Accurate proteome-wide missense variant effect prediction with AlphaMissense

STAT5B leukemic mutations, altering SH2 tyrosine 665, have opposing impacts on immune gene programs

STAT5B is a vital transcription factor for lymphocytes. Here, the function of two STAT5B mutations from human T-cell leukemias: one substituting tyrosine 665 with phenylalanine (STAT5BY665F) and the other with histidine (STAT5BY665H), was interrogated. In silico modeling predicted divergent energetic effects on homodimerization with a range of pathogenicity. In primary T cells in vitro, STAT5BY665F showed gain-of-function, whereas STAT5BY665H demonstrated loss-of-function. Introducing the mutation into the mouse genome illustrated that the gain-of-function Stat5b Y665F mutation resulted in accumulation of CD8+ effector and memory and CD4+ regulatory T cells, altering CD8+/CD4+ ratios. In contrast, STAT5BY665H "knock-in" mice showed diminished CD8+ effector and memory and CD4+ regulatory T cells. In contrast to WT STAT5B, the STAT5BY665F variant displayed greater STAT5 phosphorylation, DNA binding, and transcriptional activity after cytokine activation, whereas the STAT5BY665H variant resembled a null. The work exemplifies how joining in silico and in vivo studies of single nucleotides deepens our understanding of disease-associated variants, revealing structural determinants of altered function, defining mechanistic roles, and, specifically here, identifying a gain-of-function variant that does not directly induce hematopoietic malignancy.

Computer aided design of CGA-N9 derived peptides based on oligopeptide transporters and their antifungal evaluations

CGA-N9 is an antifungal peptide that primarily targets Candida spp. with a mild activity. Our preceding research confirmed that the CGA-N9 crosses cell membrane with the assistance of C. tropicalis oligopeptide transporter (CtOPT) -1 and - 9. In this study, CGA-N9-derived peptides were designed following the molecular docking results with CtOPT-1 and -9. Compared with CGAN9, they exhibit higher transmembrane efficiency with the assistance of CtOPT-1 during the early phase of transmembrane processes and CtOPT-9 in the late phase. And they displayed significantly enhanced antifungal activity, with lower minimum inhibitory concentrations (MICs) against C. tropicalis, C. albicans, and C. parapsilosis, as well as improved biosafety. Among them, CGAN93 was the most optimizing, with a therapeutic index of 145.33. Furthermore, in a mouse model of systemic candidiasis, CGAN93 demonstrated a therapeutic effect comparable to fluconazole, significantly improving the survival rate of mice, attenuating organ damage, and enhancing the immune organ index. In conclusion, OPTs-based computer aided design is an effective strategy for enhancing the activities of antimicrobial peptides (AMPs) by improving transmembrane transport efficiency. CGAN93 is a promising drug candidate for treating Candidiasis.

Detection of protein structural hotspots using AI distillation and explainability: application to the DAX-1 protein

AlphaMissense is a valuable resource for discerning important functional regions within proteins, providing pathogenicity heatmaps that highlight the pathogenic risk of specific mutations along the protein sequence. However, due to protein folding and long-range interactions, the actual structural alterations with functional implications may be occurring at a distance from the mutation site. As a result, the identification of the most sensitive structural regions for protein function may be hampered by the presence of mutations that indirectly affect the critical regions from a distance. In this study, we illustrate how the use of AlphaMissense predictions to train an XGBoost regression model on structural features extracted from the structures of protein variants predicted by OmegaFold enables the definition of a new explainability metric: a residue-based importance score that highlights the most critical structural domains within a protein sequence. To verify the accuracy of our approach, we applied it to the extensively studied protein DAX-1 and successfully identified critical structural domains. Notably, as this score only requires knowledge of the protein's amino acid sequence, it is valuable in guiding experimental investigations aimed at discovering functionally crucial regions in proteins that have been poorly characterized.

Analysis of Rare Coding Variants in 470,000 UK Biobank Participants Reveals Genetic Associations With Childhood Asthma Predisposition

Previous studies of genetic contributions to risk of childhood asthma have implicated common variants with small effect sizes. Some studies using exome sequence data have reported associations with rare coding variants having larger effects on risk, notably an analysis of 145,000 subjects which found association with loss of function (LOF) variants in FLG, the gene coding for filaggrin. Here, we report the results of an analysis of rare nonsynonymous and LOF variants, carried out on the full UK Biobank cohort of 470,000 exome-sequenced participants. The phenotype of childhood asthma was defined as reporting asthma with onset before 18. Regression analysis was applied to gene-wise tests for association of LOF and nonsynonymous variants. Forty-five tests using different pathogenicity predictors were applied to the first cohort of 200,000 participants. Subsequently, the 100 genes showing strongest evidence for association were analysed in the second cohort of 270,000 participants, using only the best-performing predictor for each gene. For FLG, separate analyses were carried out for participants with atopic dermatitis. Three genes achieved statistical significance after correction for testing these 100 genes: FLG, IL33 and PRKCQ. The effects on asthma risk and frequencies of variants in different functional categories were characterised for these genes. Damaging coding variants were associated with increased risk of asthma in FLG and IL33 but reduced risk in PRKCQ. FLG LOF variants were also associated with the risk of atopic dermatitis, and their effect on asthma risk was higher in people who reported a diagnosis of atopic dermatitis. Rare coding variants in a small number of genes have important effects on asthma risk. Further study of individual variant effects might elucidate mechanisms of pathogenesis. This research has been conducted using the UK Biobank Resource.

A systematic evaluation of the therapeutic potential of endogenous-ADAR editors in cancer prevention and treatment

Adenosine deaminases acting on RNA (ADAR) enzymes constitute a natural cellular mechanism that induces A-to-I(G) editing, introducing genetic changes at the RNA level. Recently, interest in the endogenous-ADAR editor has emerged for correcting genetic mutations, consisting of a programmed oligonucleotide that attracts the native ADAR, thereby offering opportunities for medical therapy. Here, we systematically chart the scope of cancer mutations that endogenous-ADAR can correct. First, analyzing germline single nucleotide variants in cancer predisposition genes, we find that endogenous-ADAR can revert a fifth of them, reducing the risk of cancer development later in life. Second, examining somatic mutations across various cancer types, we find that it has the potential to correct at least one driver mutation in over a third of the samples, suggesting a promising future treatment strategy. We also highlight key driver mutations that are amenable to endogenous-ADAR, and are thus of special clinical interest. As using endogenous-ADAR entails delivering relatively small payloads, the prospects of delivering endogenous-ADAR to various cancers seem promising. We expect that the large scope of correctable mutations that are systematically charted here for the first time will pave the way for a new era of cancer treatment options.

Enhanced genotype-phenotype analysis using multimodal adaptive optics and 3D protein structure in Bietti Crystalline Dystrophy

Purpose

Deep phenotyping of genetic retinal disease using multimodal adaptive optics ophthalmoscopy and protein structure variant analysis.

Observations

In a patient with extensive atrophy of the retinal pigment epithelium and yellow deposits in the retina, genetic testing identified two CYP4V2 variants: c.802-8_810delinsGC and c.1169G > A, p.Arg390His. AI-generated protein structures indicated loss of CYP4V2 function. Reflectance confocal and multiple-scattering Adaptive Optics Scanning Light Ophthalmoscopy (AOSLO) captured crystalline deposits throughout the retina as well as previously unreported cyst-like structures that were mainly independent from the crystalline deposits. Sequential AOSLO imaging was conducted and revealed anatomical and morphological changes in the cysts and surrounding cellular structures.

Conclusions and importance

Cyst-like changes may represent a new BCD degenerative feature. Characterizing retinal genetic disease variants with protein structural modeling and phenotyping with AOSLO represents an advanced approach for clinical diagnosis and may serve as a biomarker of disease progression.

Holistic Exome-Based Genetic Testing in Adults With Epilepsy

Background and Objectives

Exome sequencing (ES) is increasingly used in the diagnostic workup of epilepsies. While its utility has been extensively demonstrated in children, its role in adults remains to be defined. In this study, we evaluate the outcomes of a holistic exome-based approach in adults with epilepsy.

Methods

We included 106 adults with epilepsy and a presumed genetic etiology between January 2015 and December 2023 at the Medical University of Vienna, Austria. Diagnostic ES, including copy number variation (CNV) and mitochondrial analyses, was performed. We report on diagnostic outcomes, phenotype expansions, and research findings. Furthermore, we compared the diagnostic outcomes with 3 comprehensive gene panels.

