logistic regression (LR), ensemble tree methods, and deep learning models for this purpose. LR displayed remarkable stability across various subsets of data, outshining deep learning approaches, which showed greater variability in performance. Additionally, ML methods demonstrated an ability to maintain optimal performance despite correlated genomic features due to linkage disequilibrium. When comparing the performance of polygenic risk score (PRS) with ML methods, PRS consistently performed at an average level. By employing explainability tools in the ML models of MS, we found that the results confirmed the polygenicity of this disease. The highest-prioritized genomic variants in MS were identified as expression or splicing quantitative trait loci located in non-coding regions within or near genes associated with the immune response, with a prevalence of human leukocyte antigen (HLA) gene annotations. Our findings shed light on both the potential and the challenges of employing ML to capture complex genomic patterns, paving the way for improved predictive models.

PON-P3: Accurate Prediction of Pathogenicity of Amino Acid Substitutions

Different types of information are combined during variation interpretation. Computational predictors, most often pathogenicity predictors, provide one type of information for this purpose. These tools are based on various kinds of algorithms. Although the American College of Genetics and the Association for Molecular Pathology guidelines classify variants into five categories, practically all pathogenicity predictors provide binary pathogenic/benign predictions. We developed a novel artificial intelligence-based tool, PON-P3, on the basis of a carefully selected training dataset, meticulous feature selection, and optimization. We started with 1526 features describing variations, their sequence and structural context, and parameters for the affected genes and proteins. The final random boosting method was tested and compared with a total of 23 predictors. PON-P3 performed better than recently introduced predictors, which utilize large language models or structural predictions. PON-P3 was better than methods that use evolutionary data alone or in combination with different gene and protein properties. PON-P3 classifies cases into three categories as benign, pathogenic, and variants of uncertain significance (VUSs). When binary test data were used, some metapredictors performed slightly better than PON-P3; however, in real-life situations, with patient data, those methods overpredict both pathogenic and benign cases. We predicted with PON-P3 all possible amino acid substitutions in all human proteins encoded from MANE transcripts. The method was also used to predict all unambiguous VUSs (i.e., without conflicts) in ClinVar. A total of 12.9% were predicted to be pathogenic, and 49.9% were benign.

Identification of mutations associated with congenital cataracts in nineteen Chinese families

BACKGROUND

Congenital cataracts (CC) are one of the leading causes of impaired vision or blindness in children, with approximately 8.3-25% being inherited. The aim of this study is to investigate the mutation spectrum and frequency of 9 cataract-associated genes in 19 Chinese families with congenital cataracts.

PURPOSE

To identify the gene variants associated with congenital cataracts.

METHODS

This study included a total of 58 patients from 19 pedigrees with congenital cataracts. All probands were initially screened by whole-exome sequencing(WES), and then validated by co-segregation analysis using Sanger sequencing.

RESULTS

Likely pathogenic variants were detected in 8 families, with a positivity rate of 42.1%. Variants in various genes were identified, including GJA3, CRYGD, CRYBA4, BFSP2, IARS2, CRYAA, CRYBA1, ARL2 and CRYBB3. Importantly, this study identified compound heterozygous variants of IARS2 in one family.

CONCLUSIONS

Our research findings have revealed multiple gene variants associated with cataracts, providing clinical guidance for improved molecular diagnosis of congenital cataracts in the era of precision medicine.

Integrative analysis of KCNQ1 variants reveals molecular mechanisms of type 1 long QT syndrome pathogenesis

Loss-of-function (LOF) pathogenic variants in KCNQ1 encoding a cardiac potassium channel predispose to sudden cardiac death in type 1 congenital long QT syndrome (LQT1). To determine the spectrum of molecular mechanisms responsible for this life-threatening condition, we used an integrative approach to determine the biophysical, functional, and trafficking properties of 61 KCNQ1 variants distributed throughout all domains of the channel. Impaired trafficking to the plasma membrane was the most common cause of LOF across all channel domains, often but not always coinciding with protein instability. However, many LOF variants, particularly in transmembrane domains, trafficked normally, but when coexpressed with KCNE1 exhibited impaired conductance, altered voltage dependence, or abnormal gating kinetics, highlighting diverse pathogenic mechanisms. This indicates a need for personalized treatment approaches for LQT1. Use of our data to benchmark variant pathogenicity prediction methods demonstrated that prediction accuracy depends on the exact mechanism of pathogenicity associated with a given variant.

