Biology subjects

Becker, S.

Publications and source records attributed to Becker, S..

At least 19 recordsLinked to original sources

bioRxiv2026

Bilayer-bound structure of SARS-CoV-2 Envelope protein in complex with a host PDZ domain

Small multifunctional viral membrane proteins called viroporins are attractive drug targets but are difficult to structurally characterize in a lipid-bilayer environment. We have successfully reconstituted the full-length Envelope (E) protein from SARS-CoV-2 in a lipid-bilayer, co-assembled with a native host factor that binds to E, the PDZ domain 2 of ZO1. We have leveraged the enhancement afforded by dynamic nuclear polarization NMR to determine that the ensemble-averaged oligomer state of E is a pentamer. Complexing the protein with a host PDZ domain did not impact the oligomer state, but resulted in structuring of the last four residues of E, and immobilization of the PDZ domain near the lipid surface. Combined with secondary structure information derived from the chemical shifts, these data were used to model an E protein pentamer together with the PDZ protein dimer. Inclusion of a relevant host factor, as well as a complex lipid bilayer that mimics conditions in the ER-Golgi intermediate compartment, approaches native conditions, and addresses the ongoing controversy regarding the functional oligomer state in an infected cell.

biophysics↗

bioRxiv2026

Identification of a pan-orthoebolavirus-reactive antibody from an rVSV-EBOV vaccinated individual

Orthoebolaviruses such as Ebola virus (EBOV), Sudan virus (SUDV) and Bundibugyo virus (BDBV) can cause severe disease with high case-fatality rates. While licensed EBOV vaccines and therapeutic antibodies protect against EBOV infection, no single monoclonal antibody currently provides broad protection across multiple orthoebolaviruses. Here, we analyzed the humoral immune response of an rVSV-EBOV vaccinee to identify pan-orthoebolavirus-neutralizing antibodies. Using BDBV- and SUDV-glycoproteins for single B cell-sorting, we identified B10, which neutralized authentic EBOV and SUDV, with potent activity against SUDV compared with established cross-reactive antibodies. Structural analysis mapped antibody B10 binding to the pan-orthoebolavirus conserved GP2-stalk/HR2 region, associated with asymmetric trimer destabilization and spike opening. In vivo, B10 showed significant prophylactic efficacy in an EBOV mouse model and partial protection with antiviral activity in a SUDV mouse model. Together, these findings demonstrate that rVSV-EBOV vaccination induced the development of a broadly orthoebolavirus-neutralizing antibody that holds exeptional therapeutic potential.

immunology↗

bioRxiv2026

Multivalent Tau-Fyn interactions cooperatively arrest postsynaptic density condensate dynamics

Tau-Fyn signaling at postsynaptic sites is implicated in synaptic dysfunction and neuronal hyperexcitability in Alzheimers disease and related tauopathies, yet the molecular mechanisms by which Tau-Fyn interactions influence postsynaptic signaling assemblies remain poorly understood. Here, we reconstitute biomolecular condensates and membrane-associated assemblies that mimic the postsynaptic density (PSD) to investigate how Tau and Fyn regulate PSD organization and dynamics. We show that Tau and Fyn co-partition into PSD condensates and N-methyl D-aspartate receptor subtype 2B (NR2B)-associated PSD clusters, with both proteins preferentially enriching in the condensed phase relative to core PSD scaffold proteins. Enrichment of Fyn within PSD condensates is recapitulated by an SH3-SH2 fragment lacking the kinase domain, indicating that the kinase domain is dispensable for condensate partitioning. Fluorescence recovery after photobleaching reveals that Tau markedly reduces Fyn mobility and, together with Fyn, cooperatively drives a pronounced dynamic arrest of PSD condensate components, including the central scaffold protein PSD-95. Disruption of a major Tau-Fyn interaction hotspot by P216A/P219A mutations in Tau nearly completely restores the mobility of both Fyn and PSD-95, demonstrating that cooperative Tau-Fyn interactions are required for dynamic arrest. NMR spectroscopy identifies a multivalent interaction network between the Tau proline-rich region and the Fyn SH3 domain involving multiple binding motifs, with residues P216 and P219 acting as major interaction determinants. A 1.4 [A] crystal structure of the Fyn SH3 domain bound to a Tau peptide reveals a ternary 2:1 Fyn:Tau assembly, providing structural support for multivalent Tau- Fyn engagement. Together, our findings establish a condensate-based mechanism by which multivalent Tau-Fyn interactions cooperatively regulate PSD condensate dynamics and provide a molecular framework linking Tau-Fyn signaling to synaptic dysfunction in tauopathies.

