Essentiality, protein-protein interactions and evolutionary properties are key predictors for identifying cancer-associated genes using machine learning.

The distinctive nature of cancer as a disease prompts an exploration of the special characteristics the genes implicated in cancer exhibit. The identification of cancer-associated genes and their characteristics is crucial to further our understanding of this disease and enhanced likelihood of therapeutic drug targets success. However, the rate at which cancer genes are being identified experimentally is slow. Applying predictive analysis techniques, through the building of accurate machine learning models, is potentially a useful approach in enhancing the identification rate of these genes and their characteristics. Here, we investigated gene essentiality scores and found that they tend to be higher for cancer-associated genes compared to other protein-coding human genes. We built a dataset of extended gene properties linked to essentiality and used it to train a machine-learning model; this model reached 89% accuracy and > 0.85 for the Area Under Curve (AUC). The model showed that essentiality, evolutionary-related properties, and properties arising from protein-protein interaction networks are particularly effective in predicting cancer-associated genes. We were able to use the model to identify potential candidate genes that have not been previously linked to cancer. Prioritising genes that score highly by our methods could aid scientists in their cancer genes research.

Clinical, genetic, epidemiologic, evolutionary, and functional delineation of TSPEAR-related autosomal recessive ectodermal dysplasia 14.

TSPEAR variants cause autosomal recessive ectodermal dysplasia (ARED) 14. The function of TSPEAR is unknown. The clinical features, the mutation spectrum, and the underlying mechanisms of ARED14 are poorly understood. Combining data from new and previously published individuals established that ARED14 is primarily characterized by dental anomalies such as conical tooth cusps and hypodontia, like those seen in individuals with WNT10A-related odontoonychodermal dysplasia. AlphaFold-predicted structure-based analysis showed that most of the pathogenic TSPEAR missense variants likely destabilize the ß-propeller of the protein. Analysis of 100000 Genomes Project (100KGP) data revealed multiple founder TSPEAR variants across different populations. Mutational and recombination clock analyses demonstrated that non-Finnish European founder variants likely originated around the end of the last ice age, a period of major climatic transition. Analysis of gnomAD data showed that the non-Finnish European population TSPEAR gene-carrier rate is ∼1/140, making it one of the commonest AREDs. Phylogenetic and AlphaFold structural analyses showed that TSPEAR is an ortholog of drosophila Closca, an extracellular matrix-dependent signaling regulator. We, therefore, hypothesized that TSPEAR could have a role in enamel knot, a structure that coordinates patterning of developing tooth cusps. Analysis of mouse single-cell RNA sequencing (scRNA-seq) data revealed highly restricted expression of Tspear in clusters representing enamel knots. A tspeara -/-;tspearb -/- double-knockout zebrafish model recapitulated the clinical features of ARED14 and fin regeneration abnormalities of wnt10a knockout fish, thus suggesting interaction between tspear and wnt10a. In summary, we provide insights into the role of TSPEAR in ectodermal development and the evolutionary history, epidemiology, mechanisms, and consequences of its loss of function variants.

Assessing the Pathogenicity of In-Frame CACNA1F Indel Variants Using Structural Modeling.

Small in-frame insertion-deletion (indel) variants are a common form of genomic variation whose impact on rare disease phenotypes has been understudied. The prediction of the pathogenicity of such variants remains challenging. X-linked incomplete congenital stationary night blindness type 2 (CSNB2) is a nonprogressive, inherited retinal disorder caused by variants in CACNA1F, encoding the Cav1.4α1 channel protein. Here, structural analysis was used through homology modeling to interpret 10 disease-correlated and 10 putatively benign CACNA1F in-frame indel variants. CSNB2-correlated changes were found to be more highly conserved compared with putative benign variants. Notably, all 10 disease-correlated variants but none of the benign changes were within modeled regions of the protein. Structural analysis revealed that disease-correlated variants are predicted to destabilize the structure and function of the Cav1.4α1 channel protein. Overall, the use of structural information to interpret the consequences of in-frame indel variants provides an important adjunct that can improve the diagnosis for individuals with CSNB2.

Improving the clinical interpretation of missense variants in X linked genes using structural analysis.