Results

In our cohort, the diagnostic yield was 30.2%, outperforming all 3 simulated gene panels. A developmental and epileptic encephalopathy phenotype was associated with receiving a genetic diagnosis. Overall, 27 distinct molecular etiologies were identified. Eight patients had pathogenic CNVs, and 2 had mitochondrial DNA variants. Molecular diagnoses had potential clinical implications in 8 of 32 solved cases (25%), which were eventually exerted in 5 patients (15.6%). Tailored treatment changes were successfully applied in SCN1A-related epilepsy (discontinuation of sodium channel blockers) and GLUT1 deficiency (ketogenic diet). Three patients with mitochondrial diseases were referred for preventive screening investigations after the genetic diagnosis. Our findings expand the clinical spectrum of 3 known epilepsy genes. In addition, explorative variant prioritization identified heterozygous truncating variants in CLASP1 in 2 unrelated patients with focal epilepsy, suggesting it as a candidate gene.

Discussion

Our study strongly supports the use of holistic genetic approaches, encompassing CNV and mitochondrial analyses, in adults with epilepsy. Similar to pediatric cohorts, results may inform clinical care. Moreover, we report on phenotype expansions and a candidate gene discovery.

An ultra-rare missense variant in the KIF1B gene linked to autoinflammatory Menière's disease

Menière's disease (MD) is an inner ear disorder characterised by episodes of vertigo, sensorineural hearing loss and tinnitus linked to autoinflammation and/or type 2 immune response. We hypothesise that rare variation in immune response genes could drive the autoinflammatory phenotype in MD. We retrieved differentially expressed genes (DEG) from single-cell RNAseq and epigenomic datasets to search for rare variants in the MD exome (N = 454) and genome (N = 511) sequencing datasets. The variant chr1:10374335 C > T in the KIF1B gene was found in three MD unrelated individuals and was predicted to be likely pathogenic. According to differential transcript usage, transcript ENST00000622724.3 was found in MD samples, but absent in controls. Furthermore, this variant may influence splicing through the generation of exonic enhancers and silencers, potentially changing transcription factor binding at the promoter. These findings support that this KIF1B gene rare variant is associated with the MD autoinflammatory phenotype and may up-regulate its expression in monocytes.

A probabilistic graphical model for estimating selection coefficients of nonsynonymous variants from human population sequence data

Accurately predicting the effect of missense variants is important in discovering disease risk genes and clinical genetic diagnostics. Commonly used computational methods predict pathogenicity, which does not capture the quantitative impact on fitness in humans. We develop a method, MisFit, to estimate missense fitness effect using a graphical model. MisFit jointly models the effect at a molecular level ( d ) and a population level (selection coefficient, s ), assuming that in the same gene, missense variants with similar d have similar s . We train it by maximizing probability of observed allele counts in 236,017 individuals of European ancestry. We show that s is informative in predicting allele frequency across ancestries and consistent with the fraction of de novo mutations in sites under strong selection. Further, s outperforms previous methods in prioritizing de novo missense variants in individuals with neurodevelopmental disorders. In conclusion, MisFit accurately predicts s and yields new insights from genomic data.

Unveiling the influence of fastest nobel prize winner discovery: alphafold's algorithmic intelligence in medical sciences

CONTEXT

AlphaFold's advanced AI technology has transformed protein structure interpretation. By predicting three-dimensional protein structures from amino acid sequences, AlphaFold has solved the complex protein-folding problem, previously challenging for experimental methods due to numerous possible conformations. Since its inception, AlphaFold has introduced several versions, including AlphaFold2, AlphaFold DB, AlphaFold Multimer, Alpha Missense, and AlphaFold3, each further enhancing protein structure prediction. Remarkably, AlphaFold is recognized as the fastest Nobel Prize winner in science history. This technology has extensive applications, potentially transforming treatment and diagnosis in medical sciences by reducing drug design costs and time, while elucidating structural pathways of human body systems. Numerous studies have demonstrated how AlphaFold aids in understanding health conditions by providing critical information about protein mutations, abnormal protein-protein interactions, and changes in protein dynamics. Researchers have also developed new technologies and pipelines using different versions of AlphaFold to amplify its potential. However, addressing existing limitations is crucial to maximizing AlphaFold's capacity to redefine medical research. This article reviews AlphaFold's impact on five key aspects of medical sciences: protein mutation, protein-protein interaction, molecular dynamics, drug design, and immunotherapy.

METHODS

This review examines the contributions of various AlphaFold versions AlphaFold2, AlphaFold DB, AlphaFold Multimer, Alpha Missense, and AlphaFold3 to protein structure prediction. The methods include an extensive analysis of computational techniques and software used in interpreting and predicting protein structures, emphasizing advances in AI technology and its applications in medical research.

Onkopus: precise interpretation and prioritization of sequence variants for biomedical research and precision medicine

One of the major challenges in precision oncology is the identification of pathogenic, actionable variants and the selection of personalized treatments. We present Onkopus, a variant interpretation framework based on a modular architecture, for interpreting and prioritizing genetic alterations in cancer patients. A multitude of tools and databases are integrated into Onkopus to provide a comprehensive overview about the consequences of a variant, each with its own semantic, including pathogenicity predictions, allele frequency, biochemical and protein features, and therapeutic options. We present the characteristics of variants and personalized therapies in a clear and concise form, supported by interactive plots. To support the interpretation of variants of unknown significance (VUS), we present a protein analysis based on protein structures, which allows variants to be analyzed within the context of the entire protein, thereby serving as a starting point for understanding the underlying causes of variant pathogenicity. Onkopus has the potential to significantly enhance variant interpretation and the selection of actionable variants for identifying new targets, drug screens, drug testing using organoids, or personalized treatments in molecular tumor boards. We provide a free public instance of Onkopus at https://mtb.bioinf.med.uni-goettingen.de/onkopus.

Perspectives on the Current and Future State of Artificial Intelligence in Medical Genetics

Artificial intelligence (AI) is rapidly transforming numerous aspects of daily life, including clinical practice and biomedical research. In light of this rapid transformation, and in the context of medical genetics, we assembled a group of leaders in the field to respond to the question about how AI is affecting, and especially how AI will affect, medical genetics. The authors who contributed to this collection of essays intentionally represent different areas of expertise, career stages, and geographies, and include diverse types of clinicians, computer scientists, and researchers. The individual pieces cover a wide range of areas related to medical genetics; we expect that these pieces may provide helpful windows into the ways in which AI is being actively studied, used, and considered in medical genetics.

Mechanisms of PFBA toxicity in Chlorella vulgaris: Photosynthesis, oxidative stress, and antioxidant impairment

Perfluorobutanoic acid (PFBA), an emerging alternative to perfluorooctanoic acid (PFOA), has become increasingly prevalent in aquatic ecosystems, yet its ecotoxicological impacts remain poorly understood. This study investigated the aquatic toxicity of PFBA using the freshwater algae Chlorella vulgaris (C. vulgaris) as a model organism, employing a 96h pre-exposure assay to determine the median effective concentration followed by acute toxicity experiments analyzing multiple endpoints including growth, photosynthetic parameters, oxidative stress markers, and antioxidant enzyme activities. Computer simulation techniques were utilized to illustrate the underlying molecular mechanisms of PFBA toxicity. The results showed that the 96h-EC50 value of PFBA was 154.88 mg/L, which is comparable to conventional per- and polyfluoroalkyl substances (PFAS). Acute toxicity experiments revealed a biphasic dose-response relationship to the algal growth with the hormetic effects at the lower concentrations (30.97-92.93 mg/L) but inhibition at the higher levels (123.91-185.86 mg/L) of PFBA. High dosages of PFBA significantly decreased the maximum photosynthetic yield (Fv/Fm) and relative electron transfer rate (rETR), while inducing oxidative stress and inhibiting superoxide dismutase (SOD) and catalase (CAT) activities. Future AlphaFold2 modeling and molecular docking simulations demonstrated the potential binding of PFBA to photosystem II D1 C-terminal processing protease (PSII D1 protein), SOD, and CAT. These findings reveal a complex toxicity mechanism of PFBA on C. vulgaris involving photosynthetic disruption, oxidative stress, and antioxidant system impairment, contributing to the understanding of short-chain PFAS alternative ecotoxicity in aquatic ecosystems.