biochemistry↗

bioRxiv2026

Amino acid Mutations at the Interface of Sudan Virus VP40 alter transport and budding

The matrix protein VP40 of orthoebolaviruses coordinates virion release and downregulates viral RNA synthesis through distinct oligomeric states, including dimers, octamers, and filamentous assemblies. To dissect the contributions of two oligomeric interface residues, L117 and W95, in the Sudan virus (SUDV) VP40 (sVP40), we created variants carrying alanine substitutions and assessed their structural and functional properties. sVP40 L117A failed to form dimers and was predominantly monomeric showing increased structural flexibility, reduced thermal stability together with loss of plasma membrane transport, budding activity, and the ability to regulate viral RNA synthesis. VP40 W95A preserved dimerization but also exhibited increased structural flexibility and reduced thermal stability. Functionally, sVP40 W95A more strongly inhibited viral RNA synthesis and markedly enhanced budding. However, in a transcription- and replication-competent virus-like particle (trVLP) assay, trVLPs produced with sVP40 W95A induced substantially reduced reporter activity in target cells, indicating impaired particle infectivity or functionality and suggesting possible defects in minigenome packaging, entry, or early post-entry steps. These results demonstrate that mutations at key oligomerization interfaces exert distinct structural and functional effects and highlight the requirement for precise oligomerization in coordinating sVP40s dual roles in genome regulation and virion release. By defining the contributions of L117 and W95, this study advances mechanistic understanding of sVP40 function and identifies processes that may serve as targets for antiviral intervention. ImportanceSudan virus (SUDV) causes regular outbreaks in Sub-Sahara Africa with unusually high lethality rates. However, in contrast to the more often occurring Zaire ebolavirus (EBOV), no monoclonal antibodies or vaccines are available and SUDV is generally understudied. The matrix protein VP40 is responsible for the downregulation of viral genome replication and transcription as well as budding. Here, we present structural and functional characterization of the SUDV VP40 interface residues L117 and W95 and show that while both amino acids are crucial for VP40s structural integrity, their functional effects are dramatically different ranging from complete abolishment to improving regulatory and budding activities.

microbiology↗

bioRxiv2026

Pericyte Loss Reprogrammes Capillary Endothelium and Drives White Matter Injury in Small Vessel Disease

Pericytes are critical regulators of cerebrovascular homeostasis, yet their contribution to small vessel disease (SVD) and white matter injury remains incompletely understood. Here, we use an inducible Atp13a5-driven genetic strategy to selectively label and ablate brain pericytes, enabling integrated morphological, functional, and transcriptomic analyses across the neurogliovascular unit. Single-cell RNA sequencing and tissue-level mapping identified distinct pericyte subtypes distributed along the vascular tree and revealed subtype-specific vulnerability following depletion. Moderate pericyte loss induced transient cerebrovascular dysfunction characterised by reduced cerebral blood flow and increased blood-brain barrier (BBB) permeability, accompanied by delayed white matter abnormalities, including altered diffusion MRI metrics, oligodendrocyte progenitor cell responses, and myelin loss. At the molecular level, pericyte depletion reprogrammed capillary endothelial cells toward an activated venular-like state characterised by VCAM-1 induction, reduced expression of the BBB-associated transporter MFSD2A, and activation of type I interferon signalling pathways. Cross-species analyses revealed enrichment of human white matter hyperintensity-associated SVD gene signatures across endothelial subtypes following pericyte depletion. Together, these findings identify pericyte dysfunction as a driver of endothelial inflammatory remodelling and white matter injury and establish mechanistic links between microvascular pericyte loss and human SVD.

neuroscience↗

bioRxiv2026

A Heart Disease-Associated TSPO Variant Alters Transmembrane Helix Dynamics

The 18-kDa translocator protein TSPO is an outer mitochondrial membrane protein involved in cholesterol transport, stress response, and cellular metabolism. Although its five-helix architecture is conserved across species, the structural and dynamic features of human TSPO remain unresolved. Using solution NMR spectroscopy, we characterize the conformational dynamics of human TSPO in complex with a third-generation diagnostic ligand. We identify a dynamic N-terminal segment of the first transmembrane helix that does not form a stably populated helix but instead defines a flexible boundary between the cytosolic region and the transmembrane core. The common disease-associated A14V variant reduces this conformational heterogeneity by introducing short-range contacts and redistributing backbone dynamics across the protein. These changes preserve the overall fold while locally stabilizing the N-terminal transmembrane helix toward the VDAC interaction interface. Our findings reveal a human-specific dynamic architecture of TSPO and link variant-induced stabilization to modulation of transmembrane helix dynamics.