BACKGROUND: Improving the clinical interpretation of missense variants can increase the diagnostic yield of genomic testing and lead to personalised management strategies. Currently, due to the imprecision of bioinformatic tools that aim to predict variant pathogenicity, their role in clinical guidelines remains limited. There is a clear need for more accurate prediction algorithms and this study aims to improve performance by harnessing structural biology insights. The focus of this work is missense variants in a subset of genes associated with X linked disorders. METHODS: We have developed a protein-specific variant interpreter (ProSper) that combines genetic and protein structural data. This algorithm predicts missense variant pathogenicity by applying machine learning approaches to the sequence and structural characteristics of variants. RESULTS: ProSper outperformed seven previously described tools, including meta-predictors, in correctly evaluating whether or not variants are pathogenic; this was the case for 11 of the 21 genes associated with X linked disorders that met the inclusion criteria for this study. We also determined gene-specific pathogenicity thresholds that improved the performance of VEST4, REVEL and ClinPred, the three best-performing tools out of the seven that were evaluated; this was the case in 11, 11 and 12 different genes, respectively. CONCLUSION: ProSper can form the basis of a molecule-specific prediction tool that can be implemented into diagnostic strategies. It can allow the accurate prioritisation of missense variants associated with X linked disorders, aiding precise and timely diagnosis. In addition, we demonstrate that gene-specific pathogenicity thresholds for a range of missense prioritisation tools can lead to an increase in prediction accuracy.

Insight into the stacking and the species-ordering dependences of interlayer bonding in SiC/GeC polar heterostructures.

Two-dimensional (2D) polar materials experience an in-plane charge transfer between different elements due to their electron negativities. When they form vertical heterostructures, the electrostatic force triggered by such charge transfer plays an important role in the interlayer bonding beyond van der Waals (vdW) interaction. Our comprehensive first principle study on the structural stability of the 2D SiC/GeC hybrid bilayer heterostructure has found that the electrostatic interlayer interaction can induce theπ-πorbital hybridization between adjacent layers under different stacking and out-of-plane species ordering, with strong hybridization in the cases of Si-C and C-Ge species orderings but weak hybridization in the case of the C-C ordering. In particular, the attractive electrostatic interlayer interaction in the cases of Si-C and C-Ge species orderings mainly controls the equilibrium interlayer distance and the vdW interaction makes the system attain a lower binding energy. On the contrary, the vdW interaction mostly controls the equilibrium interlayer distance in the case of the C-C species ordering and the repulsive electrostatic interlayer force has less effect. Interesting finding is that the band structure of the SiC/GeC hybrid bilayer is sensitive to the layer-layer stacking and the out-of-plane species ordering. An indirect band gap of 2.76 eV (or 2.48 eV) was found under the AA stacking with Si-C ordering (or under the AB stacking with C-C ordering). While a direct band gap of 2.00-2.88 eV was found under other stacking and species orderings, demonstrating its band gap tunable feature. Furthermore, there is a charge redistribution in the interfacial region leading to a built-in electric field. Such field will separate the photo-generated charge carriers in different layers and is expected to reduce the probability of carrier recombination, and eventually give rise to the electron tunneling between layers.

Recurrent KCNT2 missense variants affecting p.Arg190 result in a recognizable phenotype.

KCNT2 variants resulting in substitutions affecting the Arg190 residue have been shown to cause epileptic encephalopathy and a recognizable facial gestalt. We report two additional individuals with intellectual disability, dysmorphic features, hypertrichosis, macrocephaly and the same de novo KCNT2 missense variants affecting the Arg190 residue as previously described. Notably, neither patient has epilepsy. Homology modeling of these missense variants revealed that they are likely to disrupt the stabilization of a closed channel conformation of KCNT2 resulting in a constitutively open state. This is the first report of pathogenic variants in KCNT2 causing a developmental phenotype without epilepsy.

EVI1 phosphorylation at S436 regulates interactions with CtBP1 and DNMT3A and promotes self-renewal.