Computational modeling of human genetic variants in mice

Mouse models represent a powerful platform to study genes and variants associated with human diseases. While genome editing technologies have increased the rate and precision of model development, predicting and installing specific types of mutations in mice that mimic the native human genetic context is complicated. Computational tools can identify and align orthologous wild-type genetic sequences from different species; however, predictive modeling and engineering of equivalent mouse variants that mirror the nucleotide and/or polypeptide change effects of human variants remains challenging. Here, we present H2M (human-to-mouse), a computational pipeline to analyze human genetic variation data to systematically model and predict the functional consequences of equivalent mouse variants. We show that H2M can integrate mouse-to-human and paralog-to-paralog variant mapping analyses with precision genome editing pipelines to devise strategies tailored to model specific variants in mice. We leveraged these analyses to establish a database containing > 3 million human-mouse equivalent mutation pairs, as well as in silico-designed base and prime editing libraries to engineer 4,944 recurrent variant pairs. Using H2M, we also found that predicted pathogenicity and immunogenicity scores were highly correlated between human-mouse variant pairs, suggesting that variants with similar sequence change effects may also exhibit broad interspecies functional conservation. Overall, H2M fills a gap in the field by establishing a robust and versatile computational framework to identify and model homologous variants across species while providing key experimental resources to augment functional genetics and precision medicine applications. The H2M database (including software package and documentation) can be accessed at https://human2mouse.com.

Antenatal phenotype associated with PAK2 pathogenic variants: bilateral pleural effusion as a warning sign

Fetal pleural effusions can arise in various contexts with different prognosis. They have been reported in fetuses presenting with hereditary or acquired conditions. One particularly rare genetic disorder, known as Knobloch syndrome, seems to emerge as a potential new cause of fetal pleural effusions, associated with severe outcomes. Knobloch syndrome 1 can be caused by biallelic variants in COL18A1. It is primarily characterized by its ophthalmic features, including severe vitreoretinal degeneration with retinal detachment and macular abnormalities. Neurological defects such as encephalocele and developmental delay, along with skeletal and renal malformations, are also associated with the syndrome. The Knobloch syndrome 2 is caused by monoallelic variants in the kinase domain of PAK2. It is less described and seems to also be associated with cardiac and respiratory damage in addition to the Knobloch syndrome 1 phenotype. PAK2 is a ubiquitous protein with a major implication in regulation and remodeling of the cytoskeleton and numerous other cellular pathways. Knobloch-associated variants seem to cause a loss of the kinase function of the protein. Even if the ophthalmic defects are almost constant, PAK2-associated Knobloch syndrome has slightly different features from Knobloch syndrome 1 in which pulmonary and lymphatic damages are still unseen. In a prenatal trio exome sequencing, we identified a novel de novo PAK2 missense variant, NM_002577.4:c.836 A > C, p.(Gln279Pro), classified as likely pathogenic in a 24 weeks of gestation fetus whose only sign was severe bilateral pleural effusion. From a literature review of patients, we recognize this sign as an important antenatal indicator of Knobloch syndrome 2, as it was the first sign identifiable in 2 out of 5 patients. This adds new evidence for the implication of this gene in fetal pleural effusions, with potentially severe outcomes.

Leveraging protein structural information to improve variant effect prediction

Despite massive sequencing efforts, understanding the difference between human pathogenic and benign variants remains a challenge. Computational variant effect predictors (VEPs) have emerged as essential tools for assessing the impact of genetic variants, although their performance varies. Initially, sequence-based methods dominated the field, but recent advances, particularly in protein structure prediction technologies like AlphaFold, have led to an increased utilization of structural information by VEPs aimed at scoring human missense variants. This review highlights the progress in integrating structural information into VEPs, showcasing novel models such as AlphaMissense, PrimateAI-3D, and CPT-1 that demonstrate improved variant evaluation. Structural data offers more interpretability, especially for non-loss-of-function variants, and provides insights into complex variant interactions in vivo. As the field advances, utilizing biomolecular complex structures will be pivotal for future VEP development, with recent breakthroughs in protein-ligand and protein-nucleic acid complex prediction offering new avenues.

