Accurate proteome-wide missense variant effect prediction with AlphaMissense

Contribution of Rare and Potentially Functionally Relevant Sequence Variants in Schizophrenia Risk-Locus Xq28,distal

Duplications of the Xq28,distal locus have been described in male and female patients with schizophrenia (SCZ) or intellectual disability. The Xq28,distal locus spans eight protein-coding genes (F8, CMC4, MTCP1, BRCC3, VBP1, FUNDC2, CLIC2, and RAB39B) and is flanked by recurrent genomic breakpoints. Thus, the issue of which gene/s at this locus is/are relevant in terms of SCZ pathogenesis remains unclear. The aim of this study was to investigate the contribution of rare and potentially functionally relevant sequence variants within the Xq28,distal locus to SCZ risk using the single-molecule molecular inversion probes (smMIP) method. Targeted sequencing was performed in a cohort of 1935 patients with SCZ and 1905 controls of European ancestry. The consecutive statistical analysis addressed two main areas. On the level of the individual variants, allele counts in the patient and control cohort were systematically compared with a Fisher's exact test: (i) for the entire present study cohort; (ii) for patients and controls separated by sex; and (iii) in combination with data published by the Schizophrenia Exome Meta-Analysis (SCHEMA) consortium. On the gene-wise level, a burden analysis was performed using the X-chromosomal model of the Optimal Unified Sequence Kernel Association Test (SKAT-O), with adjustment for possible sex-specific effects. Targeted sequencing identified a total of 13 rare and potentially functional variants in four patients and 11 controls. However, neither at the level of individual rare and potentially functional variants nor at the level of the eight protein-coding genes at the Xq28,distal locus was a statistically significant enrichment in patients compared to controls observed. Although inconclusive, the present findings represent a step toward improved understanding of the contribution of X-chromosomal risk factors in neuropsychiatric disorder development, which is an underrepresented aspect of genetic studies in this field.

Human genetic variants in SLC39A8 impact uptake and steady-state metal levels within the cell

The human SLC39A8 (hSLC39A8) gene encodes a plasma membrane protein SLC39A8 (ZIP8) that mediates the specific uptake of the metals Cd2+, Mn2+, Zn2+, Fe2+, Co2+, and Se4+ Pathogenic variants within hSLC39A8 are associated with congenital disorder of glycosylation type 2 (CDG type II) or Leigh-like syndrome. However, numerous mutations of uncertain significance are also linked to different conditions or benign traits. Our study characterized 21 hSLC39A8 variants and measured their impact on protein localization and intracellular levels of Cd2+, Zn2+, and Mn2+ We identified four variants that disrupt protein expression, five variants with high retention in the endoplasmic reticulum, and 12 variants with localization to the plasma membrane. From the 12 variants with plasma membrane localization, we identified three with complete loss of detectable ion uptake by the cell and five with differential uptake between metal ions. Further in silico analysis on protein stability identified variants that may affect the stability of homodimer interfaces. This study elucidates the variety of effects of hSLC39A8 variants on ZIP8 and on diseases involving disrupted metal ion homeostasis.

Structure-resolved dynamics of type 2M von Willebrand disease

BACKGROUND

Genetically determined amino acid substitutions in the platelet adhesive A1 domain alter von Willebrand factor's (VWF) platelet agglutination competence, resulting in both gain- (type 2B) and loss-of-function (type 2M) phenotypes of von Willebrand disease. Prior studies of variants in both phenotypes revealed defects in secondary structure that altered stability and folding of the domain. An intriguing observation was that loss of function arose from both misfolding of A1 and, in a few cases, hyperstabilization of the native structure.

OBJECTIVES

To fully understand the 2M phenotype, we thoroughly investigated the structure/function relationships of 15 additional type 2M variants and 2 polymorphisms in the A1 domain.

METHODS

These variants were characterized using circular dichroism, fluorescence, calorimetry, hydrogen-deuterium exchange mass spectrometry, surface plasmon resonance, and platelet adhesion under shear flow.

RESULTS

Six variants were natively folded, with 4 being hyperstabilized. Nine variants disordered A1, causing a loss in α-helical structure and unfolding enthalpy. GPIbα binding affinity and platelet adhesion dynamics were highly correlated to helical structure. Hydrogen-deuterium exchange resolved specific C-terminal secondary structure elements that differentially diminish the GPIbα binding affinity of A1. These localized structural perturbations were highly correlated to GPIbα binding affinity and shear-dependent platelet adhesion.

CONCLUSION

While hyperstabilized dynamics in A1 do impair stable platelet attachment to VWF under flow, variant-induced localized disorder in specific regions of the domain misfolds A1 and abrogates platelet adhesion. These 2 opposing conformational properties represent 2 structural classes of VWF that drive the loss-of-function phenotype that is type 2M von Willebrand disease.

Exome-wide association analysis identifies novel risk loci for alcohol-associated hepatitis

BACKGROUND AND AIMS

Alcohol-associated hepatitis (AH) is a clinically severe, acute disease that afflicts only a fraction of patients with alcohol use disorder. Genomic studies of alcohol-associated cirrhosis (AC) have identified several genes of large effect, but the genetic and environmental factors that lead to AH and AC, and their degree of genetic overlap, remain largely unknown. This study aims to identify genes and genetic variations that contribute to the development of AH.

APPROACH AND RESULTS

Exome-sequencing of patients with AH (N=784) and heavy drinking controls (N=951) identified an exome-wide significant association for AH at patalin-like phospholipase domain containing 3, as previously observed for AC in genome-wide association study, although with a much lower effect size. Single nucleotide polymorphisms (SNPs) of large effect size at inducible T cell costimulatory ligand ( ICOSLG ) (Chr 21) and TOX4/RAB2B (Chr 14) were also exome-wide significant. ICOSLG encodes a co-stimulatory signal for T-cell proliferation and cytokine secretion and induces B-cell proliferation and differentiation. TOX high mobility group box family member 4 ( TOX4 ) was previously implicated in diabetes and immune system function. Other genes previously implicated in AC did not strongly contribute to AH, and the only prominently implicated (but not exome-wide significant) gene overlapping with alcohol use disorder was alcohol dehydrogenase 1B ( ADH1B ). Polygenic signals for AH were observed in both common and rare variant analysis and identified genes with roles associated with inflammation.

CONCLUSIONS

This study has identified 2 new genes of high effect size with a previously unknown contribution to alcohol-associated liver disease and highlights both the overlap in etiology between liver diseases and the unique origins of AH.

Applying artificial intelligence to uncover the genetic landscape of coagulation factors

Artificial intelligence (AI) is rapidly advancing our ability to identify and interpret genetic variants associated with coagulation factor deficiencies. This review introduces AI, with a specific focus on machine learning (ML) methods, and examines its applications in the field of coagulation genetics over the past decade. We observed a significant increase in AI-related publications, with a focus on hemophilia A and B. ML approaches have shown promise in predicting the functional impact of genetic variants and establishing genotype-phenotype correlations, exemplified by tools like "Hema-Class" for factor VIII variants. However, some challenges remain, including the need to expand variant selection beyond missense mutations (which is now the standard of most studies). For the future, the integration of AI in calling, detecting, and interpreting genetic variants can significantly improve our ability to process large-scale genomic data. In this frame, we discuss various AI/ML-based tools for genetic variant detection and interpretation, highlighting their strengths and limitations. As the field evolves, the synergistic application of multiple AI models, coupled with rigorous validation strategies, will be crucial in advancing our understanding of coagulation disorders and for personalizing treatment approaches.