The transcriptional regulator EVI1 has an essential role in early development and haematopoiesis. However, acute myeloid leukaemia (AML) driven by aberrantly high EVI1 expression has very poor prognosis. To investigate the effects of post-translational modifications on EVI1 function, we carried out a mass spectrometry (MS) analysis of EVI1 in AML and detected dynamic phosphorylation at serine 436 (S436). Wild-type EVI1 (EVI1-WT) with S436 available for phosphorylation, but not non-phosphorylatable EVI1-S436A, conferred haematopoietic progenitor cell self-renewal and was associated with significantly higher organised transcriptional patterns. In silico modelling of EVI1-S436 phosphorylation showed reduced affinity to CtBP1, and CtBP1 showed reduced interaction with EVI1-WT compared with EVI1-S436A. The motif harbouring S436 is a target of CDK2 and CDK3 kinases, which interacted with EVI1-WT. The methyltransferase DNMT3A bound preferentially to EVI1-WT compared with EVI1-S436A, and a hypomethylated cell population associated by EVI1-WT expression in murine haematopoietic progenitors is not maintained with EVI1-S436A. These data point to EVI1-S436 phosphorylation directing functional protein interactions for haematopoietic self-renewal. Targeting EVI1-S436 phosphorylation may be of therapeutic benefit when treating EVI1-driven leukaemia.

Using an integrative machine learning approach utilising homology modelling to clinically interpret genetic variants: CACNA1F as an exemplar.

Advances in DNA sequencing technologies have revolutionised rare disease diagnostics and have led to a dramatic increase in the volume of available genomic data. A key challenge that needs to be overcome to realise the full potential of these technologies is that of precisely predicting the effect of genetic variants on molecular and organismal phenotypes. Notably, despite recent progress, there is still a lack of robust in silico tools that accurately assign clinical significance to variants. Genetic alterations in the CACNA1F gene are the commonest cause of X-linked incomplete Congenital Stationary Night Blindness (iCSNB), a condition associated with non-progressive visual impairment. We combined genetic and homology modelling data to produce CACNA1F-vp, an in silico model that differentiates disease-implicated from benign missense CACNA1F changes. CACNA1F-vp predicts variant effects on the structure of the CACNA1F encoded protein (a calcium channel) using parameters based upon changes in amino acid properties; these include size, charge, hydrophobicity, and position. The model produces an overall score for each variant that can be used to predict its pathogenicity. CACNA1F-vp outperformed four other tools in identifying disease-implicated variants (area under receiver operating characteristic and precision recall curves = 0.84; Matthews correlation coefficient = 0.52) using a tenfold cross-validation technique. We consider this protein-specific model to be a robust stand-alone diagnostic classifier that could be replicated in other proteins and could enable precise and timely diagnosis.

Variability in Gene Expression is Associated with Incomplete Penetrance in Inherited Eye Disorders.

Inherited eye disorders (IED) are a heterogeneous group of Mendelian conditions that are associated with visual impairment. Although these disorders often exhibit incomplete penetrance and variable expressivity, the scale and mechanisms of these phenomena remain largely unknown. Here, we utilize publicly-available genomic and transcriptomic datasets to gain insights into variable penetrance in IED. Variants in a curated set of 340 IED-implicated genes were extracted from the Human Gene Mutation Database (HGMD) 2019.1 and cross-checked with the Genome Aggregation Database (gnomAD) 2.1 control-only dataset. Genes for which >1 variants were encountered in both HGMD and gnomAD were considered to be associated with variable penetrance (n = 56). Variability in gene expression levels was then estimated for the subset of these genes that was found to be adequately expressed in two relevant resources: the Genotype-Tissue Expression (GTEx) and Eye Genotype Expression (EyeGEx) datasets. We found that genes suspected to be associated with variable penetrance tended to have significantly more variability in gene expression levels in the general population (p = 0.0000015); this finding was consistent across tissue types. The results of this study point to the possible influence of cis and/or trans-acting elements on the expressivity of variants causing Mendelian disorders. They also highlight the potential utility of quantifying gene expression as part of the investigation of families showing evidence of variable penetrance.

Severe type I interferonopathy and unrestrained interferon signaling due to a homozygous germline mutation in STAT2.

Excessive type I interferon (IFNα/ß) activity is implicated in a spectrum of human disease, yet its direct role remains to be conclusively proven. We investigated two siblings with severe early-onset autoinflammatory disease and an elevated IFN signature. Whole-exome sequencing revealed a shared homozygous missense Arg148Trp variant in STAT2, a transcription factor that functions exclusively downstream of innate IFNs. Cells bearing STAT2R148W in homozygosity (but not heterozygosity) were hypersensitive to IFNα/ß, which manifest as prolonged Janus kinase-signal transducers and activators of transcription (STAT) signaling and transcriptional activation. We show that this gain of IFN activity results from the failure of mutant STAT2R148W to interact with ubiquitin-specific protease 18, a key STAT2-dependent negative regulator of IFNα/ß signaling. These observations reveal an essential in vivo function of STAT2 in the regulation of human IFNα/ß signaling, providing concrete evidence of the serious pathological consequences of unrestrained IFNα/ß activity and supporting efforts to target this pathway therapeutically in IFN-associated disease.