Geographic and age-related variations in mutational processes in colorectal cancer

Colorectal cancer incidence rates vary geographically and have changed over time. Notably, in the past two decades, the incidence of early-onset colorectal cancer, affecting individuals under the age of 50 years, has doubled in many countries. The reasons for this increase are unknown. Here, we investigate whether mutational processes contribute to geographic and age-related differences by examining 981 colorectal cancer genomes from 11 countries. No major differences were found in microsatellite unstable cancers, but variations in mutation burden and signatures were observed in the 802 microsatellite-stable cases. Multiple signatures, most with unknown etiologies, exhibited varying prevalence in Argentina, Brazil, Colombia, Russia, and Thailand, indicating geographically diverse levels of mutagenic exposure. Signatures SBS88 and ID18, caused by the bacteria-produced mutagen colibactin, had higher mutation loads in countries with higher colorectal cancer incidence rates. SBS88 and ID18 were also enriched in early-onset colorectal cancers, being 3.3 times more common in individuals diagnosed before age 40 than in those over 70, and were imprinted early during colorectal cancer development. Colibactin exposure was further linked to APC driver mutations, with ID18 responsible for about 25% of APC driver indels in colibactin-positive cases. This study reveals geographic and age-related variations in colorectal cancer mutational processes, and suggests that early-life mutagenic exposure to colibactin-producing bacteria may contribute to the rising incidence of early-onset colorectal cancer.

X-linked Deficiency in ELF4 in Females with Skewed X Chromosome Inactivation

Deficiency in ELF4, X-linked (DEX) is a newly identified monogenic autoinflammatory disease. Most reported cases are male, leading to the recognition of DEX being primarily limited to male patients. Here we described 3 pediatric female patients with DEX from 3 unrelated families, who are all heterozygous for ELF4 mutations (c.320_c.321insA, c.329delA and c.685 A > G). Similar to reported male DEX patients, the main clinical features include recurring oral ulcers, abdominal pain and diarrhea with colonoscopy showing ulcers in the colon. Meanwhile, novel and effective treatment strategies, such as the use of the biologic vedolizumab and exclusive enteral nutrition (EEN), have provided additional options for the treatment of DEX. Finally, we observed skewed X chromosome inactivation patterns in all three female patients, with over-inactivation of the X chromosome carrying the wild-type allele confirmed in two of them.

Rewiring protein sequence and structure generative models to enhance protein stability prediction

Predicting changes in protein thermostability due to amino acid substitutions is essential for understanding human diseases and engineering useful proteins for clinical and industrial applications. While recent advances in protein generative models, which learn probability distributions over amino acids conditioned on structural or evolutionary sequence contexts, have shown impressive performance in predicting various protein properties without task-specific training, their strong unsupervised prediction ability does not extend to all protein functions. In particular, their potential to improve protein stability prediction remains underexplored. In this work, we present SPURS, a novel deep learning framework that adapts and integrates two general-purpose protein generative models-a protein language model (ESM) and an inverse folding model (ProteinMPNN)-into an effective stability predictor. SPURS employs a lightweight neural network module to rewire per-residue structure representations learned by ProteinMPNN into the attention layers of ESM, thereby informing and enhancing ESM's sequence representation learning. This rewiring strategy enables SPURS to harness evolutionary patterns from both sequence and structure data, where the sequence like-lihood distribution learned by ESM is conditioned on structure priors encoded by ProteinMPNN to predict mutation effects. We steer this integrated framework to a stability prediction model through supervised training on a recently released mega-scale thermostability dataset. Evaluations across 12 benchmark datasets showed that SPURS delivers accurate, rapid, scalable, and generalizable stability predictions, consistently outperforming current state-of-the-art methods. Notably, SPURS demonstrates remarkable versatility in protein stability and function analyses: when combined with a protein language model, it accurately identifies protein functional sites in an unsupervised manner. Additionally, it enhances current low- N protein fitness prediction models by serving as a stability prior model to improve accuracy. These results highlight SPURS as a powerful tool to advance current protein stability prediction and machine learning-guided protein engineering workflows. The source code of SPURS is available at https://github.com/luo-group/SPURS .