Genomic profiling at a single center cracks the code in inborn errors of immunity

Inborn errors of immunity (IEI) entail a diverse group of disorders resulting from hereditary or de novo mutations in single genes, leading to immune dysregulation. This study explores the clinical utility of next-generation sequencing (NGS) techniques in diagnosing monogenic immune defects. Eight patients attending the immunodeficiency clinic and with unclassified antibody deficiency were included in the analysis. Clinical records, immune characteristics, and family histories were reviewed, and a target gene panel (TGP) sequencing was performed to identify pathogenic variants. TGPs identified seven variants in TNFRSF13B (TACI), CARMIL2, STAT1, STAT3, and ORAI1 genes. These findings provided definitive diagnoses and proper prognostic assessment. Patients exhibited a wide range of clinical manifestations, including recurrent infections, autoimmune cytopenias, and organ-specific complications. The genetic diversity observed highlights the importance of genetic testing in diagnosing IEIs and tailoring treatments. This study underscores the role of TGPs in diagnosing IEIs, revealing significant genetic heterogeneity and phenotypic variability. They offer a precise tool for identifying underlying genetic defects, facilitating personalized medicine approaches, and eventually improving patient outcomes. The findings emphasize the need for comprehensive genetic testing to uncover novel pathogenic variants, enhancing our understanding of immune system dysfunction. NGS is a critical tool for the management of IEI, enabling precise diagnosis and personalized treatment strategies. Despite resource limitations, the progressive affordability is likely to expand its clinical utility, ultimately improving patient care and advancing the field of immunology. In the meantime, accurate phenotypic assessment is essential for resource optimization and case prioritization.

Proteome Size Is Positively Correlated with Lifespan in Mammals but Negatively Correlated with Lifespan in Birds

The central dogma describes the unidirectional flow of genetic information from DNA to proteins, leading to an underappreciation of the potential for the information contained in proteomes (the full set of proteins in an organism) to reflect broader biological processes such as lifespan. Here, this is addressed by examining how the size and composition of 276 proteomes from four vertebrate classes are related to lifespan. After accounting for the relationship between body weight and lifespan, lifespan is negatively correlated with proteome size in birds and, to a weaker extent, in fish, and positively correlated with lifespan in mammals. Proteome composition varies amongst the four vertebrate classes, but there is no evidence that any specific amino acid correlated with lifespan. The findings in relation to the role of dietary amino acid restriction are discussed on lifespan extension and raise questions about evolutionary and structural forces shaping proteome composition across species.

Chinese Family With Knobloch Syndrome Associated With a Novel PAK2 Variant Leading to Reduced Phosphorylation Levels

BACKGROUND

Biallelic variants of COL18A1 cause Knobloch syndrome (KNO), a rare genetic disorder, characterized by oculopathy and structural defects. Recently, several studies have suggested that novel de novo missense variants in PAK2 may be associated with KNO; however, there are few case reports. This study aimed to investigate a patient with KNO who initially presented with seizures and expand the PAK2 genotype and phenotype spectrum.

METHODS

This study included a Chinese family with a proband who primarily presented with epilepsy and developmental delay. Whole-exome sequencing and Sanger sequencing were performed to analyze potential variants. Structural modeling was performed to assess the impact of the variant on the protein structure. In vitro, a mutant plasmid was constructed and transfected into 293T cells to conduct phosphorylation assays, and phosphorylation levels at Ser141 of PAK2 were assessed. The PAK kinase inhibitor FRAX597 was used to confirm the specificity of the western blot results.

RESULTS

A de novo variant of PAK2 gene, NM_002577.4: c.1049G>A (p.Arg350Lys) was found in the patient but not in his parents or sister. This variant was found to be located in the kinase domain and may alter the hydrogen-bond network, potentially affecting kinase activity. In vitro functional experiments demonstrated that the variant may lead to reduced protein levels. Moreover, Western blot analysis showed a significant decrease in the phosphorylation level at Ser141 compared to the wild-type plasmid, indicating that the variant may lead to decreased PAK2 phosphorylation levels.

CONCLUSION

The clinical manifestations in this patient may be associated with a novel PAK2 variant, and the atypical presentation of KNO suggests that PAK2-related KNO may have a broader phenotypic spectrum.

CHIME Syndrome in a Child With Homozygous PIGL p.Leu167Pro Variant

CHIME syndrome is a variable condition characterized by ichthyosiform dermatosis, accompanied by intellectual disability, ocular colobomas, ear anomalies, and heart defects. It is an autosomal recessive condition caused by biallelic pathogenic variants in the PIGL gene. Until now, all reports of individuals affected with CHIME syndrome showed the PIGL c.500T>C p.Leu167Pro DNA variant on one allele of the PIGL gene, in combination with another PIGL DNA variant on the other allele. This has led to the hypothesis that the p.Leu167Pro variant determines to a mild phenotypic effect only and that the core phenotype is determined by the second PIGL DNA variant. We report the first individual with CHIME syndrome, a 6-year-old girl, with homozygous PIGL p.Leu167Pro variants, defusing this hypothesis as she is not mildly affected. As CHIME is a very rare condition, it is expected that a significant proportion of cases will be due to homozygous gene variants, especially of founder DNA variants, and offspring of consanguineous parents. We speculate that the lack of homozygous p.Leu167Pro DNA variants so far has been due to chance and that other homozygous cases will be identified in future reports of affected individuals.

Systems Human Immunology and AI: Immune Setpoint and Immune Health

The immune system, critical for human health and implicated in many diseases, defends against pathogens, monitors physiological stress, and maintains tissue and organismal homeostasis. It exhibits substantial variability both within and across individuals and populations. Recent technological and conceptual progress in systems human immunology has provided predictive insights that link personal immune states to intervention responses and disease susceptibilities. Artificial intelligence (AI), particularly machine learning (ML), has emerged as a powerful tool for analyzing complex immune data sets, revealing hidden patterns across biological scales, and enabling predictive models for individualistic immune responses and potentially personalized interventions. This review highlights recent advances in deciphering human immune variation and predicting outcomes, particularly through the concepts of immune setpoint, immune health, and use of the immune system as a window for measuring health. We also provide a brief history of AI; review ML modeling approaches, including their applications in systems human immunology; and explore the potential of AI to develop predictive models and personal immune state embeddings to detect early signs of disease, forecast responses to interventions, and guide personalized health strategies.

Classification and prediction of variants associated with hearing loss using sequence information in the vicinity of mutation sites

Hearing impairment is a major global health problem, affecting more than 5 % of the world's population at various ages, from neonates to the elderly. Among the common genetic variations in humans, single nucleotide variations and small insertions or deletions predominate. The study of hearing loss resulting from these variations is proving invaluable in the analysis and diagnosis of hearing disorders. The identification of pathogenic mutations is frequently a lengthy and laborious process. Existing computational prediction tools have been developed primarily for common diseases and genome-wide analyses, with less focus on deafness. This study proposes a novel approach that focuses on the regions surrounding mutation sites. Mutation sites associated with deafness and their flanking regions of different lengths were extracted from relevant databases and combined into seven distinct segments of different lengths. The information-theoretic features of these segments were computed. Five machine learning algorithms were then used for training, resulting in the construction of a model capable of classifying and predicting deafness-related mutations. For fragments encompassing the 250 bp regions upstream and downstream of the mutations, the average AUC of the five classifiers on the independent test set is 0.89 and the average ACC is 0.85, indicating that the model has a high recognition rate of the pathogenic deafness mutation site. An ensemble approach was also applied to predict variants of uncertain significance (VUS) that may be associated with deafness. These variants were then scored and ranked to assess their likelihood of contributing to the condition.

Chronic Catatonia in an Individual With a De Novo Missense SHANK1 Variant

SHANK1 encodes a scaffolding protein of the SHANK family that includes SHANK1, SHANK2 and SHANK3. All of the SHANK proteins are enriched at the post-synaptic density of excitatory synapses. Here, we present an 11-year-old boy with a history of developmental delays and no family history of psychiatric disorders who developed catatonia. MRI of his brain and spine were negative as was a workup for autoimmune encephalitis. The proband's genetic testing revealed a de novo heterozygous SHANK1 missense variant. Although catatonia has been reported previously in individuals with SHANK3 loss-of-function mutations, this is the first time catatonia has been described in an individual with a SHANK1 variant.

Clinical and genetic spectrum of patients with IRF2BPL syndrome

Interferon regulatory factor 2 binding protein-like (IRF2BPL) is a single-exon gene that is ubiquitously expressed in various tissues, including the brain. IRF2BPL encodes a transcription factor with two zinc-finger domains that potentially downregulate WNT signaling in the nervous system. Pathogenic IRF2BPL variants have been reported to cause developmental delay, seizures, myoclonus epilepsies, autistic spectrum disorder, and other neurodevelopmental disorders. Exome sequencing of 10 patients with developmental delay and/or epilepsy from nine families revealed nine pathogenic IRF2BPL variants, of which eight were novel: five missense, one in-frame indel, and three truncating variants. Using reported pathogenic and benign variants, we highlight here several regions of IRF2BPL that deviate in the frequency of pathogenic and benign variants. This study of detailed clinical and genetic information shows that IRF2BPL missense and in-frame indel variants are often associated with seizures and developmental delay.