biophysics↗

bioRxiv2026

Lethal Sudan virus infection in IFNAR-/- mice reveals hallmarks of a cytokine storm

Sudan virus (SUDV) is a member of the family Filoviridae, which comprises highly pathogenic viruses associated with unusually high case fatality rates. The development of medical countermeasures against filoviruses, including antivirals, vaccines, and therapeutic antibodies, requires preclinical evaluation in suitable animal models. C57BL/6J IFNAR-/- mice, which lack the type I interferon (IFN-/{beta}) receptor, have been reported to be susceptible to filovirus infections, although their impaired innate immune response may represent a potential limitation of the model. Here, we show that IFNAR-/- mice constitute a suitable model for SUDV infection. Following infection, animals developed a clear clinical disease characterized by significant weight loss and pronounced changes in behaviour and appearance. Mice reached the predefined clinical endpoint 3-5 days post infection. Post mortem analysis of terminal samples revealed high viral loads and viral genome copies in all tested organs as well as in serum, indicating widespread systemic dissemination. Importantly, infection was associated with a marked increase in several key chemokines and cytokines linked to systemic inflammation, consistent with the development of a cytokine storm-like response. Together, these findings demonstrate that SUDV infection in IFNAR-/- mice induces systemic viral dissemination and a pronounced inflammatory response, supporting the suitability of this model for investigating filovirus pathogenesis and infection-associated immune dysregulation.

microbiology↗

bioRxiv2026

Pluripotency Factors Modulate Interferon Signaling in Embryonic Stem Cells

Despite lacking a robust interferon response, pluripotent stem cells remain highly resistant to viral infection, in part through the constitutive expression of immune genes traditionally classified as interferon-stimulated genes. While interferon signaling has been shown to be incompatible with the maintenance of pluripotency, the molecular mechanisms underlying this relationship remain poorly understood. Here, we investigate the transcriptional response of human embryonic stem cells (hESCs) to infection with a potent activator of the interferon response, an influenza A virus mutant lacking the viral NS1 protein. Single-cell RNA sequencing revealed that while most hESCs remain unresponsive to infection, a distinct subpopulation expresses type I and III interferons. Notably, only interferon-expressing cells mounted a robust antiviral response, characterized by strong induction of interferon-stimulated genes. In contrast to the bulk hESC population, interferon responding cells exhibited reduced expression of core pluripotency factors as well as negative regulators of interferon signaling, such as SOCS1 and SPRY4. Depletion of SOCS1 enabled hESCs to respond robustly to interferon stimulation, showing that this negative regulator is a key suppressor of interferon signaling in pluripotent stem cells. We further show that SOCS1 and additional negative regulators of IFN signaling are intrinsically expressed in hESCs and are transcriptionally controlled by pluripotency factors, such as NANOG, SOX2 and OCT4. Together, our findings support a model in which pluripotency factors regulate intrinsic immune gene expression, including negative regulators of interferon signaling, thereby suppressing canonical interferon signaling to preserve pluripotency while maintaining antiviral resistance. IMPORTANCEBy combining single-cell transcriptomics with functional studies, we demonstrate that the pluripotency transcriptional program active in pluripotent stem cells coordinately regulates pluripotency factors, antiviral genes, and negative regulators of interferon signaling. This integrated control enables pluripotent stem cells to achieve effective protection against viral infection while preserving their differentiation potential, providing new insights into how innate immunity is selectively constrained in pluripotent stem cells. These findings have important implications for stem cell-based therapies, where transient modulation of antiviral responses without disrupting pluripotency could improve therapeutic efficacy. More broadly, this work advances our understanding of interferon regulation that could inform the development of antiviral strategies that enhance protective immune responses while limiting harmful or unwanted inflammatory signaling.

immunology↗

bioRxiv2026

Alzheimer's Disease Brain Organoids as a Source of Disease Relevant Amyloid-Beta Oligomers