RNA methyltransferase SPOUT1/CENP-32 links mitotic spindle organization with the neurodevelopmental disorder SpADMiSS

SPOUT1/CENP-32 encodes a putative SPOUT RNA methyltransferase previously identified as a mitotic chromosome associated protein. SPOUT1/CENP-32 depletion leads to centrosome detachment from the spindle poles and chromosome misalignment. Aided by gene matching platforms, here we identify 28 individuals with neurodevelopmental delays from 21 families with bi-allelic variants in SPOUT1/CENP-32 detected by exome/genome sequencing. Zebrafish spout1/cenp-32 mutants show reduction in larval head size with concomitant apoptosis likely associated with altered cell cycle progression. In vivo complementation assays in zebrafish indicate that SPOUT1/CENP-32 missense variants identified in humans are pathogenic. Crystal structure analysis of SPOUT1/CENP-32 reveals that most disease-associated missense variants are located within the catalytic domain. Additionally, SPOUT1/CENP-32 recurrent missense variants show reduced methyltransferase activity in vitro and compromised centrosome tethering to the spindle poles in human cells. Thus, SPOUT1/CENP-32 pathogenic variants cause an autosomal recessive neurodevelopmental disorder: SpADMiSS (SPOUT1 Associated Development delay Microcephaly Seizures Short stature) underpinned by mitotic spindle organization defects and consequent chromosome segregation errors.

Comparative sequencing study of mismatch repair and homology-directed repair genes in endometrial cancer and breast cancer patients from Kazakhstan

Endometrial cancer has been associated with pathogenic variants in mismatch repair (MMR) genes, especially in the context of the hereditary Lynch Syndrome. More recently, pathogenic variants in genes of homology-directed repair (HDR) have also been suggested to contribute to a subset of endometrial cancers. In the present hospital-based study, we investigated the relative distribution of pathogenic MMR or HDR gene variants in a series of 342 endometrial cancer patients from the Oncology Clinic in Almaty, Kazakhstan. In comparison, we also sequenced 178 breast cancer patients from the same population with the same gene panel. Identified variants were classified according to ClinVar, ESM1b, and AlphaMissense prediction tools. We found 10 endometrial cancer patients (2.9%) carrying pathogenic or likely pathogenic variants in MMR genes (7 MSH6, 1 MSH2, 2 MUTYH), while 14 endometrial cancer patients (4.1%) carried pathogenic variants in HDR genes (4 BRCA2, 3 BRCA1, 3 FANCM, 2 SLX4, 1 BARD1, 1 BRIP1). In the breast cancer series, we found 8 carriers (4.5%) of pathogenic or likely pathogenic variants in MMR genes (2 MSH2, 2 MSH6, 4 MUTYH) while 12 patients (6.7%) harbored pathogenic or likely pathogenic HDR gene variants (5 BRCA1, 3 BRCA2, 1 BRIP1, 1 ERRC4, 1 FANCM, 1 SLX4). One patient who developed breast cancer first and endometrial cancer later carried a novel frameshift variant in MSH6. Our results indicate that MMR and HDR gene variants with predicted pathogenicity occur at substantial frequencies in both breast and endometrial cancer patients from the Kazakh population.

Navigating Uncertainty: Assessing Variants of Uncertain Significance in the CDKL5 Gene for Developmental and Epileptic Encephalopathy Using In Silico Prediction Tools and Computational Analysis

Mutations in the CDKL5 gene are associated with developmental and epileptic encephalopathy (DEE), a severe disorder characterized by developmental delay and epileptic activity. In genetic analyses of DEEs, variants classified as pathogenic confirm the diagnosis of the disease while Variants of Uncertain Significance (VUS) remain in a gray area due to insufficient evidence. This study aimed to optimize the interpretation of VUS in the CDKL5 gene by evaluating the performance of 22 in silico prediction tools using 186 known pathogenic or benign missense variants from the ClinVar database. The best-performing tools were then applied to analyze CDKL5 VUS variants, complemented by the evaluation of evolutionary conservation, structural analyses, and molecular dynamics simulations to assess their impact on protein structure and function. The results identified SNPred as the most reliable tool, achieving 100% accuracy, sensitivity, and specificity. Other high-performing tools, including ESM-1v, AlphaMissense, EVE, and ClinPred, demonstrated over 98% accuracy. Among 44 CDKL5 VUS variants evaluated, 20 were initially classified as pathogenic by these tools. However, further evaluation using stringent criteria-incorporating conservation scores, structural disruptions identified by Missense3D and PyMol, and molecular dynamics simulation results-led to the reclassification of 8 VUS variants as "potentially pathogenic" and the remaining 12 as "variants with conflicting data". This comprehensive approach provides a robust framework for the classification of VUS in the CDKL5 gene, offering critical insights for accurate diagnosis and treatment strategies in DEE. These findings will serve as a valuable resource for clinicians and geneticists in resolving the diagnostic ambiguity associated with VUS.