Protein-Variant-Phenotype Study of NBAS Using AlphaFold in the Aspect of SOPH Syndrome

NBAS gene variants cause phenotypically distinct and nonoverlapping conditions, SOPH syndrome and ILFS2. NBAS is a so-called "moonlighting" protein responsible for retrograde membrane trafficking and nonsense-mediated decay. However, its three-dimensional model and the nature of its possible interactions with other proteins have remained elusive. Here, we used AlphaFold to predict protein-protein interaction (PPI) sites and mapped them to NBAS pathogenic variants. We repeated in silico milestone studies of the NBAS protein to explain the multisystem phenotype of its variants, with particular emphasis on the SOPH variant (p.R1914H). We revealed the putative binding sites for the main interaction partners of NBAS and assessed the implications of these binding sites for the subdomain architecture of the NBAS protein. Using AlphaFold, we disclosed the far-reaching impact of NBAS variants on the development of each phenotypic trait in patients with NBAS-related pathologies.

Search for germline gene variants in colorectal cancer families presenting with multiple primary colorectal cancers

A double primary colorectal cancer (CRC) in a familial setting signals a high risk of CRC. In order to identify novel CRC susceptibility genes, we whole-exome sequenced germline DNA from nine persons with a double primary CRC and a family history of CRC. The detected variants were processed by bioinformatics filtering and prioritization, including STRING protein-protein interaction and pathway analysis. A total of 150 missense, 19 stop-gain, 22 frameshift and 13 canonical splice site variants fulfilled our filtering criteria. The STRING analysis identified 20 DNA repair/cell cycle proteins as the main cluster, related to genes CHEK2, EXO1, FAAP24, FANCI, MCPH1, POLL, PRC1, RECQL, RECQL5, RRM2, SHCBP1, SMC2, XRCC1, in addition to CDK18, ENDOV, ZW10 and the known mismatch repair genes. Another STRING network included extracellular matrix genes and TGFβ signaling genes. In the nine whole-exome sequenced patients, eight harbored at least two candidate DNA repair/cell cycle/TGFβ signaling gene variants. The number of families is too small to provide evidence for individual variants but, considering the known role of DNA repair/cell cycle genes in CRC, the clustering of multiple deleterious variants in the present families suggests that these, perhaps jointly, contributed to CRC development in these families.

Single-point mutations in disordered proteins: Linking sequence, ensemble, and function

Mutations in genomic DNA often result in single-point missense mutations in proteins. For folded proteins, the functional effect of these missense mutations can often be understood by their impact on structure. However, missense mutations in intrinsically disordered protein regions (IDRs) remain poorly understood. In IDRs, function can depend on the structural ensemble- the collection of accessible, interchanging conformations that is encoded in their amino acid sequence. We argue that, analogously to folded proteins, single-point mutations in IDRs can alter their structural ensemble, and consequently alter their biological function. To make this argument, we first provide experimental evidence from the literature showcasing how single-point missense mutations in IDRs affect their ensemble dimensions. Then, we use genomic data from patients to show that disease-linked missense mutations occurring in IDRs can, in many cases, significantly alter IDR structural ensembles. We hope this analysis prompts further study of disease-linked, single-point mutations in IDRs.

Efficacy of EGFR tyrosine kinase inhibitors in patients with non-small cell lung cancer with EGFR exon 19 insertions: clinical-genomic, preclinical analysis through LC-SCRUM-Asia (multi-institutional genomic screening registry)

BACKGROUND

EGFR exon 19 insertions (EGFRex19ins) are rare EGFR mutations. Their clinical-genomic characteristics and outcomes with EGFR-tyrosine kinase inhibitors (TKIs) remain uncertain.

METHODS

We evaluated the clinical-genomic characteristics and outcomes of EGFR-TKIs for EGFRex19ins in the multi-institutional prospective lung cancer genomic screening project (LC-SCRUM-Asia). We also studied preclinical Ba/F3 models expressing EGFR-K745_E746insIPVAIK (Ba/F3-IPVAIK) to investigate their sensitivity to 1st-, 2nd-, 3rd-generation, and EGFR exon 20 insertion-active TKIs.

RESULTS

In LC-SCRUM-Asia, 16,204 NSCLC patients were enrolled from March 2015 to December 2023. EGFRex19ins were detected in 13 samples (0.1 % of NSCLC). The median age was 72 years (range, 38-80); most patients were female (77 %), had adenocarcinoma (92 %), and were never-smokers (62 %). Twelve patients (93 %) had EGFR-K745_E746insIPVAIK, while one (7 %) had EGFR-K745_E746insVPVAIK. The most frequent co-mutation was TP53 (62 %); no patients had other driver alterations. Six patients (46 %) tested positive for EGFR exon 19 deletions with PCR-based Cobas EGFR test, likely due to cross-reactivity arising from sequence homology. Twelve patients received EGFR-TKIs; five (42 %) experienced partial response. In the preclinical study, Ba/F3-IPVAIK showed the highest sensitivity to 2nd-generation EGFR-TKIs compared to other EGFR-TKIs. Structural studies supported these consistent results. When broken down by EGFR-TKI generations, response rates for 1st-, 2nd-, and 3rd-generation TKIs were 50 % (1/2), 80 % (4/5), and 0 % (0/5), respectively. The median PFS for 1st-, 2nd-, and 3rd-generation TKIs were 8.7 (95 % CI, 7.4-NR), 14.7 (95 % CI, 8.0-NR), and 4.4 (95 % CI, 3.4-NR) months, respectively.

CONCLUSION

Our preclinical, structural, and clinical findings indicate 2nd-generation EGFR-TKIs are more effective for EGFRex19ins compared to other TKIs.

Reassessing the role of the p.(Arg304Gln) missense AIP variant in pituitary tumorigenesis

OBJECTIVE

Heterozygous germline loss-of-function variants in AIP are associated with young-onset growth hormone and/or prolactin-secreting pituitary tumours. However, the pathogenic role of the c.911G > A; p.(Arg304Gln) (R304Q) AIP variant has been controversial. Recent data from public exome/genome databases show this variant is not infrequent. The objective of this work was to reassess the pathogenicity of R304Q based on clinical, genomic, and functional assay data.

DESIGN

Data were collected on published R304Q pituitary neuroendocrine tumour cases and from International Familial Isolated Pituitary Adenoma Consortium R304Q cases (n = 38, R304Q cohort). Clinical features, population cohort frequency, computational analyses, prediction models, presence of loss-of-heterozygosity, and in vitro/in vivo functional studies were assessed and compared with data from pathogenic/likely pathogenic AIP variant patients (AIPmut cohort, n = 184).

RESULTS

Of 38 R304Q patients, 61% (23/38) had growth hormone excess, in contrast to 80% of AIPmut cohort (147/184, P < .001). R304Q cohort was older at disease onset and diagnosis than the AIPmut cohort (median [quartiles] onset: 25 y [16-35] vs 16 y [14-23], P < .001; median [quartiles] diagnosis: 36 y [24-44] vs 21 y [15-29], P < .001). R304Q is present in gnomADv2.1 (0.31%) and UK Biobank (0.16%), including three persons with homozygous R304Q. No loss-of-heterozygosity was detected in four R304Q pituitary neuroendocrine tumour samples. In silico predictions and experimental data were conflicting.

CONCLUSIONS

Evidence suggests that R304Q is not pathogenic for pituitary neuroendocrine tumour. We recommend changing this variant classification to likely benign and do not recommend pre-symptomatic genetic testing of family members or follow-up of already identified unaffected individuals with the R304Q variant.