Amyloid plaques are a hallmark of Alzheimers disease (AD) progression; however, the early stages of plaque formation and the specific amyloid-beta (A{beta}) species involved remain difficult to study. While post-mortem tissue provides insight into end-stage mature plaques, therapeutic development requires targeting the earliest A{beta} oligomers to arrest plaque formation. Furthermore, inherently toxic soluble A{beta} oligomers off-pathway from plaque formation are implicated as a driving force of AD pathology. It also remains unclear if the specific nature of key disease-relevant species can be accurately replicated in preparations of synthetic peptides.. To bridge this gap, we utilize brain organoids carrying AD mutations as a biologically authentic source for A{beta} peptides and oligomers. We demonstrate that these mutations do not disrupt organoid development and that the resulting conditioned media contains A{beta} oligomers with disease-relevant structures. Finally, we show that these oligomers can be concentrated and segregated via differential ultracentrifugation for further experimental applications.

neuroscience↗

bioRxiv2026

Peripheral Nociceptor Activity and Placebo Hypoalgesia: A Proof-of-Concept Study

Placebo effects illustrate clearly how cognitive states can influence physiological processes. While central mechanisms of placebo hypoalgesia, i.e. reduce sensitivity to painful stimuli due to expectations, are well established, it remains unclear whether such expectations modulate also peripheral activity. In this proof-of-principle study, we combined a well-established placebo paradigm with microneurography to directly assess responses of single nerve fibers in awake humans. Six healthy participants (18 fibers) underwent each a placebo and a control condition in a randomized crossover design. Placebo hypoalgesia was induced using verbal suggestion and classical conditioning. Manipulation checks confirmed that participants expected hypoalgesic effects of a placebo gel after conditioning. At the peripheral level, data on the recovery cycle of the nerve fibers indicated changes in the excitability of sleeping nociceptors (CMi) in the placebo compared to control condition, namely with enhanced and longer-lasting early subnormality. While other microneurographic outcome measures showed no consistent placebo effects, rather strong time-dependent reductions in fiber responsiveness independent of the experimental conditions were observed. These findings provide preliminary evidence that modulation of peripheral nociceptor activity by cognitive states, specifically expectations, is possible. Such modulation of peripheral activity appears to be fiber-type specific. Time effects have been neglected in microneurographic protocols. The pronounced time effects found here highlight the need for optimized such protocols.

neuroscience↗

bioRxiv2026

The SynMuvA lin-15A licenses natural transdifferentiation by antagonizing identity safeguarding mechanisms

The mechanisms that restrict or enable latent cellular plasticity have attracted growing interest over the past decade, with important implications for cancer and regenerative therapies. However, the diversity of both pro- and anti-plasticity mechanisms remains incompletely understood. Here, we identify the THAP domain gene lin-15A as a novel factor involved in the natural rectal-to-neuronal Y-to-PDA transdifferentiation in Caenorhabditis elegans. We found that, unlike previously described essential factors, lin-15A is not a Driver of transdifferentiation. Instead, it antagonizes several chromatin-modifying complexes known to safeguard differentiated cell identities. We also show that lin-15A is not a core plasticity factor per se but acts as one specifically in the Y cell context. Together, our findings support a model in which diverse molecular players coordinate controlled cell identity conversions: plasticity factors function as Drivers, while others like lin-15A which we propose to term Licensers attenuate identity safeguarding mechanisms, thereby facilitating transdifferentiation.

developmental biology↗

bioRxiv2026

Towards understanding nitrate-reducing Fe(II) oxidation in Ferrigenium straubiae: spotlight on old and new key protein candidates

3.Autotrophic Fe(II) oxidation is a key microbial process linking iron cycling to global carbon and nitrogen turnover in anoxic and low-oxygen environments. While Fe(II) oxidation, and electron utilization by canonical denitrification or oxygen-reducing pathways have been proposed, the whole electron transport chain via the periplasm is unknown. To address this gap, we compared the transcriptome of Ferrigenium straubiae strain KS (= KCTC 25982; = DSM 118991) cultivated under autotrophic conditions using Fe(II) as the electron donor and either nitrate or oxygen as the terminal electron acceptor. Additionally, we reanalyzed the genome using bioinformatic tools (FeGenie, MHCScan, Foldseek, AlphaFold, COFACTOR, and InterProScan) to classify known, putative, and previously uncharacterized proteins. Using these analyses, we (i) identified a form II RubisCO-mediated, sedoheptulose-7-phosphate-forming transaldolase variant of the Calvin-Benson-Bassham cycle, (ii) revealed the metabolic possibility of a reversed oxidative TCA cycle and an alternative CO2 fixation route via 10-formyl-tetrahydrofolate, (iii) found novel structural features of the proposed Fe(II) oxidases (Cyc2 and MtoA/B: {beta}-barrel outer membrane cytochromes), (iv) uncovered a nitric oxide reductase (eNOR-family) and two redox complexes specifically upregulated under nitrate-reducing Fe(II)-oxidizing conditions, including one comprising a sphaeroides heme protein homolog, and (v) reveal a cytochrome c rich gene area encoding ten heme proteins and (iv) proposed a conserved nonaheme cytochrome c to link Fe(II) oxidation to quinone reduction. 2. RepositoriesGenBank whole genome sequence of strain KS: JAGRPI00000000 IMG Taxon ID of strain KS: 2878407288 NCBI BioProject ID of RNA-Seq data: PRJNA1399691 Strain availability: KCTC 25982 and DSM 118991 Normalized transcripts and Log2FoldChanges (Excel file), and supplementary information are available with the online version of this article.

