Biology subjects

Schneidman-Duhovny, D.

Publications and source records attributed to Schneidman-Duhovny, D..

At least 19 recordsLinked to original sources

bioRxiv2026

A structural antibody benchmark for leakage-aware deep-learning evaluation

Deep-learning methods for antibody structure prediction, antibody-antigen interaction modelling and design are advancing rapidly. However, comparisons across studies remain difficult because training and test sets are often constructed independently, and a temporal cutoff alone does not prevent train-test leakage. We present SABLE (Structural Antibody Benchmark for deep-Learning Evaluation), a versioned structural antibody resource that couples a fixed training collection with a leakage-controlled held-out test set for reproducible machine-learning development and evaluation. SABLE combines 16,511 experimental training entries with 327 manually reviewed test entries and 3,274 high-confidence, patent-derived AlphaFold3 models spanning 476 antigens. Candidate test structures were selected after the AlphaFold3 temporal cutoff and filtered against the training set using antibody and antigen sequence similarity filters. Each test entry records its nearest training set neighbour, allowing users to quantify remaining relatedness and stratify performance by similarity. Versioned releases provide metadata, processed structures, CDR annotations, redundancy labels and model-confidence fields. A Python/PyTorch API and standardised benchmark metrics provide reproducible database access and evaluation code without requiring additional antibody-structure processing.

bioinformatics↗

bioRxiv2026

CliPepPI: Scalable prediction of domain-peptide specificityusing contrastive learning

Domain-peptide interactions mediate a significant fraction of cellular protein networks, yet accurately predicting their specificity remains challenging. Peptide motifs typically have short, fuzzy sequence profiles, and their interactions are often weak and transient, limiting the size, coverage, and quality of experimentally validated domain-peptide datasets. Since true non-binders are rarely known, constructing negative examples often introduces bias. While structure-based prediction methods can achieve high accuracy, they are computationally demanding and difficult to scale to the proteome level. We introduce CLIPepPI, a dual-encoder model that leverages contrastive learning to embed domains and peptides into a shared space directly from sequence. Both encoders are initialized from a protein language model (ESM-C) and fine-tuned using lightweight LoRA adapters, enabling parameter-efficient training on positive pairs alone. To overcome data scarcity, we augment [~]3K protein-peptide complexes from PPI3D with [~]150K domain-peptide pairs derived from protein-protein interfaces. CLIPepPI further injects structural information by marking interface residues in the domain sequence, thus guiding the encoders toward binding regions and linking sequence-level learning with structural context. Competitive performance is achieved across three independent benchmarks: domain-peptide complexes from PPI3D, large-scale phage-library data from ProP-PD, and a curated dataset of nuclear export signal (NES) sequences. We demonstrate scalability and generalization through two applications: (i) proteome-wide NES scanning, and (ii) variant-effect prediction, where score changes in domain-peptide interactions between wild-type and mutant sequences discriminate pathogenic from benign variants. Together, CLIPepPI offers a scalable, structure-informed model for predicting domain-peptide specificity and generating meaningful embeddings suited for large-scale proteomic analyses. CLIPepPI is available at: https://bio3d.cs.huji.ac.il/webserver/clipeppi/.

bioinformatics↗

bioRxiv2025

Orthosteric Molecular Glue Inhibits COP9 Signalosome with Substrate-Dependent Potency

Orthosteric inhibitors block enzyme active sites and prevent substrates from binding. Enhancing their specificity through substrate dependence seems inherently unlikely, as their mechanism hinges on direct competition rather than selective recognition. Here, we show that a molecular glue mechanism unexpectedly imparts substrate-dependent potency to CSN5i-3, an orthosteric inhibitor of the COP9 signalosome (CSN). We first confirm that CSN5i-3 inhibits CSN, which catalyzes NEDD8 deconjugation from the cullin-RING ubiquitin ligases (CRLs), by occupying the active site of its catalytic subunit, CSN5, and directly competing with the iso-peptide bond substrate. Curiously, the orthosteric inhibitor binds free CSN with only micromolar affinity, yet achieves nanomolar potency in blocking its deneddylase activity. Cryo-EM structures of the enzyme-substrate-inhibitor complex reveal that active site-engaged CSN5i-3 occludes the substrate iso-peptide linkage while simultaneously extending an NEDD8-binding exosite of CSN5, acting as a molecular glue to cement the NEDD8-CSN5 interaction. The cooperativity of this tri-molecular CSN5i-3-NEDD8-CSN5 assembly, in turn, sequesters CSN5i-3 at its binding site, conferring high potency to the orthosteric inhibitor despite its low affinity for the free enzyme. Together, our findings highlight the modest affinity requirements of molecule glues for individual target proteins and establish "orthosteric molecular glue inhibitors" as a new class of substrate-dependent enzyme antagonists.