Reliable Generation of Native-Like Decoys Limits Predictive Ability in Fragment-Based Protein Structure Prediction.

Our previous work with fragment-assembly methods has demonstrated specific deficiencies in conformational sampling behaviour that, when addressed through improved sampling algorithms, can lead to more reliable prediction of tertiary protein structure when good fragments are available, and when score values can be relied upon to guide the search to the native basin. In this paper, we present preliminary investigations into two important questions arising from more difficult prediction problems. First, we investigated the extent to which native-like conformational states are generated during multiple runs of our search protocols. We determined that, in cases of difficult prediction, native-like decoys are rarely or never generated. Second, we developed a scheme for decoy retention that balances the objectives of retaining low-scoring structures and retaining conformationally diverse structures sampled during the course of the search. Our method succeeds at retaining more diverse sets of structures, and, for a few targets, more native-like solutions are retained as compared to our original, energy-based retention scheme. However, in general, we found that the rate at which native-like structural states are generated has a much stronger effect on eventual distributions of predictive accuracy in the decoy sets, as compared to the specific decoy retention strategy used. We found that our protocols show differences in their ability to access native-like states for some targets, and this may explain some of the differences in predictive performance seen between these methods. There appears to be an interaction between fragment sets and move operators, which influences the accessibility of native-like structures for given targets. Our results point to clear directions for further improvements in fragment-based methods, which are likely to enable higher accuracy predictions.

Mutations in PCYT2 disrupt etherlipid biosynthesis and cause a complex hereditary spastic paraplegia.

CTP:phosphoethanolamine cytidylyltransferase (ET), encoded by PCYT2, is the rate-limiting enzyme for phosphatidylethanolamine synthesis via the CDP-ethanolamine pathway. Phosphatidylethanolamine is one of the most abundant membrane lipids and is particularly enriched in the brain. We identified five individuals with biallelic PCYT2 variants clinically characterized by global developmental delay with regression, spastic para- or tetraparesis, epilepsy and progressive cerebral and cerebellar atrophy. Using patient fibroblasts we demonstrated that these variants are hypomorphic, result in altered but residual ET protein levels and concomitant reduced enzyme activity without affecting mRNA levels. The significantly better survival of hypomorphic CRISPR-Cas9 generated pcyt2 zebrafish knockout compared to a complete knockout, in conjunction with previously described data on the Pcyt2 mouse model, indicates that complete loss of ET function may be incompatible with life in vertebrates. Lipidomic analysis revealed profound lipid abnormalities in patient fibroblasts impacting both neutral etherlipid and etherphospholipid metabolism. Plasma lipidomics studies also identified changes in etherlipids that have the potential to be used as biomarkers for ET deficiency. In conclusion, our data establish PCYT2 as a disease gene for a new complex hereditary spastic paraplegia and confirm that etherlipid homeostasis is important for the development and function of the brain.

Biochemical and structural analyses suggest that plasminogen activators coevolved with their cognate protein substrates and inhibitors.

Protein sequences of members of the plasminogen activation system are present throughout the entire vertebrate phylum. This important and well-described proteolytic cascade is governed by numerous protease-substrate and protease-inhibitor interactions whose conservation is crucial to maintaining unchanged protein function throughout evolution. The pressure to preserve protein-protein interactions may lead to either co-conservation or covariation of binding interfaces. Here, we combined covariation analysis and structure-based prediction to analyze the binding interfaces of urokinase (uPA):plasminogen activator inhibitor-1 (PAI-1) and uPA:plasminogen complexes. We detected correlated variation between the S3-pocket-lining residues of uPA and the P3 residue of both PAI-1 and plasminogen. These residues are known to form numerous polar interactions in the human uPA:PAI-1 Michaelis complex. To test the effect of mutations that correlate with each other and have occurred during mammalian diversification on protein-protein interactions, we produced uPA, PAI-1, and plasminogen from human and zebrafish to represent mammalian and nonmammalian orthologs. Using single amino acid point substitutions in these proteins, we found that the binding interfaces of uPA:plasminogen and uPA:PAI-1 may have coevolved to maintain tight interactions. Moreover, we conclude that although the interaction areas between protease-substrate and protease-inhibitor are shared, the two interactions are mechanistically different. Compared with a protease cleaving its natural substrate, the interaction between a protease and its inhibitor is more complex and involves a more fine-tuned mechanism. Understanding the effects of evolution on specific protein interactions may help further pharmacological interventions of the plasminogen activation system and other proteolytic systems.