AlphaFold-Guided Bespoke Gene Editing Enhances Field-Grown Soybean Oil Contents

Enhancing the oil or protein content of soybean, a major crop for oil and protein production is highly desirable. GmSWEET10a encodes a sugar transporter that is strongly selected during domestication and breeding, increasing seed size and oil content. GmSWEET10b is functionally similar to GmSWEET10a, yet has not been artificially selected. Here, AlphaFold is used to find that C-terminal variants of GmSWEET10a can endow enhanced or reduced transport activity. Guided by AlphaFold, the functionality is improved for GmSWEET10a in terms of oil content through gene editing. Furthermore, novel GmSWEET10b haplotypes possessing strengthened or weakened sugar-transport capabilities that are absent in nature are engineered. Consequently, soybean oil content or protein content in independent GmSWEET10b gene-edited lines during multi-year and multi-site field trials is consistently increased, without negatively affecting yield. The study demonstrates that the combination of AlphaFold-guided protein design and gene editing has the potential to generate novel beneficial alleles, which can optimize protein function in the context of crop breeding.

Genome-wide allele-specific expression in multi-tissue samples from healthy male baboons reveals the transcriptional complexity of mammals

Allele-specific expression (ASE) is pivotal in understanding the genetic underpinnings of phenotypic variation within species, differences in disease susceptibility, and responses to environmental factors. We processed 11 different tissue types collected from 12 age-matched healthy olive baboons (Papio anubis) for genome-wide ASE analysis. By sequencing their genomes at a minimum depth of 30×, we identified over 16 million single-nucleotide variants (SNVs). We also generated long-read sequencing data, enabling the phasing of all variants present within the coding regions of 96.5% of assayable protein-coding genes as a single haplotype block. Given the extensive heterozygosity of baboons relative to humans, we could quantify ASE across 72% of the total annotated protein-coding gene set. We identified genes that exhibit ASE and affect specific tissues and genotypes. We discovered ASE SNVs that also exist in human populations with identical alleles and that are designated as pathogenic by both the PrimateAI-3D and AlphaMissense models.

DiffInvex identifies evolutionary shifts in driver gene repertoires during tumorigenesis and chemotherapy

Somatic cells can transform into tumors due to mutations, and the tumors further evolve towards increased aggressiveness and therapy resistance. We develop DiffInvex, a framework for identifying changes in selection acting on individual genes in somatic genomes, drawing on an empirical mutation rate baseline derived from non-coding DNA that accounts for shifts in neutral mutagenesis during cancer evolution. We apply DiffInvex to >11,000 somatic whole-genome sequences from ~30 cancer types or healthy tissues, identifying genes where point mutations are under conditional positive or negative selection during exposure to specific chemotherapeutics, suggesting drug resistance mechanisms occurring via point mutation. DiffInvex identifies 11 genes exhibiting treatment-associated selection for different classes of chemotherapies, linking selected mutations in PIK3CA, APC, MAP2K4, SMAD4, STK11 and MAP3K1 with drug exposure. Various gene-chemotherapy associations are further supported by differential functional impact of mutations pre- versus post-therapy, and are also replicated in independent studies. In addition to nominating drug resistance genes, we contrast the genomes of healthy versus cancerous cells of matched human tissues. We identify noncancerous expansion-specific drivers, including NOTCH1 and ARID1A. DiffInvex can also be applied to diverse analyses in cancer evolution to identify changes in driver gene repertoires across time or space.

X-linked competition - implications for human development and disease

During early mammalian female development, X chromosome inactivation leads to random transcriptional silencing of one of the two X chromosomes. This inactivation is maintained through subsequent cell divisions, leading to intra-individual diversity, whereby cells express either the maternal or paternal X chromosome. Differences in X chromosome sequence content can trigger competitive interactions between clones that may alter organismal development and skew the representation of X-linked sequence variants in a cell-type-specific manner - a recently described phenomenon termed X-linked competition in analogy to existing cell competition paradigms. Skewed representation can define the phenotypic impact of X-linked variants, for example, the manifestation of disease in female carriers of X-linked disease alleles. Here, we review what is currently known about X-linked competition, reflect on what remains to be learnt and map out the implications for X-linked human disease.

EVOLVE: A Web Platform for AI-Based Protein Mutation Prediction and Evolutionary Phase Exploration

While predicting structure-function relationships from sequence data is fundamental in biophysical chemistry, identifying prospective single-point and collective mutation sites in proteins can help us stay ahead in understanding their potential effects on protein structure and function. Addressing the challenges of large sequence-space analysis, we present EVOLVE, a web tool enabling researchers to explore prospective mutation sites and their collective behavior. EVOLVE integrates a statistical mechanics-guided machine learning algorithms to predict probable mutational sites, with statistical mechanics calculating mutational entropy to accurately identify mutational hotspots. Validation against a number of viral protein sequences confirms its ability to predict mutations and their functional consequences. By leveraging statistical mechanics of phase transition concept, EVOLVE also quantifies mutational entropy fluctuations, offering a quantitative foundation for identifying Variants of Concern (VOC) or Variants under Monitoring (VUM) as per World Health Organization (WHO) guidelines. EVOLVE streamlines data upload and analysis with a user-friendly interface and comprehensive tutorials. Access EVOLVE free at https://evolve-iiserkol.com.

Computing pathogenicity of mutations in human cytochrome P450 superfamily

Cytochrome P450 (CYPs) are crucial heme-containing enzymes that metabolize drugs and endogenous compounds. In humans, 57 CYP isoforms have been identified, with over 200 mutations linked to severe disorders. Our comprehensive computational study assessed the reason for the pathogenicity of mutations by comparing pathogenic and non-pathogenic variants. We analyzed 25,94,151 mutations across 26 CYP structures using structure- and sequence-based methods, revealing a meaningful stability pattern: non-pathogenic > all > pathogenic mutation datasets. Notably, pathogenic mutations were predominantly buried within CYP structures, indicating a higher potential for pathogenesis. We identified three key amino acid properties affected by mutations: Gibbs free energy, isoelectric point, and volume. Furthermore, diseased mutations significantly reduced positive residue content, particularly due to arginine mutations, which directly influenced the isoelectric point. Our findings indicate a greater likelihood of pathogenic mutations occurring at conserved sites, disrupting CYP function. A higher frequency of pathogenic mutations was observed in heme sites, primarily involving arginine, which may interfere with arginine-heme interactions. Molecular docking revealed a differential binding of heme in wild-type and pathogenic CYPs. This study provides a foundational analysis of mutation effects across multiple CYPs. It models the chemical basis of CYP-related pathogenicity, facilitating the development of a semi-quantitative disease prediction model.

Protein-truncating variants in UQCRC1 are associated with Parkinson's disease: evidence from half-million people

Recent studies have suggested a potential but inconsistent link between UQCRC1 and Parkinson's disease (PD). For the first time, we systematically investigated the association between non-synonymous variants in UQCRC1 and PD risk using data from the UK Biobank with half-million participants, which provide evidence supporting the role of UQCRC1 Protein-truncating variants (PTVs) in PD (P = 1.20 × 10-6, OR = 6.59) and highlight the importance of large-scale population studies in identifying rare genetic risk factors.

Androgens mediate sexual dimorphism in Pilarowski-Bjornsson Syndrome

Sex-specific penetrance in autosomal dominant Mendelian conditions is largely understudied. The neurodevelopmental disorder Pilarowski-Bjornsson syndrome (PILBOS) was initially described in females. Here, we describe the clinical and genetic characteristics of the largest PILBOS cohort to date, showing that both sexes can exhibit PILBOS features, although males are overrepresented. A mouse model carrying a human-derived Chd1 missense variant (Chd1 R616Q/+) displays female-restricted phenotypes, including growth deficiency, anxiety and hypotonia. Orchiectomy unmasks a growth deficiency phenotype in male Chd1 R616Q/+ mice, while testosterone rescues the phenotype in females, implicating androgens in phenotype modulation. In the gnomAD and UK Biobank databases, rare missense variants in CHD1 are overrepresented in males, supporting a male protective effect. We identify 33 additional highly constrained autosomal genes with missense variant overrepresentation in males. Our results support androgen-regulated sexual dimorphism in PILBOS and open novel avenues to understand the mechanistic basis of sexual dimorphism in other autosomal Mendelian disorders.