The C-terminal structure of the N6-methyladenosine deaminase YerA and its role in deamination

The N6-methyladenine (6mA) modification is an essential epigenetic marker and plays a crucial role in processes, such as DNA repair, replication, and gene expression regulation. YerA from Bacillus subtilis is considered a novel class of enzymes that are capable of catalyzing the deamination of 6mA to produce hypoxanthine. Despite the significance of this type of enzymes in bacterial self-defense system and potential applications as a gene-editing tool, the substrate specificity, catalytic mechanism, and physiological function of YerA are currently unclear due to the lack of structural information. In the present study, we expressed the recombinant enzyme and conducted its reconstitution to yield the active form. Our deamination assays showed that N6-methyladenosine (N6-mAdo) served as a more favorable substrate than its base derivative 6mA. Here, we report the high-resolution structure of the C-terminal region of YerA, which exhibited a compact architecture composed of two antiparallel β-sheets with no obvious close structural homologs in Protein Data Bank. We also created docking models to investigate the ligand-binding pattern and found that more favorable contacts of N6-mAdo with the enzyme-binding pocket lead to its preference for N6-mAdo over 6mA. Finally, structural comparison of the N6-methyladenosine monophosphate deaminase allowed us to propose that a plausible role for this C-terminal region is to shield the active site from solvent and protect the intermediate during catalysis. Taken together, this study sheds light on the catalytic mechanism and evolutionary pathways of the promiscuous enzyme YerA, thereby contributing to our molecular understanding of epigenetic nucleoside metabolism.

Functional characterization of all CDKN2A missense variants and comparison to in silico models of pathogenicity

Interpretation of variants identified during genetic testing is a significant clinical challenge. In this study, we developed a high-throughput CDKN2A functional assay and characterized all possible CDKN2A missense variants. We found that 17.7% of all missense variants were functionally deleterious. We also used our functional classifications to assess the performance of in silico models that predict the effect of variants, including recently reported models based on machine learning. Notably, we found that all in silico models performed similarly when compared to our functional classifications with accuracies of 39.5-85.4%. Furthermore, while we found that functionally deleterious variants were enriched within ankyrin repeats, we did not identify any residues where all missense variants were functionally deleterious. Our functional classifications are a resource to aid the interpretation of CDKN2A variants and have important implications for the application of variant interpretation guidelines, particularly the use of in silico models for clinical variant interpretation.

Unraveling diversity by isolating peptide sequences specific to distinct taxonomic groups

The identification of succinct, universal fingerprints that enable the characterization of individual taxonomies can reveal insights into trait development and can have widespread applications in pathogen diagnostics, human healthcare, ecology and the characterization of biomes. Here, we investigated the existence of peptide k-mer sequences that are exclusively present in a specific taxonomy and absent in every other taxonomic level, termed taxonomic quasi-primes. By analyzing proteomes across 24,073 species, we identified quasi-prime peptides specific to superkingdoms, kingdoms, and phyla, uncovering their taxonomic distributions and functional relevance. These peptides exhibit remarkable sequence uniqueness at six- and seven-amino-acid lengths, offering insights into evolutionary divergence and lineage-specific adaptations. Moreover, we show that human quasi-prime loci are more prone to harboring pathogenic variants, underscoring their functional significance. This study introduces taxonomic quasi-primes and offers insights into their contributions to proteomic diversity, evolutionary pathways, and functional adaptations across the tree of life, while emphasizing their potential impact on human health and disease.