molecular biology↗

bioRxiv2025

Calibrated Variant Effect Prediction at the Residue Level Using Conditional Score Distributions

Effective clinical use of variant effect prediction (VEP) requires models that are both accurate and well-calibrated. Calibration refers to a model's ability to produce meaningful and reliable probability estimates. Benchmarking 24 VEPs, we show that while models appear well-calibrated on average, they remain markedly miscalibrated within specific variant subgroups. We propose VEP calibration at the residue level and introduce a calibration approach based on differential score mapping per variant subgroup. When calibration targets are chosen appropriately, our calibration approach can also improve model discrimination. Leveraging these insights, we develop RaCoon (Residue-aware Calibration via Conditional Distributions), implemented on ESM1b, which provides multicalibrated and interpretable predictions. RaCoon introduces Gaussian mixture models-based sampling fitted on model score distribution which substantially reduces labeled-data requirements for calibration. The method maintains low calibration error across variant subgroups and datasets while also improving AUROC on multiple benchmarks. Our calibration strategy is readily transferable to other VEPs.

genomics↗

bioRxiv2025

Deciphering the mechanistic basis for the pathological effect of the Gαo E246K mutation in neurodevelopmental disorder

Mutations in the GNAO1 gene, which encodes for Go, a major neuronal G protein, are associated with neurodevelopmental disorders, epilepsy, and movement disorders. We identified and characterized a spontaneous heterozygous GNAO1 E246K mutation in an Israeli female infant with complex developmental delays and substantial motor difficulties. This mutation has been reported in other cases as a prevalent pathogenic mutation in patients with motor dysfunction and a broad range of neurological outcomes. To investigate the molecular and functional consequences of the Go E246K mutation, we employed structural modeling and analysis, biochemical assays, mass spectrometry-based proteomics, and cellular functional assays. We show that this mutation does not affect nucleotide binding, nor basal or RGS- accelerated GTP hydrolysis. Despite the E246 position located within a predicted effector binding region, proteomics analysis did not identify any new cellular partners. Instead, we demonstrate that the E246K mutation disrupts the Go regulatory GTPase cycle by directly impairing G{beta}{gamma} dissociation. This impairment overrides the presence of wild-type Go, explaining the dominant effect of the severe neurogenetic phenotype in the heterozygous background. These findings establish a new molecular mechanism for a GNAO1 mutation with dominant-negative effects on the GTPase regulatory cycle. The insights gained from studying this mechanism of action provide a basis for developing specific and personalized therapeutic strategies based on the outcome of a missense mutation in GNAO1.

neuroscience↗

bioRxiv2024

DockFormer: Efficient Multi-Modal Receptor-Ligand Interaction Prediction using Pair Transformer

Protein-small molecule interactions, or receptor-ligand interactions, are essential for understanding biological processes and advancing drug design. Despite advancements, existing prediction models of these interactions still lack capabilities and accuracy needed to replace traditional screening. In this paper, we introduce DockFormer, a method that leverages multi-modal learning to predict both the binding affinity and structure of these interactions. DockFormer employs fully flexible docking, where no part of the receptor remains rigid, by adapting the AlphaFold2 architecture. Instead of relying on protein sequences and Multiple Sequence Alignments, DockFormer uses predicted receptor structures as input. This modification enables the model to concentrate on ligand docking prediction rather than protein folding, while preserving full receptor flexibility. The stream-lined design also reduces the model size to just 8 layers, compared to AlphaFold2s 48 layers, greatly accelerating the inference process and making it more efficient for large-scale screening. When evaluated on affinity benchmarks such as CASF-2016, PLINDER, and the recently released CASP16 ligand screening benchmark, DockFormer performs comparably to or better than state-of-the-art methods, which typically rely on templates or bound structures as input. On structural benchmarks such as Posebusters and PLINDER, DockFormer demonstrated success rate of 20% and 15%, respectively.