Machine learning approaches enable the discovery of therapeutics across domains

Multi-modal datasets have grown exponentially in the last decade. This has created an enormous demand for machine learning models that can predict complex outcomes by leveraging cellular, molecular, and humoral profiles. Corresponding inference of mechanisms can help to uncover new therapeutic targets. Here, we discuss how biological principles guide the design of predictive models and how interpretable machine learning can lead to novel mechanistic insights. We provide descriptions of multiple learning techniques and how suited they are to domain adaptations. Finally, we talk about broad learning capabilities of foundation models on large datasets and whether they can be used to provide meaningful inference about biological datasets.

Accurate identification and mechanistic evaluation of pathogenic missense variants with Rhapsody-2

Understanding the effects of missense mutations or single amino acid variants (SAVs) on protein function is crucial for elucidating the molecular basis of diseases/disorders and designing rational therapies. We introduce here Rhapsody-2, a machine learning tool for discriminating pathogenic and neutral SAVs, significantly expanding on a precursor limited by the availability of structural data. With the advent of AlphaFold2 as a powerful tool for structure prediction, Rhapsody-2 is trained on a significantly expanded dataset of 117,525 SAVs corresponding to 12,094 human proteins reported in the ClinVar database. Adopting a broad set of descriptors composed of sequence evolutionary, structural, dynamic, and energetics features in the training algorithm, Rhapsody-2 achieved an AUROC of 0.94 in 10-fold cross-validation when all SAVs of a particular test protein (mutant) were excluded from the training set. Benchmarking against a variety of testing datasets demonstrated the high performance of Rhapsody-2. While sequence evolutionary descriptors play a dominant role in pathogenicity prediction, those based on structural dynamics provide a mechanistic interpretation. Notably, residues involved in allosteric communication and those distinguished by pronounced fluctuations in the high-frequency modes of motion or subject to spatial constraints in soft modes usually give rise to pathogenicity when mutated. Overall, Rhapsody-2 provides an efficient and transparent tool for accurately predicting the pathogenicity of SAVs and unraveling the mechanistic basis of the observed behavior, thus advancing our understanding of genotype-to-phenotype relations.

CRISPR genome editing using a combined positive and negative selection system

The clustered regularly interspaced short palindromic repeats (CRISPR)/Cas system is a powerful genome editing tool that has revolutionized research. Single nucleotide polymorphisms (SNPs) are the most common form of genetic variation in humans. Only a subset of these SNPs has been shown to be linked to genetic diseases, while the biological relevance of the majority remains unclear. Investigating these variants of unknown significance could provide valuable insights into their roles in biological processes, disease susceptibility, and treatment responses. While CRISPR/Cas has emerged as a transformative technology, its ability to make single nucleotide substitutions remains a significant limitation. Other techniques in single nucleotide editing, such as base editing and prime editing, offer promising possibilities to complement CRISPR/Cas systems, though they also have their own limitations. Hence, alternative approaches are necessary to overcome the limitations of CRISPR. Here, to improve the feasibility of generating single base edits in the genome, we provide a protocol that introduces a multiple expression and dual selection (MEDS) system, which, alongside CRISPR, utilizes the opposing roles of cytosine deaminase/uracil phosphoribosyltransferase (CD/UPRT) for negative selection and neomycin phosphotransferase II (NPT II) for positive selection. As a proof of concept and to demonstrate feasibility of the method, we used MEDS, along with traditional CRISPR-Cas9, to generate sickle hemoglobin by introducing a point mutation (A → T) in the sixth codon of the hemoglobin beta gene.

PRESCOTT: a population aware, epistatic, and structural model accurately predicts missense effects

Predicting the functional impact of point mutations is a critical challenge in genomics. PRESCOTT reconstructs complete mutational landscapes, identifies mutation-sensitive regions, and categorizes missense variants as benign, pathogenic, or variants of uncertain significance. Leveraging protein sequences, structural models, and population-specific allele frequencies, PRESCOTT surpasses existing methods in classifying ClinVar variants, the ACMG dataset, and over 1800 proteins from the Human Protein Dataset. Its online server facilitates mutation effect predictions for any protein and variant, and includes a database of over 19,000 human proteins, ready for population-specific analyses. Open access to residue-specific scores offers transparency and valuable insights for genomic medicine.

Evaluating variant pathogenicity prediction tools to establish African inclusive guidelines for germline genetic testing

BACKGROUND

Genetic germline testing is restricted for African patients. Lack of ancestrally relevant genomic data perpetuated by African diversity has resulted in European-biased curated clinical variant databases and pathogenic prediction guidelines. While numerous variant pathogenicity prediction tools (VPPTs) exist, their performance has yet to be established within the context of African diversity.

METHODS

To address this limitation, we assessed 54 VPPTs for predictive performance (sensitivity, specificity, false positive and negative rates) across 145,291 known pathogenic or benign variants derived from 50 Southern African and 50 European men matched for advanced prostate cancer. Prioritising VPPTs for optimal ancestral performance, we screened 5.3 million variants of unknown significance for predicted functional and oncogenic potential.

RESULTS

We observe a 2.1- and 4.1-fold increase in the number of known and predicted rare pathogenic or benign variants, respectively, against a 1.6-fold decrease in the number of available interrogated variants in our European over African data. Although sensitivity was significantly lower for our African data overall (0.66 vs 0.71, p = 9.86E-06), MetaSVM, CADD, Eigen-raw, BayesDel-noAF, phyloP100way-vertebrate and MVP outperformed irrespective of ancestry. Conversely, MutationTaster, DANN, LRT and GERP-RS were African-specific top performers, while MutationAssessor, PROVEAN, LIST-S2 and REVEL are European-specific. Using these pathogenic prediction workflows, we narrow the ancestral gap for potentially deleterious and oncogenic variant prediction in favour of our African data by 1.15- and 1.1-fold, respectively.

CONCLUSION

Although VPPT sensitivity favours European data, our findings provide guidelines for VPPT selection to maximise rare pathogenic variant prediction for African disease studies.

Identifying genetic errors of immunity due to mosaicism

Inborn errors of immunity are monogenic disorders of the immune system that lead to immune deficiency and/or dysregulation in patients. Identification of precise genetic causes of disease aids diagnosis and advances our understanding of the human immune system; however, a significant portion of patients lack a molecular diagnosis. Somatic mosaicism, genetic changes in a subset of cells, is emerging as an important mechanism of immune disease in both young and older patients. Here, we review the current landscape of somatic genetic errors of immunity and methods for the detection and validation of somatic variants.

A series of reviews in familial cancer: genetic cancer risk in context variants of uncertain significance in MMR genes: which procedures should be followed?

Interpreting variants of uncertain significance (VUS) in mismatch repair (MMR) genes remains a major challenge in managing Lynch syndrome and other hereditary cancer syndromes. This review outlines recommended VUS classification procedures, encompassing foundational and specialized methodologies tailored for MMR genes by expert organizations, including InSiGHT and ClinGen's Hereditary Colorectal Cancer/Polyposis Variant Curation Expert Panel (VCEP). Key approaches include: (1) functional data, encompassing direct assays measuring MMR proficiency such as in vitro MMR assays, deep mutational scanning, and MMR cell-based assays, as well as techniques like methylation-tolerant assays, proteomic-based approaches, and RNA sequencing, all of which provide critical functional evidence supporting variant pathogenicity; (2) computational data/tools, including in silico meta-predictors and models, which contribute to robust VUS classification when integrated with experimental evidence; and (3) enhanced variant detection to identify the actual causal variant through whole-genome sequencing and long-read sequencing to detect pathogenic variants missed by traditional methods. These strategies improve diagnostic precision, support clinical decision-making for Lynch syndrome, and establish a flexible framework that can be applied to other OMIM-listed genes.