Genomics of pediatric cardiomyopathy

Cardiomyopathy in children is a leading cause of heart failure and cardiac transplantation. Disease-associated genetic variants play a significant role in the development of the different subtypes of disease. Genetic testing is increasingly being recognized as the standard of care for diagnosing this heterogeneous group of disorders, guiding management, providing prognostic information, and facilitating family-based risk stratification. The increase in clinical and research genetic testing within the field has led to new insights into this group of disorders. Mutations in genes encoding sarcomere, cytoskeletal, Z-disk, and sarcolemma proteins appear to play a major role in causing the overlapping clinical phenotypes called cardioskeletal myopathies through "final common pathway" links. For myocarditis, the high frequency of infectious exposures and wide spectrum of presentation suggest that genetic factors mediate the development and course of the disease, including genetic risk alleles, an association with cardiomyopathy, and undiagnosed arrhythmogenic cardiomyopathy. Finally, while we have made strides in elucidating the genetic architecture of pediatric cardiomyopathy, understanding the clinical implications of variants of uncertain significance remains a major issue. The need for continued genetic innovation in this field remains great, particularly as a basis to drive forward targeted precision medicine and gene therapy efforts. IMPACT: Cardiomyopathy and skeletal myopathy can occur in the same patient secondary to gene mutations that encode for sarcomeric or cytoskeletal proteins, which are expressed in both muscle groups, highlighting that there are common final pathways of disease. The heterogeneous presentation of myocarditis is likely secondary to a complex interaction of multiple environmental and genetic factors, suggesting a utility to genetic testing in pediatric patients with myocarditis, particularly those in higher risk groups. Given the high prevalence of variants of uncertain significance in genetic testing, better bioinformatic tools and pipelines are needed to resolve their clinical meaning.

Unlocking precision medicine: clinical applications of integrating health records, genetics, and immunology through artificial intelligence

Artificial intelligence (AI) has emerged as a transformative force in precision medicine, revolutionizing the integration and analysis of health records, genetics, and immunology data. This comprehensive review explores the clinical applications of AI-driven analytics in unlocking personalized insights for patients with autoimmune rheumatic diseases. Through the synergistic approach of integrating AI across diverse data sets, clinicians gain a holistic view of patient health and potential risks. Machine learning models excel at identifying high-risk patients, predicting disease activity, and optimizing therapeutic strategies based on clinical, genomic, and immunological profiles. Deep learning techniques have significantly advanced variant calling, pathogenicity prediction, splicing analysis, and MHC-peptide binding predictions in genetics. AI-enabled immunology data analysis, including dimensionality reduction, cell population identification, and sample classification, provides unprecedented insights into complex immune responses. The review highlights real-world examples of AI-driven precision medicine platforms and clinical decision support tools in rheumatology. Evaluation of outcomes demonstrates the clinical benefits and impact of these approaches in revolutionizing patient care. However, challenges such as data quality, privacy, and clinician trust must be navigated for successful implementation. The future of precision medicine lies in the continued research, development, and clinical integration of AI-driven strategies to unlock personalized patient care and drive innovation in rheumatology.

Systematic characterization of indel variants using a yeast-based protein folding sensor

Gene variants resulting in insertions or deletions of amino acid residues (indels) have important consequences for evolution and are often linked to disease, yet, compared to missense variants, the effects of indels are poorly understood and predicted. We developed a sensitive protein folding sensor based on the complementation of uracil auxotrophy in yeast by circular permutated orotate phosphoribosyltransferase (CPOP). The sensor reports on the folding of disease-linked missense variants and de-novo-designed proteins. Applying the folding sensor to a saturated library of single-residue indels in human dihydrofolate reductase (DHFR) revealed that most regions that tolerate indels are confined to internal loops, the termini, and a central α helix. Several indels are temperature sensitive, and folding is rescued upon binding to methotrexate. Rosetta and AlphaFold2 predictions correlate with the observed effects, suggesting that most indels destabilize the native fold and that these computational tools are useful for the classification of indels observed in population sequencing.

Two de novo UBR1 variants in trans as a cause of Johanson-Blizzard syndrome

AIMS/BACKGROUND

Johanson-Blizzard syndrome (JBS) is a rare autosomal recessive disease caused by pathogenic variants in the UBR1 gene. JBS is usually suspected based on characteristic anomalies, but only genetic testing provides a definitive diagnosis. Since most variants are inherited from the parents, we aimed to identify the causal variants in a Czech proband with clinically suspected JBS and perform segregation analysis.

METHODS

A proband with clinically suspected JBS underwent clinical exome sequencing (CES). Sanger sequencing was used for the validation, characterization, and segregation of variants in the family. The variants were also characterized using quantitative real-time PCR (qPCR) and in silico analysis.