Improved fragment-based protein structure prediction by redesign of search heuristics.

Difficulty in sampling large and complex conformational spaces remains a key limitation in fragment-based de novo prediction of protein structure. Our previous work has shown that even for small-to-medium-sized proteins, some current methods inadequately sample alternative structures. We have developed two new conformational sampling techniques, one employing a bilevel optimisation framework and the other employing iterated local search. We combine strategies of forced structural perturbation (where some fragment insertions are accepted regardless of their impact on scores) and greedy local optimisation, allowing greater exploration of the available conformational space. Comparisons against the Rosetta Abinitio method indicate that our protocols more frequently generate native-like predictions for many targets, even following the low-resolution phase, using a given set of fragment libraries. By contrasting results across two different fragment sets, we show that our methods are able to better take advantage of high-quality fragments. These improvements can also translate into more reliable identification of near-native structures in a simple clustering-based model selection procedure. We show that when fragment libraries are sufficiently well-constructed, improved breadth of exploration within runs improves prediction accuracy. Our results also suggest that in benchmarking scenarios, a total exclusion of fragments drawn from homologous templates can make performance differences between methods appear less pronounced.

EVI1 carboxy-terminal phosphorylation is ATM-mediated and sustains transcriptional modulation and self-renewal via enhanced CtBP1 association.

The transcriptional regulator EVI1 has an essential role in early hematopoiesis and development. However, aberrantly high expression of EVI1 has potent oncogenic properties and confers poor prognosis and chemo-resistance in leukemia and solid tumors. To investigate to what extent EVI1 function might be regulated by post-translational modifications we carried out mass spectrometry- and antibody-based analyses and uncovered an ATM-mediated double phosphorylation of EVI1 at the carboxy-terminal S858/S860 SQS motif. In the presence of genotoxic stress EVI1-WT (SQS), but not site mutated EVI1-AQA was able to maintain transcriptional patterns and transformation potency, while under standard conditions carboxy-terminal mutation had no effect. Maintenance of hematopoietic progenitor cell clonogenic potential was profoundly impaired with EVI1-AQA compared with EVI1-WT, in particular in the presence of genotoxic stress. Exploring mechanistic events underlying these observations, we showed that after genotoxic stress EVI1-WT, but not EVI1-AQA increased its level of association with its functionally essential interaction partner CtBP1, implying a role for ATM in regulating EVI1 protein interactions via phosphorylation. This aspect of EVI1 regulation is therapeutically relevant, as chemotherapy-induced genotoxicity might detrimentally sustain EVI1 function via stress response mediated phosphorylation, and ATM-inhibition might be of specific targeted benefit in EVI1-overexpressing malignancies.

MolProbity: More and better reference data for improved all-atom structure validation.

This paper describes the current update on macromolecular model validation services that are provided at the MolProbity website, emphasizing changes and additions since the previous review in 2010. There have been many infrastructure improvements, including rewrite of previous Java utilities to now use existing or newly written Python utilities in the open-source CCTBX portion of the Phenix software system. This improves long-term maintainability and enhances the thorough integration of MolProbity-style validation within Phenix. There is now a complete MolProbity mirror site at http://molprobity.manchester.ac.uk. GitHub serves our open-source code, reference datasets, and the resulting multi-dimensional distributions that define most validation criteria. Coordinate output after Asn/Gln/His "flip" correction is now more idealized, since the post-refinement step has apparently often been skipped in the past. Two distinct sets of heavy-atom-to-hydrogen distances and accompanying van der Waals radii have been researched and improved in accuracy, one for the electron-cloud-center positions suitable for X-ray crystallography and one for nuclear positions. New validations include messages at input about problem-causing format irregularities, updates of Ramachandran and rotamer criteria from the million quality-filtered residues in a new reference dataset, the CaBLAM Cα-CO virtual-angle analysis of backbone and secondary structure for cryoEM or low-resolution X-ray, and flagging of the very rare cis-nonProline and twisted peptides which have recently been greatly overused. Due to wide application of MolProbity validation and corrections by the research community, in Phenix, and at the worldwide Protein Data Bank, newly deposited structures have continued to improve greatly as measured by MolProbity's unique all-atom clashscore.