Further Delineation of the AUTS2 HX Repeat Domain-Related Phenotype

Haploinsufficiency of AUTS2 is associated with a neurodevelopmental disorder characterized by intellectual disability, autistic features, and spasticity. AUTS2 protein interacts with p300, encoded by EP300, through the HX repeat domain of AUTS2, thereby activating transcription. We previously reported two de novo variants in the HX repeat domain of AUTS2. These variants disrupt the AUTS2-P300 interaction, resulting in a phenotype resembling Rubinstein-Taybi Syndrome (RSTS) associated with variants in EP300/CREBBP. Here, we expand beyond the initial clinical description to delineate the HX domain-associated phenotype and compare it to the AUTS2-haploinsufficient phenotype. We reviewed clinical data, photographs, and neuroimaging studies to examine genotype-phenotype relationships. Our review of 80 individuals included 14 individuals we present here and 66 individuals with AUTS2 variants presented in the literature. The clinical features for individuals with variants in the HX repeat domain include severe intellectual disability, severe language disability, distinct craniofacial and skeletal dysmorphic features, and neuroimaging findings. Facial dysmorphisms include wide and prominent nasal bridges with complex nasal shapes and dysmorphic eyebrows. Dysmorphisms include digit anomalies: Symphalangism and hypoplasia of distal phalanges, exclusive to the HX domain variant group. Cerebellar anomalies not seen with other AUTS2 variants are seen within this group. Our report delineates a distinct and severe clinical phenotype associated with variants in the AUTS2 HX domain, including an in-depth comparison with the AUTS2 haploinsufficiency phenotype features.

Analysis of canine gene constraint identifies new variants for orofacial clefts and stature

Dog breeding promotes within-group homogeneity through conformation to strict breed standards, while simultaneously driving between-group heterogeneity. There are over 350 recognized dog breeds that provide the foundation for investigating the genetic basis of phenotypic diversity. Typically, breed standard phenotypes such as stature, pelage, and craniofacial structure are analyzed through genetic association studies. However, such analyses are limited to assayed phenotypes only, leaving difficult-to-measure phenotypic subtleties easily overlooked. We investigated coding variation from over 2000 dogs, leading to discoveries of variants related to craniofacial morphology and stature. Breed-enriched variants were prioritized according to gene constraint, which was calculated using a mutation model derived from trinucleotide substitution probabilities. Among the newly found variants is a splice-acceptor variant in PDGFRA associated with bifid nose, a characteristic trait of Çatalburun dogs, implicating the gene's role in midline closure. Two additional LCORL variants, both associated with canine body size are also discovered: a frameshift that causes a premature stop in large breeds (>25 kg) and an intronic substitution found in small breeds (<10 kg), thus highlighting the importance of allelic heterogeneity in selection for breed traits. Most variants prioritized in this analysis are not associated with genomic signatures for breed differentiation, as these regions are enriched for constrained genes intolerant to nonsynonymous variation. This indicates trait selection in dogs is likely a balancing act between preserving essential gene functions and maximizing regulatory variation to drive phenotypic extremes.

Charcot-Marie-Tooth disease type 1E: Clinical Natural History and Molecular Impact of PMP22 Variants

Charcot-Marie-Tooth disease type 1E (CMT1E) is a rare, autosomal dominant peripheral neuropathy caused by missense variants, deletions, and truncations within the peripheral myelin protein-22 (PMP22) gene. CMT1E phenotypes vary depending on the specific variant, ranging from mild to severe, and there is little natural history and phenotypic progression data on individuals with CMT1E. Patients with CMT1E were evaluated during initial and follow-up visits at sites within the Inherited Neuropathy Consortium. Clinical characteristics were obtained from history, neurological exams, and nerve conduction studies. Clinical outcome measures were used to quantify baseline and longitudinal changes, including the Rasch-modified CMT Examination Score version 2 (CMTESv2-R) and the CMT Pediatric Scale (CMTPedS). The trafficking of PMP22 variants in transfected cells was correlated to disease severity. Twenty-four, presumed disease-causing PMP22 variants were identified in 50 individuals from 35 families, including 19 missense variants, three in-frame deletions, and two truncations. Twenty-nine patients presented with delayed walking during childhood. At their baseline evaluation, the mean CMTESv2-R in 46 patients was 16 ± 7.72 (out of 32), and the mean CMTPedS from 17 patients was 28 ± 6.35 (out of 44). Six individuals presented with hearing loss, eleven with scoliosis, three with hip dysplasia, and one with both scoliosis and hip dysplasia. Twenty variants were localized within in transmembrane domains; 31 of 35 individuals with these variants had moderate to severe phenotypes. Three variants were found in the extracellular domain and were associated with milder phenotypes. Reduced expression of PMP22 at the cell surface, and the location of missense variants within in the transmembrane domain correlated with disease severity. Pathogenic PMP22 variants located within the transmembrane regions usually cause a moderate to severe clinical phenotype, beginning in early childhood, and have impaired trafficking to the plasma membrane.

Common cis-regulatory variation modifies the penetrance of pathogenic SHROOM3 variants in craniofacial microsomia

Pathogenic coding variants have been identified in thousands of genes, yet the mechanisms underlying the incomplete penetrance in individuals carrying these variants are poorly understood. In this study, in a cohort of 2009 craniofacial microsomia (CFM) patients of Chinese ancestry and 2625 Han Chinese controls, we identified multiple predicted pathogenic coding variants in SHROOM3 in both CFM patients and healthy individuals. We found that the penetrance of CFM correlates with specific haplotype combinations containing likely pathogenic-coding SHROOM3 variants and CFM-associated expression quantitative trait loci (eQTLs) of SHROOM3 expression. Further investigations implicate specific eQTL combinations, such as rs1001322 or rs344131, in combination with other significant CFM-associated eQTLs, which we term combined eQTL phenotype modifiers (CePMods). We additionally show that rs344131, located within a regulatory enhancer region of SHROOM3, demonstrates allele-specific effects on enhancer activity and thus impacts expression levels of the associated SHROOM3 allele harboring any rare coding variant. Our findings also suggest that CePMods may serve as pathogenic determinants, even in the absence of rare deleterious coding variants in SHROOM3 This highlights the critical role of allelic expression in determining the penetrance and severity of craniofacial abnormalities, including microtia and facial asymmetry. Additionally, using quantitative phenotyping, we demonstrate that both microtia and facial asymmetry are present in two separate Shroom3 mouse models, the severity of which is dependent on gene dosage. Our study establishes SHROOM3 as a likely pathogenic gene for CFM and demonstrates eQTLs as determinants of modified penetrance in the manifestation of the disease in individuals carrying likely pathogenic rare coding variants.

ACMG/AMP interpretation of BRCA1 missense variants: Structure-informed scores add evidence strength granularity to the PP3/BP4 computational evidence

Classification of missense variants is challenging. Lacking compelling clinical and/or functional data, ACMG/AMP lines of evidence are restricted to PM2 (rarity code applied at supporting level) and PP3/BP4 (computational evidence based mostly on multiple-sequence-alignment conservation tools). Currently, the ClinGen ENIGMA BRCA1/2 Variant Curation Expert Panel uses BayesDel to apply PP3/BP4 to missense variants located in the BRCA1 RING/BRCT domains. The ACMG/AMP framework does not refer explicitly to protein structure as a putative source of pathogenic/benign evidence. Here, we tested the value of incorporating structure-based evidence such as relative solvent accessibility (RSA), folding stability (ΔΔG), and/or AlphaMissense pathogenicity to the classification of BRCA1 missense variants. We used MAVE functional scores as proxies for pathogenicity/benignity. We computed RSA and FoldX5.0 ΔΔG predictions using as alternative input templates for either PDB files or AlphaFold2 models, and we retrieved pre-computed AlphaMissense and BayesDel scores. We calculated likelihood ratios toward pathogenicity/benignity provided by the tools (individually or combined). We performed a clinical validation of major findings using the large-scale BRIDGES case-control dataset. AlphaMissense outperforms ΔΔG and BayesDel, providing similar PP3/BP4 evidence strengths with lower rate of variants in the uninformative score range. AlphaMissense combined with ΔΔG increases evidence strength granularity. AlphaFold2 models perform well as input templates for ΔΔG predictions. Regardless of the tool, BP4 (but not PP3) is highly dependent on RSA, with benignity evidence provided only to variants targeting buried or partially buried residues (RSA ≤ 60%). Stratification by functional domain did not reveal major differences. In brief, structure-based analysis improves PP3/BP4 assessment, uncovering a relevant role for RSA.