bioinformatics↗

bioRxiv2024

EspH utilizes phosphoinositide and Rab binding domains to interact with plasma membrane infection sites and Rab GTPases

Enteropathogenic E. coli (EPEC) is a Gram-negative bacterial pathogen that causes persistent diarrhea. Upon attachment to the apical plasma membrane of the intestinal epithelium, the pathogen translocates virulent proteins called effectors into the infected cells. These effectors hijack numerous host processes for the pathogens benefit. Therefore, studying the mechanisms underlying their action is crucial for a better understanding of the disease. We show that translocated EspH interacts with multiple host Rab GTPases. AlphaFold predictions and site-directed mutagenesis identified glutamic acid and lysine at positions 37 and 41 as Rab interacting residues. Mutating these sites abolished the EspH ability to inhibit the Akt and mTORC1 signaling, lysosomal exocytosis, and bacterial invasion. Knocking out the endogenous Rab8a gene expression highlighted the involvement of Rab8a in Akt/mTORC1 signaling and lysosomal exocytosis. A phosphoinositide binding domain with a critical tyrosine was identified in EspH. Mutating the tyrosine abolished the localization of EspH at infection sites and its capacity to interact with Rabs. Our data suggest novel EspH-dependent mechanisms that elicit immune signaling and membrane trafficking during EPEC infection.

cell biology↗

bioRxiv2024

Predicting RNA Structure and Dynamics with Deep Learning and Solution Scattering

Advanced deep learning and statistical methods can predict structural models for RNA molecules. However, RNAs are flexible, and it remains difficult to describe their macromolecular conformations in solutions where varying conditions can induce conformational changes. Small-angle X-ray scattering (SAXS) in solution is an efficient technique to validate structural predictions by comparing the experimental SAXS profile with those calculated from predicted structures. There are two main challenges in comparing SAXS profiles to RNA structures: the absence of cations essential for stability and charge neutralization in predicted structures and the inadequacy of a single structure to represent RNAs conformational plasticity. We introduce Solution Conformation Predictor for RNA (SCOPER) to address these challenges. This pipeline integrates kinematics-based conformational sampling with the innovative deep-learning model, IonNet, designed for predicting Mg2+ ion binding sites. Validated through benchmarking against fourteen experimental datasets, SCOPER significantly improved the quality of SAXS profile fits by including Mg2+ ions and sampling of conformational plasticity. We observe that an increased content of monovalent and bivalent ions leads to decreased RNA plasticity. Therefore, carefully adjusting the plasticity and ion density is crucial to avoid overfitting experimental SAXS data. SCOPER is an efficient tool for accurately validating the solution state of RNAs given an initial, sufficiently accurate structure and provides the corrected atomistic model, including ions. The method is available from: https://github.com/dina-lab3d/IonNet Our pipeline is available for use as a web server: https://bilbomd.bl1231.als.lbl.gov/ Statement of SignificanceUnderstanding the behavior of RNA in solution is critical for deciphering its biological functions, yet predicting its macromolecular conformation remains challenging. While advanced computational methods can predict RNA structures, their accuracy in solution is often limited by the absence of stabilizing ions and the failure to account for RNAs conformational flexibility. This study presents SCOPER, an innovative tool that addresses these challenges by integrating deep-learning-based ion binding site prediction with conformational sampling, offering a more reliable approach to validate and refine RNA structures against experimental SAXS data. We provide our source code and a web server that runs the pipeline.