Alternative start codon selection shapes mitochondrial function during evolution, homeostasis, and disease

Mitochondrial endosymbiosis was a pivotal event in eukaryotic evolution, requiring core proteins to adapt to function both within the mitochondria and in the host cell. Here, we systematically profile the localization of protein isoforms generated by alternate start codon selection during translation. We identify hundreds of pairs of differentially-localized protein isoforms, many of which affect mitochondrial targeting and are essential for mitochondrial function. The emergence of dual-localized mitochondrial protein isoforms coincides with mitochondrial acquisition during early eukaryotic evolution. We further reveal that eukaryotes use diverse mechanisms-such as leaky ribosome scanning, alternative transcription, and paralog duplication-to maintain the production of dual-localized isoforms. Finally, we identify multiple isoforms that are specifically dysregulated by rare disease patient mutations and demonstrate how these mutations can help explain unique clinical presentations. Together, our findings illuminate the evolutionary and pathological relevance of alternative translation initiation, offering new insights into the molecular underpinnings of mitochondrial biology.

Comprehensive structural and functional analyses of RAD50 nsSNPs: from prediction to impact assessment

Background

The RAD50 gene on chromosome 5q3.11 plays an important role in the MRN (Mre11-Rad50-Nbs1) complex. This complex orchestrates cellular responses to the DNA double-strand breaks (DSBs) through several pathways for genome stability. This study aims to investigate the functional impact of non-synonymous single-nucleotide polymorphisms (nsSNPs) in RAD50 (a breast cancer-associated gene) and focuses on their consequences on protein structure and interaction within the MRN complex.

Methods

A total of 1,806 nsSNPs were retrieved and subjected to variant analysis using a set of computational tools and ConSurf. Pathogenicity and protein stability criteria were established based on specific tools. Highly conserved damaging nsSNPs were prioritized for the structural analysis. GOR-IV was used for secondary structure prediction, whereas AlphaFold, RoseTTAFold, and I-TASSER were used for protein structure prediction. The docking of RAD50-Mre11A complexes was performed using HADDOCK to assess the impact of nsSNPs on protein-protein interactions. Molecular dynamic simulation was performed to verify the role of mutants in molecular docking analysis.

Results

A subset of pathogenic and disease-associated nsSNPs in the RAD50 gene altered the protein stability and interactions with the Mre11A protein. Substantial alterations in the interacting profiles of mutants (A73P, V117F, L518P, L1092R, N1144S, and A1209T) suggest potential implications for DNA repair mechanisms and genome stability.

Conclusion

The study discloses the normative impact of RAD50 mutations on the pathophysiology of breast cancer. It can provide the basis to treat RAD50 mutation-deficient cells.

VUStruct: a compute pipeline for high throughput and personalized structural biology

Effective diagnosis and treatment of rare genetic disorders requires the interpretation of a patient's genetic variants of unknown significance (VUSs). Today, clinical decision-making is primarily guided by gene-phenotype association databases and DNA-based scoring methods. Our web-accessible variant analysis pipeline, VUStruct, supplements these established approaches by deeply analyzing the downstream molecular impact of variation in context of 3D protein structure. VUStruct's growing impact is fueled by the co-proliferation of protein 3D structural models, gene sequencing, compute power, and artificial intelligence. Contextualizing VUSs in protein 3D structural models also illuminates longitudinal genomics studies and biochemical bench research focused on VUS, and we created VUStruct for clinicians and researchers alike. We now introduce VUStruct to the broad scientific community as a mature, web-facing, extensible, High-Performance Computing (HPC) software pipeline. VUStruct maps missense variants onto automatically selected protein structures and launches a broad range of analyses. These include energy-based assessments of protein folding and stability, pathogenicity prediction through spatial clustering analysis, and machine learning (ML) predictors of binding surface disruptions and nearby post-translational modification sites. The pipeline also considers the entire input set of VUS and identifies genes potentially involved in digenic disease. VUStruct's utility in clinical rare disease genome interpretation has been demonstrated through its analysis of over 175 Undiagnosed Disease Network (UDN) Patient cases. VUStruct-leveraged hypotheses have often informed clinicians in their consideration of additional patient testing, and we report here details from two cases where VUStruct was key to their solution. We also note successes with academic research collaborators, for whom VUStruct has informed research directions in both computational genomics and wet lab studies.

Oxidation of retromer complex controls mitochondrial translation

Reactive oxygen species (ROS) underlie human pathologies including cancer and neurodegeneration1,2. However, the proteins that sense ROS levels and regulate their production through their cysteine residues remain ill defined. Here, using systematic base-editing and computational screens, we identify cysteines in VPS35, a member of the retromer trafficking complex3, that phenocopy inhibition of mitochondrial translation when mutated. We find that VPS35 underlies a reactive metabolite-sensing pathway that lowers mitochondrial translation to decrease ROS levels. Intracellular hydrogen peroxide oxidizes cysteine residues in VPS35, resulting in retromer dissociation from endosomal membranes and subsequent plasma membrane remodelling. We demonstrate that plasma membrane localization of the retromer substrate SLC7A1 is required to sustain mitochondrial translation. Furthermore, decreasing VPS35 levels or oxidation of its ROS-sensing cysteines confers resistance to ROS-generating chemotherapies, including cisplatin, in ovarian cancer models. Thus, we identify that intracellular ROS levels are communicated to the plasma membrane through VPS35 to regulate mitochondrial translation, connecting cytosolic ROS sensing to mitochondrial ROS production.

Single-cell multiome and enhancer connectome of human retinal pigment epithelium and choroid nominate pathogenic variants in age-related macular degeneration

Age-related macular degeneration (AMD) is a leading cause of vision loss worldwide. Genome-wide association studies (GWAS) of AMD have identified dozens of risk loci that may house disease targets. However, variants at these loci are largely noncoding, making it difficult to assess their function and whether they are causal. Here, we present a single-cell gene expression and chromatin accessibility atlas of human retinal pigment epithelium (RPE) and choroid to systematically analyze both coding and noncoding variants implicated in AMD. We employ HiChIP and Activity-by-Contact modeling to map enhancers in these tissues and predict cell and gene targets of risk variants. We further perform allele-specific self-transcribing active regulatory region sequencing (STARR-seq) to functionally test variant activity in RPE cells, including in the context of complement activation. Our work nominates new pathogenic variants and mechanisms in AMD and offers a rich and accessible resource for studying diseases of the RPE and choroid.

FDPSM: Feature-Driven Prediction Modeling of Pathogenic Synonymous Mutations

Synonymous mutations, once considered to be biologically neutral, are now recognized to affect protein expression and function by altering the RNA splicing, stability, or translation efficiency. These effects can contribute to disease, making the prediction of the pathogenicity a crucial task. Computational methods have been developed to analyze the sequence features and biological functions of synonymous mutations, but existing methods face limitations, including scarcity of labeled data, reliance on other prediction tools, and insufficient representation of feature interrelationships. Here, we present FDPSM, a novel prediction method specifically designed to predict pathogenic synonymous mutations. FDPSM was trained on a robust data set of 4251 positive and negative training samples to enhance predictive accuracy. The method leveraged a comprehensive set of features, including genomic context, conservation, splicing effects, functional effects, and epigenomics, without relying on prediction scores from other mutation pathogenicity tools. Recognizing that original features alone may not fully capture the distinctions between pathogenic and benign synonymous mutations, we enhanced the feature set by extracting effective information from the interactions and distribution of these features. The experimental results showed that FDPSM significantly outperformed existing methods in predicting the pathogenicity of synonymous mutations, offering a more accurate and reliable tool for this important task. FDPSM is available at https://github.com/xialab-ahu/FDPSM.

Insights for variant clinical interpretation based on a benchmark of 65 variant effect predictors

Single amino acid substitutions in protein sequences are generally harmless, but a certain number of these changes can lead to disease. Accurately predicting the effect of genetic variants is crucial for clinicians as it accelerates the diagnosis of patients with missense variants associated with health problems. Many computational tools have been developed to predict the pathogenicity of genetic variants with various approaches. Analysing the performance of these different computational tools is crucial to provide guidance to both future users and especially clinicians. In this study, a large-scale investigation of 65 tools was conducted. Variants from both clinical and functional contexts were used, incorporating data from the ClinVar database and bibliographic sources. The analysis showed that AlphaMissense often performed very well and was in fact one of the best options among the existing tools. In addition, as expected, meta-predictors perform well on average. Tools using evolutionary information showed the best performance for functional variants. These results also highlighted some heterogeneity in the difficulty of predicting some specific variants while others are always well categorized. Strikingly, the majority of variants from the ClinVar database appear to be easy to predict, while variants from other sources of data are more challenging. This raises questions about the use of ClinVar and the dataset used to validate tools accuracy. In addition, these results show that this variant predictability can be divided into three distinct classes: easy, moderate and hard to predict. We analyzed the parameters leading to these differences and showed that the classes are related to structural and functional information.