AlphaMissense Predictions and ClinVar Annotations: A Deep Learning Approach to Uveal Melanoma

Objective

Uveal melanoma (UM) poses significant diagnostic and prognostic challenges due to its variable genetic landscape. We explore the use of a novel deep learning tool to assess the functional impact of genetic mutations in UM.

Design

A cross-sectional bioinformatics exploratory data analysis of genetic mutations from UM cases.

Subjects

Genetic data from patients diagnosed with UM were analyzed, explicitly focusing on missense mutations sourced from the Catalogue of Somatic Mutations in Cancer (COSMIC) database.

Methods

We identified missense mutations frequently observed in UM using the COSMIC database, assessed their potential pathogenicity using AlphaMissense, and visualized mutations using AlphaFold. Clinical significance was cross-validated with entries in the ClinVar database.

Main Outcome Measures

The primary outcomes measured were the agreement rates between AlphaMissense predictions and ClinVar annotations regarding the pathogenicity of mutations in critical genes associated with UM, such as GNAQ, GNA11, SF3B1, EIF1AX, and BAP1.

Results

Missense substitutions comprised 91.35% (n = 1310) of mutations in UM found on COSMIC. Of the 151 unique missense mutations analyzed in the most frequently mutated genes, only 40.4% (n = 61) had corresponding data in ClinVar. Notably, AlphaMissense provided definitive classifications for 27.2% (n = 41) of the mutations, which were labeled as "unknown significance" in ClinVar, underscoring its potential to offer more clarity in ambiguous cases. When excluding these mutations of uncertain significance, AlphaMissense showed perfect agreement (100%) with ClinVar across all analyzed genes, demonstrating no discrepancies where a mutation predicted as "pathogenic" was classified as "benign" or vice versa.

Conclusions

Integrating deep learning through AlphaMissense offers a promising approach to understanding the mutational landscape of UM. Our methodology holds the potential to improve genomic diagnostics and inform the development of personalized treatment strategies for UM.

Financial Disclosures

Proprietary or commercial disclosure may be found in the Footnotes and Disclosures at the end of this article.

Characterizing trends in clinical genetic testing: A single-center analysis of EHR data from 1.8 million patients over two decades

A lack of structural data in electronic health records (EHRs) makes assessing the impact of genetic testing on clinical practice challenging. We extracted clinical genetic tests from the EHRs of more than 1.8 million patients seen at Vanderbilt University Medical Center from 2002 to 2022. With these data, we quantified the use of clinical genetic testing in healthcare and described how testing patterns and results changed over time. We assessed trends in types of genetic tests, tracked usage across medical specialties, and introduced a new measure, the genetically attributable fraction (GAF), to quantify the proportion of observed phenotypes attributable to a genetic diagnosis over time. We identified 104,392 tests and 19,032 molecularly confirmed diagnoses. The proportion of patients with genetic testing in their EHRs increased from 1.0% in 2002 to 6.1% in 2022, and testing became more comprehensive with the growing use of multi-gene panels. The number of unique diseases diagnosed with genetic testing increased from 51 in 2002 to 509 in 2022, and there was a rise in the number of variants of uncertain significance. The phenome-wide GAF for 6,505,620 diagnoses made in 2022 was 0.46%, and the GAF was greater than 5% for 74 phenotypes, including pancreatic insufficiency (67%), chorea (64%), atrial septal defect (24%), microcephaly (17%), paraganglioma (17%), and ovarian cancer (6.8%). Our study provides a comprehensive quantification of the increasing role of genetic testing at a major academic medical institution and demonstrates its growing utility in explaining the observed medical phenome.

PSTP: accurate residue-level phase separation prediction using protein conformational and language model embeddings

Phase separation (PS) is essential in cellular processes and disease mechanisms, highlighting the need for predictive algorithms to analyze uncharacterized sequences and accelerate experimental validation. Current high-accuracy methods often rely on extensive annotations or handcrafted features, limiting their generalizability to sequences lacking such annotations and making it difficult to identify key protein regions involved in PS. We introduce Phase Separation's Transfer-learning Prediction (PSTP), which combines conformational embeddings with large language model embeddings, enabling state-of-the-art PS predictions from protein sequences alone. PSTP performs well across various prediction scenarios and shows potential for predicting novel-designed artificial proteins. Additionally, PSTP provides residue-level predictions that are highly correlated with experimentally validated PS regions. By analyzing 160 000+ variants, PSTP characterizes the strong link between the incidence of pathogenic variants and residue-level PS propensities in unconserved intrinsically disordered regions, offering insights into underexplored mutation effects. PSTP's sliding-window optimization reduces its memory usage to a few hundred megabytes, facilitating rapid execution on typical CPUs and GPUs. Offered via both a web server and an installable Python package, PSTP provides a versatile tool for decoding protein PS behavior and supporting disease-focused research.

Early childhood height, weight, and BMI development in children with monogenic obesity: a European multicentre, retrospective, observational study

BACKGROUND

Monogenic defects in the leptin-melanocortin pathway are associated with hyperphagia and severe, early-onset obesity. Early childhood growth patterns in height, weight, and BMI, might serve as phenotypic markers for specific genetic disorders; however, reliable data are scarce. This study aimed to evaluate the natural history of height, weight, and BMI in early childhood in a large European group of individuals with monogenic obesity.

METHODS

This multicentre observational study analysed height, weight, and BMI from birth to age 5 years in individuals diagnosed with biallelic (likely) pathogenic LEP, LEPR, POMC, PCSK1, or MC4R variants or monoallelic (likely) pathogenic MC4R variants from six European centres (Berlin and Ulm, Germany; Cambridge, UK; Madrid, Spain; Paris, France; Rotterdam, Netherlands). All patient data up to May 31, 2022 were included in this analysis. All individuals had at least two height or weight measurements between birth and age 5 years. Early childhood growth trajectories were compared with those of control children with obesity without a known genetic cause, following a negative next-generation sequencing panel. Diagnostic performance of BMI as a predictor test for monogenic obesity was also evaluated.

FINDINGS

We included 147 individuals with monogenic obesity. From the age of 6 months onwards, children with biallelic variants (n=88, 55% female vs 45% male) had substantially higher BMIs than those with monoallelic MC4R variants (n=59, 53% female vs 47% male) and control children (n=113, 59% female vs 41% male). Children with biallelic LEP, LEPR, and MC4R variants showed a steep BMI increase during the first year of life, followed by a plateau until age 5 years, whereas those with biallelic POMC variants did not plateau. Accelerated linear growth was only observed in children with biallelic MC4R variants starting from age 1 year. The optimal BMI cut-off for distinguishing individuals with biallelic variants from control individuals was identified at age 2 years, with a test positivity cutoff of 24·0 kg/m2 (sensitivity: 0·96 [95% CI 0·89-1·00], specificity: 0·83 [0·74-0·90], AUC: 0·96 [0·91-0·99], p<0·0001). However, BMI had poor diagnostic performance for monoallelic MC4R variants.

INTERPRETATION

This study identified characteristic early childhood BMI trajectories for different forms of monogenic obesity. From age 6 months onwards, individuals with biallelic variants can be distinguished from those with monoallelic variants and common obesity. A BMI ≥24 kg/m2 at age 2 years had good diagnostic performance for biallelic variants, informing future recommendations for genetic screening for monogenic obesity.

FUNDING

Federal Ministry of Education and Research as part of the German Center for Child and Adolescent Health, German Research Foundation, Spanish Ministry of Health, The Wellcome Trust, Botnar Fondation, Leducq Foundation, National Institute for Health Research (NIHR) Cambridge Biomedical Research Centre, and NIHR Senior Investigator Award.

A Perspective on Artificial Intelligence for Molecular Pathologists

The widespread adoption of next-generation sequencing technology in molecular pathology has enabled us to interrogate the genome as never before. The huge quantities of data generated by sequencing, the enormous complexity of human and microbial genetics, and the need for fast answers demand increasing use of automation as we diagnose disease and guide patient treatment. Much of this automation is based on tools that fall under umbrellas that have come to be known as machine learning and artificial intelligence. This review outlines some of the broad ideas that underpin these complex computational methods. It discusses the roles of pathologists and data scientists in generating new tools and factors to keep in mind when adopting these systems for use in molecular pathology. It pays special attention to regulatory and professional society guidance for validating them in individual institutions and to possible sources of bias. Finally, it briefly discusses ongoing efforts in computer science that may dramatically impact artificial intelligence in the future.