molecular biology↗

bioRxiv2023

The Interactome of DUX4 Reveals Multiple Activation Pathways

The ectopic expression of the DUX4 protein in muscle cells is the underlying cause of Facioscapulohumeral Muscular Dystrophy (FSHD). DUX4 is a potent transcription factor that activates a large number of genes in a dysregulated manner, but the direct protein interactions involved in this activation are only partially known. Here, we tailored an affinity purification and mass spectrometry (AP-MS) analysis to the unique features and functions of DUX4, to provide a more complete view of its interactome. We also obtained and verified all-atom models for two of the major interactions by employing cross-linking and mass spectrometry (CL-MS), computational modeling, and guided mutation studies. We find that DUX4 interacts strongly with two homologous transcription activators, PTOV1 and MED25, in addition to the previously characterized CBP/p300. The interaction with the PTOV1/MED25 domain involves the wrapping of the last thirty residues of the DUX4 activation region around the domain in a very extensive interface. Hence, DUX4 has the capacity to both open the chromatin and directly recruit the Mediator complex. DUX4 also binds to all members of the RFPL4 family, which are among the strongest genes it activates. These interactions are mediated through a hitherto unrecognized motif in the DUX4 disordered linker region. This feedback mechanism suggests that DUX4 may be inhibited by its own activation products, and explains its typical pulsed expression profile. We also found SIX1 and the AP-2 complex as strong DUX4 C-terminal interactors. A separate analysis of interactions involving the N-terminal of DUX4 revealed enrichment of proteins that are involved in DNA repair following double-strand breaks. Overall, these findings reveal new activation pathways for DUX4, which may be modulated in future strategies to control its toxicity. This study also showcases the synergy between CL-MS and deep-learning based modeling for the structural elucidation of challenging protein-protein interactions.

biochemistry↗

bioRxiv2022

Structure modeling and specificity of peptide-MHC class I interactions using geometric deep learning

Major Histocompatibility Complex (MHC) plays a major role in the adaptive immune response by recognizing foreign proteins through binding to their peptides. In humans alone there are several hundred different MHC alleles, where each allele binds a specific subset of peptides. The peptide-MHC complex on a cell surface is identified by a T-cell receptor (TCR) and this binding invokes an immune response. Therefore, predicting the binding specificity of peptide-MHC pairs is necessary for understanding the immune recognition mechanism. Here, we develop an end-to-end novel deep learning model, MHCfold, that consists of structure and specificity prediction modules for simultaneous modeling of peptide-MHC class I (pMHCI) complexes and prediction of their specificity based on their modeled structure. MHCfold produces highly accurate structures of pMHCI complexes with mean C RMSD of 0.98[A] and 1.50[A] for the MHC chain and the peptide, respectively. The binding specificity is also predicted with high accuracy (mean AUC of 0.94). Furthermore, the structure modeling component is orders of magnitudes faster than state-of-the-art methods (modeling of 100,000 pMHCI pairs in four hours on a standard computer), enabling high-throughput applications for large immunopeptidomics datasets. While peptide-MHC specificity can be accurately predicted from the sequence alone, TCR specificity prediction likely requires modeling of the 3D structures. We anticipate our model can be further used in structure-based prediction of TCR specificity. MHCfold is available @https://github.com/dina-lab3D/MHCfold

bioinformatics↗

bioRxiv2022

Reduced antigenicity of Omicron lowers host serologic response

SARS-CoV-2 Omicron variant of concern (VOC) contains fifteen mutations on the receptor binding domain (RBD), evading most neutralizing antibodies from vaccinated sera. Emerging evidence suggests that Omicron breakthrough cases are associated with substantially lower antibody titers than other VOC cases. However, the mechanism remains unclear. Here, using a novel geometric deep-learning model, we discovered that the antigenic profile of Omicron RBD is distinct from the prior VOCs, featuring reduced antigenicity in its remodeled receptor binding sites (RBS). To substantiate our deep-learning prediction, we immunized mice with different recombinant RBD variants and found that the Omicrons extensive mutations can lead to a drastically attenuated serologic response with limited neutralizing activity in vivo, while the T cell response remains potent. Analyses of serum cross-reactivity and competitive ELISA with epitope-specific nanobodies revealed that the antibody response to Omicron was reduced across RBD epitopes, including both the variable RBS and epitopes without any known VOC mutations. Moreover, computational modeling confirmed that the RBS is highly versatile with a capacity to further decrease antigenicity while retaining efficient receptor binding. Longitudinal analysis showed that this evolutionary trend of decrease in antigenicity was also found in hCoV229E, a common cold coronavirus that has been circulating in humans for decades. Thus, our study provided unprecedented insights into the reduced antibody titers associated with Omicron infection, revealed a possible trajectory of future viral evolution and may inform the vaccine development against future outbreaks.