microbiology↗

bioRxiv2025

A single-cell cytokine dictionary of human peripheral blood

Cytokines orchestrate immune responses, yet we still lack a comprehensive understanding of their specific effects across human immune cells due to their pleiotropy, context dependence and extensive functional redundancy. Here, we present a Human Cytokine Dictionary, created from high-resolution single-cell transcriptomes of 9,697,974 human peripheral blood mononuclear cells (PBMC) from 12 donors stimulated in vitro with 90 different cytokines. We describe donor-specific response variation and uncover robust consensus cytokine signatures across individuals. We then delineate similarities between cytokine response profiles, and derive cytokine-induced immune programs that organize responsive genes into data-driven, biologically interpretable functional modules. By integrating cell type-specific responses with expression of cytokines, we infer higher-order cell-to-cell and cytokine-to-cytokine communication networks exemplified by an IL-32-{beta}-initiated signaling cascade, which rewires myeloid programs by inducing neutrophil-recruiting factors while suppressing Th1-responses and promoting IL-10-family cytokines. Finally, we show how the Human Cytokine Dictionary enables the interpretation of cytokine-driven immune responses in other studies and disease contexts, including systemic lupus erythematosus, multiple sclerosis, and non-small cell lung carcinoma. Together, the Human Cytokine Dictionary constitutes the first comprehensive cell type-resolved transcriptional screen of human cytokine responses and provides an essential open-access, easy-to-use community resource with accompanying software package to advance our understanding of cytokine biology in human disease and guide therapeutic discovery.

immunology↗

bioRxiv2025

Cohesin Facilitates Nucleosome Invasion by Transcription Factors

Nucleosomes present a major barrier to transcription factor (TF) binding. However, a subgroup of TFs known as pioneer factors (PFs) can recognize motifs covered by nucleosomes and initiate chromatin opening. PFs also bind nucleosomal substrates with high affinity in vitro, which may facilitate their nucleosome invasion in vivo. Here, we show that LexA, a bacterial TF with poor nucleosome binding in vitro, can rapidly invade into a well-positioned nucleosome from a motif positioned at the dyad when expressed ectopically in yeast. This striking contrast between LexA binding in vitro and in vivo raises the possibility that TFs can exploit nucleosome dynamics in vivo to access occluded sites. Surprisingly, we find that LexA-mediated chromatin opening can occur in the absence of DNA replication, chromatin remodeling, histone turnover, and a few histone chaperones. Instead, nucleosome invasion by LexA and a native PF, Cbf1, is promoted by the cohesin complex, illustrating intriguing a connection between cohesin and TF binding. Together, our results demonstrate that even non-pioneer TFs like LexA can bind and displace nucleosomes in vivo, through a process facilitated by cohesin.

molecular biology↗

bioRxiv2025

Exploration of RNA outside segmented cells in spatial transcriptomics reveals extrasomatic RNA organization

Image-based spatial transcriptomics (iST) enables visualization of RNA molecules in their spatial context, yet up to 40% of transcripts remains unassigned to cells and has been largely overlooked. In this study, we systematically analyze unassigned RNAs (uRNAs) across 14 public iST datasets and multiple technologies to characterize their nature and relevance in tissue biology. By assessing potential technical origins, in particular segmentation errors, noise, and diffusion across many tissues in both humans and mice, we find that around one third of uRNAs cannot be attributed to technical artifacts. Those non-technical uRNAs are enriched around cells with complex morphologies such as neurons, glia, and endothelial cells and reflect transcripts localized in cellular protrusions and extrasomatic compartments. Using these signals, we infer protrusion-associated transcript localization and identify cell-cell contacts beyond standard cell-centric segmentation. Our results challenge the assumption that uRNA is purely technical noise and instead highlight its potential biological relevance, particularly in relation to intracellular RNA localization and tissue architecture. To enable their systematic study, we introduce troutpy, a Python package for quantitative uRNA exploration in spatial transcriptomics data.