RESULTS

Using CES in the proband, we identified two novel causal variants in the UBR1 gene, c.3482A>C and c.3509+6T>C. Although the variants were found in trans, neither was detected in the parents. Sanger sequencing of the cDNA revealed that the novel variant c.3509+6T>C caused activation of the non-canonical GC donor splice site. The inclusion of 70 bp of the intronic sequence generated a frameshift and a premature termination codon leading to nonsense-mediated decay, as detected by qPCR. In silico protein structural analysis showed that the novel missense variant c.3482A>C in the zinc-stabilized domain RING-H2 altered a highly conserved zinc-coordinating histidine by proline.

CONCLUSION

To the best of our knowledge, we report the first molecular confirmation of JBS in the Czech Republic and the first identification of two de novo causal variants in two alleles. Our findings also expand the spectrum of pathogenic variants in the UBR1 gene.

Sam-Sam Association Between EphA2 and SASH1: In Silico Studies of Cancer-Linked Mutations

Recently, SASH1 has emerged as a novel protein interactor of a few Eph tyrosine kinase receptors like EphA2. These interactions involve the first N-terminal Sam (sterile alpha motif) domain of SASH1 (SASH1-Sam1) and the Sam domain of Eph receptors. Currently, the functional meaning of the SASH1-Sam1/EphA2-Sam complex is unknown, but EphA2 is a well-established and crucial player in cancer onset and progression. Thus, herein, to investigate a possible correlation between the formation of the SASH1-Sam1/EphA2-Sam complex and EphA2 activity in cancer, cancer-linked mutations in SASH1-Sam1 were deeply analyzed. Our research plan relied first on searching the COSMIC database for cancer-related SASH1 variants carrying missense mutations in the Sam1 domain and then, through a variety of bioinformatic tools and molecular dynamic simulations, studying how these mutations could affect the stability of SASH1-Sam1 alone, leading eventually to a defective fold. Next, through docking studies, with the support of AlphaFold2 structure predictions, we investigated if/how mutations in SASH1-Sam1 could affect binding to EphA2-Sam. Our study, apart from presenting a solid multistep research protocol to analyze structural consequences related to cancer-associated protein variants with the support of cutting-edge artificial intelligence tools, suggests a few mutations that could more likely modulate the interaction between SASH1-Sam1 and EphA2-Sam.

Responsible AI in biotechnology: balancing discovery, innovation and biosecurity risks

The integration of artificial intelligence (AI) in protein design presents unparalleled opportunities for innovation in bioengineering and biotechnology. However, it also raises significant biosecurity concerns. This review examines the changing landscape of bioweapon risks, the dual-use potential of AI-driven bioengineering tools, and the necessary safeguards to prevent misuse while fostering innovation. It highlights emerging policy frameworks, technical safeguards, and community responses aimed at mitigating risks and enabling responsible development and application of AI in protein design.

Higher throughput assays for understanding the pathogenicity of variants of unknown significance (VUS) in the RPE65 gene

Purpose

RPE65 is a key enzyme in the visual cycle that regenerates 11-cis retinal. Mutations in RPE65 cause a retinal dystrophy that is treatable with an FDA-approved gene therapy. Variants of unknown significance (VUS) on genetic testing can prevent patients from obtaining a firm genetic diagnosis and accessing gene therapy. Since most RPE65 mutations have a low protein expression level, this study developed and validated multiple methods for assessing the expression level of RPE65 variants. This functional evidence is expected to aid in reclassifying RPE65 VUS as pathogenic, which in turn can broaden the application of gene therapy for RPE65 patients.

Methods

30 different variants of RPE65 (12 pathogenic, 13 VUS, 5 benign) were cloned into lentiviral expression vectors. Protein expression levels were measured after transient transfection or in stable cell lines, using Western blots and immunostaining with flow cytometry. Then, a pooled, high throughput, fluorescence-activated cell sorting (FACS) assay with an NGS-based sequencing readout was used to assay pools of RPE65 variants.

Results

There was a high correlation between protein levels measured by Western blot, flow cytometry, and the pooled FACS assay. Using these assays, we confirm and extend RPE65 variant data, including that Pro111Ser has a low, pathogenic expression level. There was a high correlation between RPE65 expression and previously reported enzyme activity levels; further development of a high throughput enzymatic activity assay would complement this expression data.

Conclusion

This scalable approach can be used to solve patient pedigrees with VUS in RPE65, facilitating treatment and providing RPE65 structure-function information.