immunology↗

bioRxiv2021

Super-immunity by broadly protective nanobodies to sarbecoviruses

Vaccine boosters and infection can facilitate the development of SARS-CoV-2 antibodies with improved potency and breadth. Here, we observed super-immunity in a camelid extensively immunized with the SARS-CoV-2 receptor-binding domain (RBD). We rapidly isolated a large repertoire of specific ultrahigh-affinity nanobodies that bind strongly to all known sarbecovirus clades using integrative proteomics. These pan-sarbecovirus nanobodies (psNbs) are highly effective against SARS-CoV and SARS-CoV-2 variants including the Omicron, with the best median neutralization potency at single-digit ng/ml. Structural determinations of 13 psNbs with the SARS-CoV-2 spike or RBD revealed five epitope classes, providing insights into the mechanisms and evolution of their broad activities. The highly evolved psNbs target small, flat, and flexible epitopes that contain over 75% of conserved RBD surface residues. Their potencies are strongly and negatively correlated with the distance of the epitopes to the receptor binding sites. A highly potent, inhalable and bispecific psNb (PiN-31) was developed. Our findings inform on the development of broadly protective vaccines and therapeutics. One sentence summarySuccessive immunization of SARS-CoV-2 RBD in a camelid enhanced the development of super-immunity and isolation and systematic characterization of a large repertoire of ultrahigh-affinity pan-sarbecovirus single-chain VHH antibodies to understand the evolution of this potent and broad immune response.

microbiology↗

bioRxiv2021

Structural Basis for Mis18 Complex Assembly: Implications for Centromere Maintenance

The centromere, defined by the enrichment of CENP-A (a Histone H3 variant) containing nucleosomes, is a specialised chromosomal locus that acts as a microtubule attachment site. To preserve centromere identity, CENP-A levels must be maintained through active CENP-A loading during the cell cycle. A central player mediating this process is the Mis18 complex (Mis18, Mis18y and Mis18BP1), which recruits the CENP-A specific chaperone HJURP to centromeres for CENP-A deposition. Here, using a multi-pronged approach, we characterise the structure of the Mis18 complex and show that multiple hetero- and homo-oligomeric interfaces facilitate the hetero-octameric Mis18 complex assembly composed of 4 Mis18, 2 Mis18y and 2 Mis18BP1. Evaluation of structure-guided/separation-of-function mutants reveals structural determinants essential for Mis18 complex assembly and centromere maintenance. Our results provide new mechanistic insights on centromere maintenance, highlighting that while Mis18 can associate with centromeres and deposit CENP-A independently of Mis18y, the latter is indispensable for the optimal level of CENP-A loading required for preserving the centromere identity.

cell biology↗

bioRxiv2021

ScanNet: An interpretable geometric deep learning model for structure-based protein binding site prediction

Predicting the functional sites of a protein from its structure, such as the binding sites of small molecules, other proteins or antibodies sheds light on its function in vivo. Currently, two classes of methods prevail: Machine Learning (ML) models built on top of handcrafted features and comparative modeling. They are respectively limited by the expressivity of the handcrafted features and the availability of similar proteins. Here, we introduce ScanNet, an end-to-end, interpretable geometric deep learning model that learns features directly from 3D structures. ScanNet builds representations of atoms and amino acids based on the spatio-chemical arrangement of their neighbors. We train ScanNet for detecting protein-protein and protein-antibody binding sites, demonstrate its accuracy - including for unseen protein folds - and interpret the filters learned. Finally, we predict epitopes of the SARS-CoV-2 spike protein, validating known antigenic regions and predicting previously uncharacterized ones. Overall, ScanNet is a versatile, powerful, and interpretable model suitable for functional site prediction tasks. A webserver for ScanNet is available from http://bioinfo3d.cs.tau.ac.il/ScanNet/

bioinformatics↗

bioRxiv2021

NanoNet: Rapid end-to-end nanobody modeling by deep learning at sub angstrom resolution

Antibodies are a rapidly growing class of therapeutics. Recently, single domain camelid VHH antibodies, and their recognition nanobody domain (Nb) appeared as a cost-effective highly stable alternative to full-length antibodies. There is a growing need for high-throughput epitope mapping based on accurate structural modeling of the variable domains that share a common fold and differ in the Complementarity Determining Regions (CDRs). We develop a deep learning end-to-end model, NanoNet, that given a sequence directly produces the 3D coordinates of the C[a] atoms of the entire VH domain. For the Nb test set, NanoNet achieves 1.7[A] overall average RMSD and 3.0[A] average RMSD for the most variable CDR3 loops. The accuracy for antibody VH domains is even higher: overall average RMSD < 1[A] and 2.2[A] RMSD for CDR3. NanoNet runtimes allow generation of ~1M nanobody structures in less than an hour on a standard CPU computer enabling high-throughput structure modeling.