TITINdb2-expanding annotation and structural information for protein variants in the giant sarcomeric protein titin

Summary

We present TITINdb2, an update to the TITINdb database previously constructed to facilitate the identification of pathogenic missense variants in the giant protein titin, which are associated with a variety of skeletal and cardiac myopathies. The database and web portal have been substantially revised and include the following new features: (i) an increase in computational annotation from 4 to 20 variant impact predictors, available through a new custom data table dialogue; (ii) through structural coverage of single domains with AlphaFold2 predicted models; (iii) newly predicted domain-domain interface annotations; (iv) an expanded in silico saturation mutagenesis incorporating four variant impact predictors; (v) a comprehensive overhaul of available data, including population data sources and variants reported pathogenic in the literature; and (vi) a curated mapping of existing protein, transcript, and chromosomal sequence positions and a new variant conversion tool to translate variants in one format to any other format.

Availability and implementation

The database is accessible via titindb.kcl.ac.uk/TITINdb/.

Structural Evaluation of RYR2-CPVT Missense Variants and Continuous Bayesian Estimates of their Penetrance

Background

Catecholaminergic Polymorphic Ventricular Tachycardia (CPVT) is strongly associated with rare missense variants in RYR2, the gene encoding the intracellular calcium release channel RyR2. Precision medicine is complicated by incomplete penetrance, particularly in the case of RYR2-CPVT variants.

Objective

To improve structural understanding and clinical actionability of RYR2-CPVT incomplete penetrance.

Methods

We curated 179 manuscripts reviewed by three individuals to extrapolate RYR2-CPVT missense variant genotype-phenotype relationships. Purportedly neutral control variants were ascertained from RYR2 missense variants observed in gnomAD and ClinVar. We performed an RYR2-CPVT Bayesian penetrance analysis by conditioning a CPVT penetrance prior on variant-specific features (in silico and structural) calibrated by heterozygote phenotypes. We compared the calibration of our Bayesian penetrance estimates and our previous described structural density metric with in silico predictors REVEL, AlphaMissense and ClinVar annotations, using Spearman rank-order correlations, and Brier Scores. Penetrance estimates were superimposed upon a cryo-EM structure of RyR2 to investigate 'hot-spot' heterogeneity.

Results

From the literature and gnomAD, we identified 1,014 affected missense RYR2 heterozygotes (468 unique variants) among a total of 622,575 heterozygotes (5,181 unique variants). Among the predictors, our Bayesian prior score had the highest Spearman rank-order and lowest Brier scores, respectively (0.19; 0.0090), compared to ClinVar (0.083; 0.019), REVEL (0.16; 0.018), or AlphaMissense (0.18; 0.018). Penetrance estimates for all RYR2 missense variants are prospectively hosted at our Variant Browser website.

Conclusions

Bayesian penetrance scores outperform current tools in evaluating variant penetrance. We provide prospective CPVT penetrance values for 29,242 RYR2 missense variants at our online Variant Browser.

Predictomes, a classifier-curated database of AlphaFold-modeled protein-protein interactions

Protein-protein interactions (PPIs) are ubiquitous in biology, yet a comprehensive structural characterization of the PPIs underlying cellular processes is lacking. AlphaFold-Multimer (AF-M) has the potential to fill this knowledge gap, but standard AF-M confidence metrics do not reliably separate relevant PPIs from an abundance of false positive predictions. To address this limitation, we used machine learning on curated datasets to train a structure prediction and omics-informed classifier (SPOC) that effectively separates true and false AF-M predictions of PPIs, including in proteome-wide screens. We applied SPOC to an all-by-all matrix of nearly 300 human genome maintenance proteins, generating ∼40,000 predictions that can be viewed at predictomes.org, where users can also score their own predictions with SPOC. High-confidence PPIs discovered using our approach enable hypothesis generation in genome maintenance. Our results provide a framework for interpreting large-scale AF-M screens and help lay the foundation for a proteome-wide structural interactome.

Identification of IDH3G, encoding the gamma subunit of mitochondrial isocitrate dehydrogenase, as a novel candidate gene for X-linked retinitis pigmentosa

PURPOSE

Retinitis pigmentosa (RP) is a genetically heterogeneous group of retinal degenerative disorders characterized by the loss of rod and cone photoreceptors, leading to visual impairment and blindness. To date, to our knowledge, X-linked RP has been associated with variants in 3 genes (RPGR, RP2, and OFD1), whereas genetic defects at 3 loci (RP6, RP24, and RP34) are yet unidentified. The aim of this study was to identify a novel candidate gene underlying X-linked RP.

METHODS

Participants were identified from cohorts of genetically unsolved male individuals affected by RP, who underwent genome sequencing, exome sequencing, or candidate gene screening via direct Sanger sequencing at 3 referral centers. Specifically, 2 probands were identified at the National Reference Centre for Rare Retinal Diseases (Paris, France), 2 at the Massachusetts Eye and Ear Hospital (Boston, MA), and 1 at the National Reference Centre for Inherited Sensory Diseases (Montpellier, France). The pathogenicity of the identified variants was assessed using bioinformatic predictions, protein expression analyses, and mitochondrial function assays.

RESULTS

We identified 4 rare single-nucleotide variants in IDH3G (HGNC:5386), located at the RP34 locus on the X chromosome, and a complete gene deletion, in 5 unrelated male individuals affected with nonsyndromic RP. The variants segregated with the phenotype in all available family members. In all cases, the disease severity was intermediate. None had high myopia. IDH3G encodes the γ subunit of mitochondrial isocitrate dehydrogenase (IDH3), an enzyme involved in the citric acid cycle, which is expressed in the inner segments of photoreceptors. Variants in IDH3A and IDH3B, encoding the other subunits of IDH3, have already been associated with nonsyndromic autosomal recessive RP. Bioinformatic predictions and functional assays support a pathogenic role for the variants identified in this study, possibly through partial loss of enzymatic activity and mitochondrial function.

CONCLUSION

Our findings suggest that variants in IDH3G are a novel cause of X-linked RP.

Low Penetrance Sarcomere Variants Contribute to Additive Risk in Hypertrophic Cardiomyopathy

BACKGROUND

Classically, hypertrophic cardiomyopathy (HCM) has been viewed as a single-gene (monogenic) disease caused by pathogenic variants in sarcomere genes. Pathogenic sarcomere variants are individually rare and convey high risk for developing HCM (highly penetrant). Recently, important polygenic contributions have also been characterized. Low penetrance sarcomere variants (LowSVs) at intermediate frequencies and effect sizes have not been systematically investigated. We hypothesize that LowSVs may be common in HCM with substantial influence on disease risk and severity.

METHODS

Among all sarcomere variants observed in the Sarcomeric Human Cardiomyopathy Registry (SHaRe), we identified putative LowSVs defined by (1) population frequency greater than expected for highly penetrant (monogenic) HCM (allele frequency >5×10-5 in the Genome Aggregation Database, gnomAD) and (2) moderate enrichment (>2×) in patients with HCM compared with gnomAD. LowSVs were examined for their association with disease severity and clinical outcomes. Functional effects of selected LowSVs were assessed using induced pluripotent stem cell-derived cardiomyocytes. Association of LowSVs with HCM-adjacent traits in the general population was tested using UK Biobank cardiac magnetic resonance imaging data.