Non-muscle myosin IIB (Myh10) is required for epicardial function and coronary vessel formation during mammalian development.

The coronary vasculature is an essential vessel network providing the blood supply to the heart. Disruptions in coronary blood flow contribute to cardiac disease, a major cause of premature death worldwide. The generation of treatments for cardiovascular disease will be aided by a deeper understanding of the developmental processes that underpin coronary vessel formation. From an ENU mutagenesis screen, we have isolated a mouse mutant displaying embryonic hydrocephalus and cardiac defects (EHC). Positional cloning and candidate gene analysis revealed that the EHC phenotype results from a point mutation in a splice donor site of the Myh10 gene, which encodes NMHC IIB. Complementation testing confirmed that the Myh10 mutation causes the EHC phenotype. Characterisation of the EHC cardiac defects revealed abnormalities in myocardial development, consistent with observations from previously generated NMHC IIB null mouse lines. Analysis of the EHC mutant hearts also identified defects in the formation of the coronary vasculature. We attribute the coronary vessel abnormalities to defective epicardial cell function, as the EHC epicardium displays an abnormal cell morphology, reduced capacity to undergo epithelial-mesenchymal transition (EMT), and impaired migration of epicardial-derived cells (EPDCs) into the myocardium. Our studies on the EHC mutant demonstrate a requirement for NMHC IIB in epicardial function and coronary vessel formation, highlighting the importance of this protein in cardiac development and ultimately, embryonic survival.

Adaptive HIV-1 evolutionary trajectories are constrained by protein stability.

Despite the use of combination antiretroviral drugs for the treatment of HIV-1 infection, the emergence of drug resistance remains a problem. Resistance may be conferred either by a single mutation or a concerted set of mutations. The involvement of multiple mutations can arise due to interactions between sites in the amino acid sequence as a consequence of the need to maintain protein structure. To better understand the nature of such epistatic interactions, we reconstructed the ancestral sequences of HIV-1's Pol protein, and traced the evolutionary trajectories leading to mutations associated with drug resistance. Using contemporary and ancestral sequences we modelled the effects of mutations (i.e. amino acid replacements) on protein structure to understand the functional effects of residue changes. Although the majority of resistance-associated sequences tend to destabilise the protein structure, we find there is a general tendency for protein stability to decrease across HIV-1's evolutionary history. That a similar pattern is observed in the non-drug resistance lineages indicates that non-resistant mutations, for example, associated with escape from the immune response, also impacts on protein stability. Maintenance of optimal protein structure therefore represents a major constraining factor to the evolution of HIV-1.

Rapid functional and evolutionary changes follow gene duplication in yeast.

Duplication of genes or genomes provides the raw material for evolutionary innovation. After duplication a gene may be lost, recombine with another gene, have its function modified or be retained in an unaltered state. The fate of duplication is usually studied by comparing extant genomes and reconstructing the most likely ancestral states. Valuable as this approach is, it may miss the most rapid evolutionary events. Here, we engineered strains of Saccharomyces cerevisiae carrying tandem and non-tandem duplications of the singleton gene IFA38 to monitor (i) the fate of the duplicates in different conditions, including time scale and asymmetry of gene loss, and (ii) the changes in fitness and transcriptome of the strains immediately after duplication and after experimental evolution. We found that the duplication brings widespread transcriptional changes, but a fitness advantage is only present in fermentable media. In respiratory conditions, the yeast strains consistently lose the non-tandem IFA38 gene copy in a surprisingly short time, within only a few generations. This gene loss appears to be asymmetric and dependent on genome location, since the original IFA38 copy and the tandem duplicate are retained. Overall, this work shows for the first time that gene loss can be extremely rapid and context dependent.