bioinformatics↗

bioRxiv2021

Targeted in situ cross-linking mass spectrometry and integrative modeling reveal the architectures of Nsp1, Nsp2, and Nucleocapsid proteins from SARS-CoV-2

Atomic structures of several proteins from the coronavirus family are still partial or unavailable. A possible reason for this gap is the instability of these proteins outside of the cellular context, thereby prompting the use of in-cell approaches. In situ cross-linking and mass spectrometry (in situ CLMS) can provide information on the structures of such proteins as they occur in the intact cell. Here, we applied targeted in situ CLMS to structurally probe Nsp1, Nsp2, and Nucleocapsid (N) proteins from SARS-CoV-2, and obtained cross-link sets with an average density of one cross-link per twenty residues. We then employed integrative modeling that computationally combined the cross-linking data with domain structures to determine full-length atomic models. For the Nsp2, the cross-links report on a complex topology with long-range interactions. Integrative modeling with structural prediction of individual domains by the AlphaFold2 system allowed us to generate a single consistent all-atom model of the full-length Nsp2. The model reveals three putative metal binding sites, and suggests a role for Nsp2 in zinc regulation within the replication-transcription complex. For the N protein, we identified multiple intra- and inter-domain cross-links. Our integrative model of the N dimer demonstrates that it can accommodate three single RNA strands simultaneously, both stereochemically and electrostatically. For the Nsp1, cross-links with the 40S ribosome were highly consistent with recent cryo-EM structures. These results highlight the importance of cellular context for the structural probing of recalcitrant proteins and demonstrate the effectiveness of targeted in situ CLMS and integrative modeling.

biochemistry↗

bioRxiv2020

Integrative proteomics reveals exceptional diversity and versatility of mammalian humoral immunity

The humoral immune response is essential for the survival of mammals. However, we still lack a systematic understanding of the specific serologic antibody repertoire in response to an antigen. We developed a proteomic strategy to survey, at an unprecedented scale, the landscapes of antigen-engaged, serum-circulating repertoires of camelid heavy-chain antibodies (hcAbs). The sensitivity and robustness of this technology were validated using three antigens spanning orders of magnitude in immune response; thousands of divergent, high-affinity hcAb families were confidently identified and quantified. Using high-throughput structural modeling, cross-linking mass spectrometry, mutagenesis, and deep learning, we mapped and analyzed the epitopes of > 100,000 antigen-antibody complexes. Our results revealed a surprising diversity of high-affinity hcAbs for specific antigen binding on a variety of dominant epitopes. hcAbs perfect both shape and charge complementarity to target challenging antigens specifically; they can rapidly evolve to recognize a conserved, promiscuous cellular protein interaction interface, unraveling the convergent force that drives protein-protein interactions.

bioengineering↗

bioRxiv2020

A robust and versatile nanobody platform for drug delivery

Therapeutic and diagnostic efficacies of numerous small biomolecules and chemical compounds are hampered by the short half-lives. Here we report the development of a repertoire of diverse, high-affinity albumin-nanobodies (NbHSA) to facilitate drug delivery. By integrating biophysics, and hybrid structural biology, we have systematically characterized the NbHSA for albumin binding, mapped the epitopes, and resolved the architecture of a tetrameric Nb-albumin complex. We employed quantitative proteomics for accurate, multiplex Nb pharmacokinetic analysis. Using a humanized albumin mouse model, we found that the NbHSA has outstanding pharmacokinetics; the most stable NbHSA has a 771-fold T1/2 improvement compared with a control Nb. Interestingly, the pharmacokinetics of NbHSA is related to their biophysical and structural properties. To demonstrate the utility of NbHSA, we developed a highly stable NbHSA-hIL-2 cytokine conjugate "Duraleukin" and confirmed its improved anticancer properties than hIL-2 alone. We envision that this high-quality Nb resource will advance research into novel biotherapeutics.

bioengineering↗