RESULTS

Among 6045 patients and 1159 unique variants in sarcomere genes, 12 LowSVs were identified. LowSVs were collectively common in the general population (1:350) and moderately enriched in HCM (aggregate odds ratio, 14.9 [95% CI, 12.5-17.9]). Isolated LowSVs were associated with an older age of HCM diagnosis and fewer adverse events. However, LowSVs in combination with a pathogenic sarcomere variant conferred higher morbidity (eg, composite adverse event hazard ratio, 5.4 [95% CI, 3.0-9.8] versus single pathogenic sarcomere variant, 2.0 [95% CI, 1.8-2.2]; P<0.001). An intermediate functional impact was validated for 2 specific LowSVs-MYBPC3 c.442G>A (partial splice gain) and TNNT2 c.832C>T (intermediate effect on contractile mechanics). Cardiac magnetic resonance imaging analysis of the general population revealed 5 of 12 LowSVs were significantly associated with HCM-adjacent traits without overt HCM.

CONCLUSIONS

This study establishes a new class of low penetrance sarcomere variants that are relatively common in the population. When penetrant, isolated LowSVs cause mild HCM. In combination with pathogenic sarcomere variants, LowSVs markedly increase disease severity, supporting a clinically significant additive effect. Last, LowSVs also contribute to age-related remodeling even in the absence of overt HCM.

Learning the language of protein-protein interactions

Protein Language Models (PLMs) trained on large databases of protein sequences have proven effective in modeling protein biology across a wide range of applications. However, while PLMs excel at capturing individual protein properties, they face challenges in natively representing protein-protein interactions (PPIs), which are crucial to understanding cellular processes and disease mechanisms. Here, we introduce MINT, a PLM specifically designed to model sets of interacting proteins in a contextual and scalable manner. Using unsupervised training on a large curated PPI dataset derived from the STRING database, MINT outperforms existing PLMs in diverse tasks relating to protein-protein interactions, including binding affinity prediction and estimation of mutational effects. Beyond these core capabilities, it excels at modeling interactions in complex protein assemblies and surpasses specialized models in antibody-antigen modeling and T cell receptor-epitope binding prediction. MINT's predictions of mutational impacts on oncogenic PPIs align with experimental studies, and it provides reliable estimates for the potential for cross-neutralization of antibodies against SARS-CoV-2 variants of concern. These findings position MINT as a powerful tool for elucidating complex protein interactions, with significant implications for biomedical research and therapeutic discovery.

A novel seven-tier framework for the classification of MEFV missense variants using adaptive and rigid classifiers

There is a great discrepancy between the clinical categorization of MEFV gene variants and in silico tool predictions. In this study, we developed a seven-tier classification system for MEFV missense variants of unknown significance and recommended a generalized pipeline for other gene classifications. We extracted 12,017 human MEFV gene variants from the Ensembl database. After extraction, we detected 6034 missense variants. In the next step, we selected 42 in silico tools for our classification model. We determined the optimal value via the scores from three in silico tools. For the implementation of machine learning methods, we used two bagging methods and two boosting methods. After predicting known variants, we applied our model to 5507 variants of unknown significance. In the final stage, we applied the developed framework to the entire dataset to rigorously evaluate its classification performance and validate its potential clinical utility. The XGBoost model achieved the highest accuracy at 0.9882 (± 0.0295), followed by Extremely Randomized Trees (0.9835 ± 0.0335), Random Forest (0.9788 ± 0.0158), and AdaBoost (0.9671 ± 0.0815). Following the refinement of the dataset and the introduction of a novel classification and clustering methodology, the proportion of known variants increased from 6.9 to 29.4%, marking a 4.3-fold relative improvement. Furthermore, we identified two novel hotspot regions and one tolerant site, offering valuable insights into the functional structure of the pyrin protein. Rigid and adaptive classifiers offer an innovative framework for VOUS classification, integrating a grayscale interpretation system with cutting-edge in silico tools and machine learning algorithms. This approach not only improves the accuracy of MEFV gene variant classification but also identifies new hotspot regions for functional studies, paving the way for scalable applications to other genes and might contribute to advancing precision genomic medicine in the future.

Deciphering lung adenocarcinoma evolution and the role of LINE-1 retrotransposition

Understanding lung cancer evolution can identify tools for intercepting its growth. In a landscape analysis of 1024 lung adenocarcinomas (LUAD) with deep whole-genome sequencing integrated with multiomic data, we identified 542 LUAD that displayed diverse clonal architecture. In this group, we observed an interplay between mobile elements, endogenous and exogenous mutational processes, distinct driver genes, and epidemiological features. Our results revealed divergent evolutionary trajectories based on tobacco smoking exposure, ancestry, and sex. LUAD from smokers showed an abundance of tobacco-related C:G>A:T driver mutations in KRAS plus short subclonal diversification. LUAD in never smokers showed early occurrence of copy number alterations and EGFR mutations associated with SBS5 and SBS40a mutational signatures. Tumors harboring EGFR mutations exhibited long latency, particularly in females of European-ancestry (EU_N). In EU_N, EGFR mutations preceded the occurrence of other driver genes, including TP53 and RBM10. Tumors from Asian never smokers showed a short clonal evolution and presented with heterogeneous repetitive patterns for the inferred mutational order. Importantly, we found that the mutational signature ID2 is a marker of a previously unrecognized mechanism for LUAD evolution. Tumors with ID2 showed short latency and high L1 retrotransposon activity linked to L1 promoter demethylation. These tumors exhibited an aggressive phenotype, characterized by increased genomic instability, elevated hypoxia scores, low burden of neoantigens, propensity to develop metastasis, and poor overall survival. Reactivated L1 retrotransposition-induced mutagenesis can contribute to the origin of the mutational signature ID2, including through the regulation of the transcriptional factor ZNF695, a member of the KZFP family. The complex nature of LUAD evolution creates both challenges and opportunities for screening and treatment plans.

Deep learning prioritizes cancer mutations that alter protein nucleocytoplasmic shuttling to drive tumorigenesis

Genetic variants can affect protein function by driving aberrant subcellular localization. However, comprehensive analysis of how mutations promote tumor progression by influencing nuclear localization is currently lacking. Here, we systematically characterize potential shuttling-attacking mutations (SAMs) across cancers through developing the deep learning model pSAM for the ab initio decoding of the sequence determinants of nucleocytoplasmic shuttling. Leveraging cancer mutations across 11 cancer types, we find that SAMs enrich functional genetic variations and critical genes in cancer. We experimentally validate a dozen SAMs, among which R14M in PTEN, P255L in CHFR, etc. are identified to disrupt the nuclear localization signals through interfering their interactions with importins. Further studies confirm that the nucleocytoplasmic shuttling altered by SAMs in PTEN and CHFR rewire the downstream signaling and eliminate their function of tumor suppression. Thus, this study will help to understand the molecular traits of nucleocytoplasmic shuttling and their dysfunctions mediated by genetic variants.

Functionally constrained human proteins are less prone to mutational instability from single amino acid substitutions

Missense mutations that disrupt protein structural stability are a common pathogenic mechanism in human genetic disease. Here, we quantify potential disruption of protein stability due to amino acid substitution and show that functionally constrained proteins are less susceptible to large mutational changes in stability. Mechanistically, this relates to greater intrinsic disorder among constrained proteins and to increased B-factors in the ordered regions of constrained proteins. This phenomenon means that constrained proteins exhibit smaller stability effects due to missense mutations, and partly explains why overtransmission of pathogenic missense variation is less prevalent in genetic disorders characterised by protein truncations. We show that the most functionally constrained proteins are depleted of both destabilising and overly-stabilising amino acid variation in disease-free populations. Despite this, amino acid substitutions with large stability effects in functionally constrained proteins are still highly prevalent among pathogenic human genetic variation. Importantly, we observe that there are approximately five times more missense variants with large stability effects than there are unambiguous loss-of-function mutations. Missense variants with disruption of stability effects recapitulate the per-gene patterns of functional constraint observed with protein truncating loss-of-function variation, yet their relative abundance abrogates difficulties encountered when estimating functional constraint for the shortest human genes.