DEAD/DEAH-box RNA helicases shape the risk of neurodevelopmental disorders

The DEAD/DEAH-box family of RNA helicases (RHs) is among the most abundant and conserved in eukaryotes. These proteins catalyze the remodeling of RNAs to regulate their splicing, stability, localization, and translation. Rare genetic variants in DEAD/DEAH-box proteins have recently emerged as being associated with neurodevelopmental disorders (NDDs). Analyses in cellular and animal models have uncovered fundamental roles for these proteins during brain development. We discuss the genetic and functional evidence that implicates DEAD/DEAH-box proteins in brain development and NDDs, with a focus on how structural insights from paralogous genes can be leveraged to advance our understanding of the pathogenic mechanisms at play.

Big data and transformative bioinformatics in genomic diagnostics and beyond

The current era of high-throughput analysis-driven research offers invaluable insights into disease etiologies, accurate diagnostics, pathogenesis, and personalized therapy. In the field of movement disorders, investigators are facing an increasing growth in the volume of produced patient-derived datasets, providing substantial opportunities for precision medicine approaches based on extensive information accessibility and advanced annotation practices. Integrating data from multiple sources, including phenomics, genomics, and multi-omics, is crucial for comprehensively understanding different types of movement disorders. Here, we explore formats and analytics of big data generated for patients with movement disorders, including strategies to meaningfully share the data for optimized patient benefit. We review computational methods that are essential to accelerate the process of evaluating the increasing amounts of specialized data collected. Based on concrete examples, we highlight how bioinformatic approaches facilitate the translation of multidimensional biological information into clinically relevant knowledge. Moreover, we outline the feasibility of computer-aided therapeutic target evaluation, and we discuss the importance of expanding the focus of big data research to understudied phenotypes such as dystonia.

Rare predicted loss-of-function and damaging missense variants in CFHR5 associate with protection from age-related macular degeneration

Age-related macular degeneration (AMD) is a leading cause of blindness among older adults worldwide, but treatment options are limited. Genetics studies have implicated the CFH locus, containing CFH and five CFHR genes, CFHR1-5, in AMD. While CFH has been robustly linked with AMD risk, potential additional roles for the CFHR genes remain unclear, obscured by strong linkage disequilibrium across the locus. Investigating rare coding variants can help to identify causal genes in such regions. We used whole-exome sequencing data from 406,952 UK Biobank participants to examine AMD associations with genes at the CFH locus. For each gene, we used burden testing to examine associations of rare (minor-allele frequency [MAF] < 1%) predicted loss-of-function (pLoF) and predicted damaging missense variants with AMD. We considered "broadly defined AMD" (ICD-10 35.3; ncases = 10,700) and "strictly defined AMD" (dry or wet AMD; ncases = 346). Adjusting for CFH-region variants known to independently associate with AMD, we find that CFHR5 rare variant burden significantly associates with a decreased risk of broadly defined AMD (odds ratio [OR] = 0.75, p = 7 × 10-4), with this association primarily driven by pLoF variants. Furthermore, the association of CFHR5 rare variants with AMD protection is estimated to be stronger for individuals with the CFH rs1061170 AMD risk allele (p.Tyr402His [p.Y402H]; interaction p = 0.04). Corresponding analyses of strict AMD were underpowered. However, we observe that thinning of the photoreceptor layer outer segment strongly predicts strict AMD and find that CFHR5 rare variant burden is significantly associated with increased thickness of this retinal layer (+0.34 SD, p = 4 × 10-4, n = 45,365). These findings suggest CFHR5 inhibition as a potential therapeutic approach for AMD.

Structural biology in variant interpretation: Perspectives and practices from two studies

Structural biology offers a powerful lens through which to assess genetic variants by providing insights into their impact on clinically relevant protein structure and function. Due to the availability of new, user-friendly, web-based tools, structural analyses by wider audiences have become more mainstream. These new tools, including AlphaMissense and AlphaFold, have recently been in the limelight due to their initial success and projected future promise; however, the intricacies and limitations of using these tools still need to be disseminated to the more general audience that is likely to use them in variant analysis. Here, we expound on frameworks applying structural biology to variant interpretation by examining two accompanying articles. To this end, we explore the nuances of choosing the correct protein model, compare and contrast various structural approaches, and highlight both the advantages and limitations of employing structural biology in variant interpretation. Using two articles published in this issue of The American Journal of Human Genetics as a baseline, we focus on case studies in TP53 and BRCA1 to illuminate gene-specific differences in the applications of structural information, which illustrate the complexities inherent in this field. Additionally, we discuss the implications of recent advancements, such as AlphaFold, and provide practical guidance for researchers navigating variant interpretation using structural biology.

Pharmacogenomic Testing in the Clinical Laboratory: Historical Progress and Future Opportunities

Pharmacogenomics is a rapidly evolving field with a strong foundation in basic science dating back to 1960. Pharmacogenomic findings have been translated into clinical care through collaborative efforts of clinical practitioners, pharmacists, clinical laboratories, and research groups. The methods used have transitioned from targeted genotyping of relatively few variants in individual genes to multiplexed multi-gene panels, and sequencingbased methods are likely on the horizon; however, no system exists for classifying and reporting rare variants identified via sequencing-based approaches. Laboratory testing in pharmacogenomics is complex for several genes, including cytochrome P450 2D6 (CYP2D6), HLA-A, and HLA-B , owing to a high degree of polymorphisms, homology with other genes, and copy-number variation. These loci require specialized methods and familiarity with each gene, which may persist during the transition to next-generation sequencing. Increasing implementation across laboratories and clinical facilities has required cooperative efforts to develop standard testing targets, nomenclature, and reporting practices and guidelines for applying the results clinically. Beyond standardization, harmonization between pharmacogenomics and the broader field of genomic medicine may be essential for facilitating further adoption and realizing the full potential of personalized medicine. In this review, we describe the evolution of clinical laboratory testing for pharmacogenomics, including standardization efforts and the anticipated transition from targeted genotyping to sequencing-based pharmacogenomics. We speculate on potential upcoming developments, including pharmacoepigenetics, improved understanding of the impact of non-coding variants, use of large-scale functional genomics to characterize rare variants, and a renewed interest in polygenic risk or combinatorial approaches, which will drive the progression of the field.

Role of Mutual Information Profile Shifts in Assessing the Pathogenicity of Mutations on Protein Functions: The Case of Pyrin Variants Associated With Familial Mediterranean Fever

This study presents a novel method to assess the pathogenicity of pyrin protein mutations by using mutual information (MI) as a measure to quantify the correlation between residue motions or fluctuations and associated changes affecting the phenotype. The concept of MI profile shift is presented to quantify changes in MI upon mutation, revealing insights into residue-residue interactions at critical positions. We apply this method to the pyrin protein variants, which are associated with an autosomal recessively inherited disease called familial Mediterranean fever (FMF) since the available tools do not help predict the pathogenicity of the most penetrant variants. We demonstrate the utility of MI profile shifts in assessing the effects of mutations on protein stability, function, and disease phenotype. The importance of MI shifts, particularly the negative shifts observed in the pyrin example, as indicators of severe functional effects is emphasized. Additionally, the exploration of potential compensatory mechanisms suggested by positive MI shifts, which are otherwise random and inconsequential, is highlighted. The study also discusses challenges in relating MI profile changes to disease severity and advocates for comprehensive analysis considering genetic, environmental, and stochastic factors. Overall, this study provides insights into the molecular mechanisms underlying the pathogenesis of FMF and offers a framework for identifying potential therapeutic targets based on MI profile changes induced by mutations.

Integration of protein stability and AlphaMissense scores improves bioinformatic impact prediction for p53 missense and in-frame amino acid deletion variants

The clinical classification of germline missense variants and single-amino-acid deletions is challenging. The BayesDel and Align-GVGD bioinformatic prediction tools currently used for ClinGen TP53 variant curation expert panel (VCEP) classification do not directly capture changes in protein folding stability, measured using computed destabilization energies (ΔΔG scores). The AlphaMissense tool recently developed by Google DeepMind to predict pathogenicity for all human proteome missense variants is trained in part using AlphaFold2 architecture. Our study investigated whether protein folding stability and/or AlphaMissense scores could improve impact prediction for p53 missense and single-amino-acid deletion variants. ΔΔG scores were calculated for missense variants using FoldX and for single-amino-acid deletions using an AlphaFold2/RosettaRelax protocol. Residue surface exposure was categorized using relative solvent accessibility (RSA) measures. The predictive values of ΔΔG scores, AlphaMissense, BayesDel, and Align-GVGD were examined using Boruta and binary logistic regression based on functionally defined reference sets. The likelihood ratio (LR) toward pathogenicity was estimated and used to refine optimal categories for predicting variant pathogenicity for different RSA values. We showed that current VCEP predictive approaches for missense variants were improved by integrating ΔΔG scores ≥2.5 kcal/mol for partially buried and buried residues, but better performance was achieved using AlphaMissense with ΔΔG and RSA. For deletion variants, ΔΔG scores ≥4.8 Rosetta energy unit (REU) in buried residues outperformed currently used predictive approaches. Future TP53 VCEP specifications for p53 missense impact prediction may consider AlphaMissense, ΔΔG score, and RSA combined for substitution variants and ΔΔG score alone for deletion variants.

Recent developments in cystic fibrosis drug discovery: where are we today?

INTRODUCTION

The advent of variant-specific disease-modifying drugs into clinical practice has provided remarkable benefits for people with cystic fibrosis (PwCF), a multi-organ life-limiting inherited disease. However, further efforts are needed to maximize therapeutic benefits as well as to increase the number of PwCF taking CFTR modulators.

AREA COVERED

The authors discuss some of the key limitations of the currently available CFTR modulator therapies (e.g. adverse reactions) and strategies in development to increase the number of available therapeutics for CF. These include novel methods to accelerate theratyping and identification of novel small molecules and cellular targets representing alternative or complementary therapies for CF.

EXPERT OPINION

While the CF therapy development pipeline continues to grow, there is a critical need to optimize strategies that will accelerate testing and approval of effective therapies for (ultra)rare CFTR variants as traditional assays and trials are not suitable to address such issues. Another major barrier that needs to be solved is the restricted access to currently available modulator therapies, which remains a significant burden for PwCF who are from racial and ethnic minorities and/or living in underprivileged regions.

GJA8-associated developmental eye disorders: a new multicentre study highlights mutational hotspots and genotype-phenotype correlations

Variants in gap junction protein alpha 8 (GJA8), the gene encoding connexin 50 (Cx50), are primarily associated with developmental cataract, although some are associated with severe structural eye anomalies, such as aphakia (absent lens), microphthalmia (small eyes), and sclerocornea. To further define the relationship of GJA8 variants to ocular developmental disorders, we screened four large international cohorts with structural eye anomalies, including anophthalmia, microphthalmia, and coloboma (AMC) or cataracts. We identified 15 new families carrying 14 different heterozygous GJA8 variants (12 missense variants and two 1q21 microdeletions). The missense variants comprised 10 previously reported alterations in cases with eye anomalies [p.(Gly22Ser), p.(Val44Met), p.(Asp67Gly), p.(Arg76Cys), p.(Pro88Leu), p.(Gly94Glu), p.(Gly94Arg), p.(His98Arg), p.(Pro189Ser), and p.(Arg198Trp)] and two not yet linked with disease [p.(Thr39Met) and p.(Tyr66Asp)]. Their associated phenotypes ranged from isolated cataracts to a combination of microphthalmia and cataract with/without sclerocornea. Our study confirms GJA8 variants as an important source of genetic diagnoses for families with structural eye anomalies in addition to cataract and highlights specific mutational hotspots. Furthermore, we confirm an important genotype-phenotype correlation between sclerocornea and the p.(Gly94Arg) variant, and detail intra- and inter-familial phenotypic variability, which is important for clinical assessment and genetic counselling.

Expression and immunological role of FUNDC2 in pan-cancer

FUNDC2 is a novel mitochondrial protein and is highly involved in various cancers. However, expression pattern and possible role and mechanism of FUNDC2 in pan-cancer remain to be investigated. TIMER 2.0 was used to investigate the expression patterns and immune infiltration of FUNDC2. GEPIA was applied to study the relationship between level of FUNDC2 and prognosis of the patients with pan-cancer. STRING was employed to analyze the potential interacting proteins of FUNDC2. The phosphorylation sites were predicted by cBioPortal and PhosphoNet. Furthermore, variations of FUNDC2 in cancers were investigated by cBioPortal. Finally, AlphaFold was used to predict the structure of FUNDC2. The data show that there were significant differences in the expression levels of FUNDC2 between cancer tissues and controls. Specifically, the levels of FUNDC2 in 8 cancers were significantly lower than the respective controls. The survival time of the cancer patients with higher levels of FUNDC2 was longer than that of lower FUNDC2 in most different types of cancers. The pattern of FUNDC2 was significantly related to immune infiltration of B cells of cancer patients. STRING analysis revealed that FUNDC2 can interact with FUNDC1, et al. Fifteen phosphorylation sites were predicted by PhosphoNet and cBioPortal, of which the S167 also overlapped with the mutation sites of FUNDC2. These data collectively show that the mitochondrial protein FUNDC2 may serve as a possible prognostic biomarker across various cancers and the mechanism may include immune infiltration.

Profiling of Protein-Coding Missense Mutations in Mendelian Rare Diseases: Clues from Structural Bioinformatics

The growing availability of protein structural data from experimental methods and accurate predictive models provides the opportunity to investigate the molecular origins of rare diseases (RDs) reviewed in the Orpha.net database. In this study, we analyzed the topology of 5728 missense mutation sites involved in Mendelian RDs (MRDs), forming the basis of our structural bioinformatics investigation. Each mutation site was characterized by side-chain position within the overall 3D protein structure and side-chain orientation. Atom depth quantitation, achieved by using SADIC v2.0, allowed the classification of all the mutation sites listed in our database. Particular attention was given to mutations where smaller amino acids replaced bulky, outward-oriented residues in the outer structural layers. Our findings reveal that structural features that could lead to the formation of void spaces in the outer protein region are very frequent. Notably, we identified 722 cases where MRD-associated mutations could generate new surface pockets with the potential to accommodate pharmaceutical ligands. Molecular dynamics (MD) simulations further supported the prevalence of cryptic pocket formation in a subset of drug-binding protein candidates, underscoring their potential for structure-based drug discovery in RDs.

Expanding the Allelic and Clinical Heterogeneity of Movement Disorders Linked to Defects of Mitochondrial Adenosine Triphosphate Synthase

BACKGROUND

Defects of mitochondrial ATP synthase (ATPase) represent an emerging, yet incompletely understood group of neurodevelopmental diseases with abnormal movements.

OBJECTIVE

The aim of this study was to redefine the phenotypic and mutational spectrum of movement disorders linked to the ATPase subunit-encoding genes ATP5F1A and ATP5F1B.

METHODS

We recruited regionally distant patients who had been genome or exome sequenced. Fibroblast cultures from two patients were established to perform RNA sequencing, immunoblotting, mass spectrometry-based high-throughput quantitative proteomics, and ATPase activity assays. In silico three-dimensional missense variant modeling was performed.

RESULTS

We identified a patient with developmental delay, myoclonic dystonia, and spasticity who carried a heterozygous frameshift c.1404del (p.Glu469Serfs*3) variant in ATP5F1A. The patient's cells exhibited significant reductions in ATP5F1A mRNA, underexpression of the α-subunit of ATPase in association with other aberrantly expressed ATPase components, and compromised ATPase activity. In addition, a novel deleterious heterozygous ATP5F1A missense c.1252G>A (p.Gly418Arg) variant was discovered, shared by three patients from two families with hereditary spastic paraplegia (HSP). This variant mapped to a functionally important intersubunit communication site. A third heterozygous variant, c.1074+1G>T, affected a canonical donor splice site of ATP5F1B and resulted in exon skipping with significantly diminished ATP5F1B mRNA levels, as well as impaired ATPase activity. The associated phenotype consisted of cerebral palsy (CP) with prominent generalized dystonia.

CONCLUSIONS

Our data confirm and expand the role of dominant ATP5F1A and ATP5F1B variants in neurodevelopmental movement disorders. ATP5F1A/ATP5F1B-related ATPase diseases should be considered as a cause of dystonia, HSP, and CP. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.