The Kelch 3 motif on gigaxonin mediates the interaction with NUDCD3 and regulates vimentin filament morphology

Gigaxonin is an intermediate filament (IF)-interacting partner belonging to the Kelch-like (KLHL) protein family. Gigaxonin is encoded by the KLHL16 gene, which is mutated in Giant Axonal Neuropathy (GAN). The lack of functional gigaxonin in GAN patient cells impairs IF proteostasis, leading to focal abnormal accumulations of IFs and compromised neuronal function. We hypothesized that gigaxonin forms molecular interactions via specific sequence motifs to regulate IF proteostasis. The goal of this study was to examine how distinct Kelch motifs on gigaxonin regulate IF protein degradation and filament morphology. We analyzed vimentin IFs in HEK293 cells overexpressing wild type (WT) gigaxonin, or gigaxonin lacking each of the six individual Kelch motifs: K1 (aa274-326), K2 (aa327-374), K3 (aa376-421), K4 (aa422-468), K5 (aa470-522), and K6 (aa528-574). All six gigaxonin deletion mutants (ΔK1-ΔK6) promoted the degradation of soluble vimentin. The ΔK3 gigaxonin mutant exhibited soluble vimentin degradation and promoted the bundling of vimentin IFs relative to WT gigaxonin. Using mass spectrometry proteomic analysis we found that, relative to WT gigaxonin, ΔK3 gigaxonin had increased associations with ubiquitination-associated and mitochondrial proteins and lost the association with the NudC domain-containing protein 3 (NUDCD3), a molecular chaperone enriched in the nervous system. Collectively, our cell biological data show the induction of an abnormal GAN-like IF phenotype in cells expressing ΔK3-gigaxonin, while our mass spectrometry profiling links the loss of gigaxonin-NUDCD3 interactions with defective IF proteostasis, revealing NUDCD3 as a potential new target in GAN.

Insulin receptor variants: Extending the traditional Mendelian spectrum

PURPOSE

INSR encodes the insulin receptor, the essential entrainer of growth and metabolism to nutritional cues. INSR variants cause a spectrum of monogenic insulin resistance (IR) syndromes, namely, type A insulin resistance, Rabson-Mendenhall, and Donohue syndromes. However, to our knowledge, no large cohort studies focused on variant classification and its diagnostic value have been described.

METHODS

This multicentric cohort study included 73 patients carrying INSR variants, referred for IR by 52 centers from 6 countries. Variants were classified using new bioinformatic tools relying on different prediction mechanisms and the American College of Medical Genetics and Genomics guidelines.

RESULTS

Besides expanding the INSR mutational spectrum, this study suggested a semidominant inheritance in several Donohue/Rabson-Mendenhall syndrome families. Questioning strictly Mendelian inheritance, heterozygous loss-of-function (LoF) variants were mostly found in overweight patients, with a higher LoF frequency in IR patients than in the general population (odds ratio 5.77). Diagnostic challenges arose when trying to refine classification criteria for variants of uncertain significance. Among the variant effect predictors assessed, MISTIC and AlphaMissense outperformed REVEL.

CONCLUSION

The spectrum of INSR-related disorders extends beyond traditional entities. Heterozygous INSR LoF variants may increase IR susceptibility. International collaboration and functional assays are needed to drive precision medicine forward.

Congenital Fibrinogen Deficiencies: Not So Rare

Congenital fibrinogen deficiencies (CFDs), traditionally considered rare monogenic disorders, are now recognized as more prevalent and genetically complex than previously thought. Indeed, the symptoms manifested in CFD patients, such as bleeding and thrombosis, are likely to result from variation in several genes rather than solely driven by variants in one of the three fibrinogen genes, FGB, FGA, and FGG. This review highlights recent advances in understanding the genetic causes of CFD and their variability, facilitated by the growing use and availability of next-generation sequencing data. Using gnomAD v4.1.0. data, which includes more than 800,000 individuals, we provide updated global prevalence estimates for CFDs based on frequencies of predicted deleterious variants in FGB, FGA, and FGG. Recessively inherited fibrinogen deficiencies (homozygous genotypes) could be present in around 29 individuals per million, while dominantly inherited deficiencies (heterozygous genotypes) may be present in up to 15,000 per million. These increased estimates can be attributed to the inclusion of broader, more diverse genetic datasets in the new version of gnomAD, thus capturing a greater range of rare variants and homozygous cases.

Point mutations of the mitochondrial chaperone TRAP1 affect its functions and pro-neoplastic activity

The mitochondrial chaperone TRAP1 is a key regulator of cellular homeostasis and its activity has important implications in neurodegeneration, ischemia and cancer. Recent evidence has indicated that TRAP1 mutations are involved in several disorders, even though the structural basis for the impact of point mutations on TRAP1 functions has never been studied. By exploiting a modular structure-based framework and molecular dynamics simulations, we investigated the effect of five TRAP1 mutations on its structure and stability. Each mutation differentially impacts long-range interactions, intra and inter-protomer dynamics and ATPase activity. Changes in these parameters influence TRAP1 functions, as revealed by their effects on the activity of the TRAP1 interactor succinate dehydrogenase (SDH). In keeping with this, TRAP1 point mutations affect the growth and migration of aggressive sarcoma cells, and alter sensitivity to a selective TRAP1 inhibitor. Our work provides new insights on the structure-activity relationship of TRAP1, identifying crucial amino acid residues that regulate TRAP1 proteostatic functions and pro-neoplastic activity.

OMEGA-guided DNA polymerases enable random mutagenesis in a tunable window

Targeted random mutagenesis is crucial for breeding, directed evolution, and gene function studies, yet efficient tools remain scarce. Here, we present obligate mobile element guided activity (OMEGA)-R, an innovative targeted random mutagenesis system that integrates SpyCatcher-enIscB and PolI3M-TBD-SpyTag, outperforming existing state-of-the-art technologies in key metrics, such as protein size, mutagenesis efficiency, window length, and continuity. OMEGA-R achieves a dramatic enhancement of on-target mutagenesis, reaching a rate of 1.4 × 10-5 base pairs (bp) per generation (bpg), with minimal off-target effects, in both Escherichia coli and Bacillus subtilis. The system also demonstrates exceptional compatibility with high-throughput screening (HTS) technologies, including fluorescence-activated droplet sorting (FADS) and phage-assisted continuous evolution (PACE). Utilizing OMEGA-R, we successfully identified a series of effective mutations within the T7 promoter (pT7), ribosome-binding site (RBS), superfolder GFP (sfGFP), and autocyclizing ribozyme (AR), which are invaluable for the development of high-performance biotechnology tools. These findings underscore the high efficiency and broad application potential of OMEGA-R.

Calibration of additional computational tools expands ClinGen recommendation options for variant classification with PP3/BP4 criteria

PURPOSE

We previously developed an approach to calibrate computational tools for clinical variant classification, updating recommendations for the reliable use of variant impact predictors to provide evidence strength up to Strong. A new generation of tools using distinctive approaches has since been released, and these methods must be independently calibrated for clinical application.

METHODS

Using our local posterior probability-based calibration and our established data set of ClinVar pathogenic and benign variants, we determined the strength of evidence provided by 3 new tools (AlphaMissense, ESM1b, and VARITY) and calibrated scores meeting each evidence strength.

RESULTS

All 3 tools reached the Strong level of evidence for variant pathogenicity and Moderate for benignity, although sometimes for few variants. Compared with previously recommended tools, these yielded at best only modest improvements in the trade-offs between evidence strength and false-positive predictions.

CONCLUSION

At calibrated thresholds, 3 new computational predictors provided evidence for variant pathogenicity at similar strength to the 4 previously recommended predictors (and comparable with functional assays for some variants). This calibration broadens the scope of computational tools for application in clinical variant classification. Their new approaches offer promise for future advancement of the field.

Predicting Resistance to Small Molecule Kinase Inhibitors

Drug resistance is a critical challenge in treating diseases like cancer and infectious disease. This study presents a novel computational workflow for predicting on-target resistance mutations to small molecule inhibitors (SMIs). The approach integrates genetic models with alchemical free energy perturbation (FEP+) calculations to identify likely resistance mutations. Specifically, a genetic model, RECODE, leverages cancer-specific mutation patterns to prioritize probable amino acid changes. Physics-based calculations assess the impact of these mutations on protein stability, endogenous substrate binding, and inhibitor binding. We apply this approach retrospectively to gefitinib and osimertinib, two clinical epidermal growth factor receptor (EGFR) inhibitors used to treat nonsmall cell lung cancer (NSCLC). Among hundreds of possible mutations, the pipeline accurately predicted 4 out of 11 and 7 out of 19 known binding site mutations for gefitinib and osimertinib, respectively, including the clinically relevant T790M and C797S resistance mutations. This study demonstrates the potential of integrating genetic models and physics-based calculations to predict SMI resistance mutations. This approach can be applied to other kinases and target classes, potentially enabling the design of next-generation inhibitors with improved durability of response in patients.

PCSK5M452I is a recessive hypomorph exclusive to MCF10DCIS.com cells

The most widely used cell line for studying ductal carcinoma in situ (DCIS) premalignancy is the transformed breast epithelial cell line, MCF10DCIS.com. During its original clonal isolation and selection, MCF10DCIS.com acquired a heterozygous M452I mutation in the proprotein convertase PCSK5, which has never been reported in any human cancer. The mutation is noteworthy because PCSK5 matures GDF11, a TGFβ-superfamily ligand that suppresses progression of triple-negative breast cancer. We asked here whether PCSK5M452I and its activity toward GDF11 might contribute to the unique properties of MCF10DCIS.com. Using an optimized in-cell GDF11 maturation assay, we found that overexpressed PCSK5M452I was measurably active but at a fraction of the wildtype enzyme. In a PCSK5 -/- clone of MCF10DCIS.com reconstituted with different PCSK5 alleles, PCSK5M452I was mildly defective in anterograde transport. However, the multicellular organization of PCSK5M452I addback cells in 3D matrigel cultures was significantly less compact than wildtype and indistinguishable from a PCSK5T288P null allele. Growth of intraductal MCF10DCIS.com xenografts was similarly impaired along with the frequency of comedo necrosis and stromal activation. In no setting did PCSK5M452I exhibit gain-of-function activity, leading us to conclude that it is hypomorphic and thus compensated by the remaining wildtype allele in MCF10DCIS.com.

Implications

This work reassures that an exotic PCSK5 mutation is not responsible for the salient characteristics of the MCF10DCIS.com cell line.

Saturation mapping of MUTYH variant effects using DNA repair reporters

Variants of uncertain significance (VUS) limit the actionability of genetic testing. A prominent example is MUTYH, a base excision repair factor associated with polyposis and colorectal cancer, which has a pathogenic variant carrier rate approaching 1 in 50 individuals in some populations. To systematically interrogate variant function in MUTYH, we coupled deep mutational scanning with a DNA repair reporter containing its lesion substrate, 8OG:A. Our variant-to-function map covers >97% of all possible MUTYH point variants (n=10,941) and achieves 100% accuracy classifying the pathogenicity of known clinical variants (n=247). Leveraging a large clinical registry, we observe significant associations with colorectal polyps and cancer, with more severely impaired missense variants conferring greater risk. We recapitulate known functional differences between pathogenic founder alleles, and highlight sites of complete missense intolerance, including residues that intercalate DNA and coordinate essential Zn2+ or Fe-S clusters. This map provides a resource to resolve the 1,032 existing missense VUS and 90 variants with conflicting interpretations in MUTYH, and demonstrates a scalable strategy to interrogate other clinically relevant DNA repair factors.

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.

Phenotypic pleiotropy of missense variants in human B cell-confinement receptor P2RY8

Missense variants can have pleiotropic effects on protein function and predicting these effects can be difficult. We performed near-saturation deep mutational scanning of P2RY8, a G-protein-coupled receptor that promotes germinal center B cell confinement. We assayed the effect of each variant on surface expression, migration, and proliferation. We delineated variants that affected both expression and function, affected function independently of expression, and discrepantly affected migration and proliferation. We also used cryo-electron microscopy to determine the structure of activated, ligand-bound P2RY8, providing structural insights into the effects of variants on ligand binding and signal transmission. We applied the deep mutational scanning results to both improve computational variant effect predictions and to characterize the phenotype of germline variants and lymphoma-associated variants. Together, our results demonstrate the power of integrating deep mutational scanning, structure determination, and in silico prediction to advance the understanding of a receptor important in human health.

Protective targets of PfSPZ vaccines identified from whole-genome sieve analysis of isolates from malaria vaccine efficacy trials in West Africa

Identification of antigens targeted by a protective response is a central quest in malaria vaccinology. Whole-genome sieve analysis (SAWG) in samples collected from placebo-controlled field trials of Plasmodium falciparum (Pf) sporozoite (SPZ) vaccines may enable identification of Pf pre-erythrocytic antigens. We applied SAWG to genomic data generated from Pf isolates collected during two field trials measuring the efficacy, in malaria-exposed African adults, of two PfSPZ vaccines. These randomized, double-blind, placebo-controlled trials were conducted in regions of Mali and Burkina Faso characterized by high seasonal transmission, where parasite genetic diversity is high. Genomic sites in which the vaccine allelic state was significantly underrepresented among breakthrough infections in vaccinees relative to placebo recipients were termed "target sites". Protein-coding loci containing target sites that changed amino acids were termed "target loci". The SAWG conducted on clinical trial samples from the Burkina Faso and Mali trials identified 138 and 80 single-copy protein-coding target loci in the Burkinabe and Malian data sets, respectively, with twelve common to both, a number significantly higher than expected (E = 3.9; 99%CI = [0, 9]). Among these was the thrombospondin-related anonymous protein locus, which encodes PfSSP2|TRAP, one of the most abundant and well-characterized pre-erythrocytic stage antigen as well as other genes encoding membrane-associated proteins of unknown function. These results identify SAWG as a potentially powerful tool for identifying protective vaccine antigens in recombining pathogens with large genome size and reveals potential new protective Pf antigens.

A graph neural network approach for accurate prediction of pathogenicity in multi-type variants

Accurate prediction of pathogenic variants in human disease-associated genes would have a profound effect on clinical decision-making; however, it remains a significant challenge due to the overwhelming number of these variants. We propose graph neural network for multimodal annotation-based pathogenicity prediction (GNN-MAP), a novel deep learning framework that effectively integrates multimodal annotations and similarity relationships among variants to predict the pathogenicity of multi-type variants. Trained on the ClinVar dataset, GNN-MAP exhibits superior predictive performance in internal validation and orthogonal test datasets, accurately predicting variant pathogenicity. Notably, GNN-MAP enables accurate prediction of the pathogenicity of rare variants and highly imbalanced datasets. Furthermore, it achieves high performance in the pathogenicity prediction of inherited retinal disease-specific variants, highlighting its effectiveness in disease-specific variant prediction. These findings suggest that the robust capability of GNN-MAP to predict pathogenicity across multiple variant types and datasets holds significant potential for applications in research and clinical settings.

The costimulatory domain influences CD19 CAR-T cell resistance development in B-cell malignancies

CD19-CAR-T-cells emerge as a major therapeutic option for relapsed/refractory B-cell-derived malignancies, however approximately half of patients eventually relapse. To identify resistance-driving factors, we repeatedly exposed B-cell lymphoma/B-cell acute lymphoblastic leukemia to 4-1BB/CD28-based CD19-CAR-T-cells in vitro. Generated models revealed costimulatory domain-dependent differences in CD19 loss. While CD19-4-1BB-CAR-T-cells induced combination epitope/total CD19 protein loss, CD19-CD28-CAR-T-cells did not drive antigen-escape. Consistent with observations in patients relapsing after CD19-4-1BB-CAR-T-cells, we identified CD19 frameshift/missense mutations affecting residues critical for FMC63 epitope recognition. Mathematical simulations revealed that differences between CD19-4-1BB- and CD19-CD28-CAR-T-cells activity against low-antigen-expressing tumor contribute to heterogeneous therapeutic responses. By integrating in vitro and in silico data, we propose a biological scenario where CD19-4-1BB-CAR-T-cells fail to eliminate low-antigen tumor cells, fostering CAR-resistance. These findings offer mechanistic insight into the observed clinical differences between axi-cel (CD28-based) and tisa-cel (4-1BB-based)-treated B-cell lymphoma patients and advance our understanding on CAR-T resistance. Furthermore, we underscore the need for specific FMC63 epitope detection to deliver information on antigen levels accessible for CD19-CAR-T-cells.