bioinformatics↗

bioRxiv2025

Reward-induced endogenous pain inhibition scales with action-outcome certainty in humans

Endogenous pain modulation is thought to encompass a crucial evolutionary purpose in guiding decision-making away from harm. This is well exemplified by the idea that mechanisms of learning through error-correction and endogenous pain modulation are inherently intertwined. However, adaptive behavior requires more than learning through error-correction. Biological environments are volatile, which can cause decision-makers to be uncertain about what actions lead to rewards or punishments. Evidence on how uncertainty in action-outcome distributions impacts endogenous pain modulation is lacking. In the present study, we extend and adapt a well-established paradigm for the study of endogenous pain modulation with the implementation of a reversal learning outcome schedule. Thirty healthy human volunteers took part in this probabilistic gambling task, where they had to gamble for the obtainment of pain relief and the avoidance of pain punishments. Using a computational approach, we show that decision-making in a situation of acute pain is best explained by models that account for uncertainty. Such uncertainty is associated with the observed pain modulation in the task, indicating that endogenous pain modulation may be sensitive to volatility and the perception of uncertainty. Specifically, pain inhibition from winning pain relief increases as a function of certainty in what actions lead to pain relief. Our findings emphasize the importance of considering mechanisms of uncertainty processing in reinforcement learning from painful outcomes and endogenous pain modulation. Those mechanisms could be relevant to understanding behavioral changes in chronic pain, where altered reinforcement learning has already been established.

neuroscience↗

bioRxiv2025

HM-DyadCap - Capture and Mapping of 5-Hydroxymethyl-cytosine/5-Methylcytosine CpG Dyads in Mammalian DNA

5-Methylcytosine (mC) and 5-hydroxymethylcytosine (hmC) are the main epigenetic modifications of mammalian DNA, and play crucial roles in cell differentiation, development, and tumorigenesis. Both modifications co-exist with unmodified cytosine in palindromic CpG dyads in different symmetric and asymmetric combinations across the two DNA strands, each having unique regulatory potential. To facilitate investigating the individual functions of such dyad modifications, we report HM-DyadCap. This method employs an evolved methyl-CpG-binding domain (MECP2 HM) for the direct capture and sequencing of DNA fragments containing the CpG dyad hmC/mC. Binding studies reveal a high discrimination of MECP2 HM against off-target dinucleotides. We conduct comparative mapping experiments for mESC genomes with HM-DyadCap, standard MethylCap employing wild type MECP2, as well as MeDIP and hMeDIP protocols. We find that MECP2 HM is blocked by hmC glucosylation, and conduct control enrichments with glucosylated genomes that indicate highly selective enrichment of hmC/mC dyads by MECP2 HM. Metagene profiles correlate hmC/mC marks with actively transcribed genes, and reveal global enrichment in gene bodies as well as depletion at transcription start sites. We anticipate that HM-DyadCap will enable effective enrichment and mapping of hmC/mC marks with broad applicability for unravelling the function of this dyad in chromatin biology and cancer.

genomics↗

bioRxiv2025

scConcept: Contrastive pretraining for technology-agnostic single-cell representations beyond reconstruction

Recent large-scale single-cell foundation models have shown promise for exploring cellular states, yet they often underperform compared to simpler, domain-specific methods, raising concerns about their broader applicability. A key limitation lies in their reliance on masked language modeling, which is well suited for generative language tasks but poorly aligned with learning rich cell-level embeddings required in single-cell biology. Moreover, the proliferation of transcriptomic technologies--from whole transcriptome dissociated assays to image-based targeted profiling--poses a major challenge for cross-platform generalization. Here, we align with recent advances in machine learning to move beyond reconstruction metrics, which often do not capture important sample variation. We present scConcept ("contrastive cell pre-training"), a transformer-based contrastive learning framework that directly optimizes cell embeddings by contrasting multiple views of cells. By replacing gene-level reconstruction with a cell-level identification task, scConcept learns robust representations that are invariant to count distributions and gene panel selection, across diverse assays and technologies. To highlight the capability of the proposed framework, we pretrain scConcept on a similar corpus of over 30 million single-cell RNA-seq profiles as recent foundation models. Our approach demonstrates superior performance not only compared to state-of-the-art pretrained foundation models but also domain-specific methods in various downstream tasks, including cell-type annotation, technology integration, dissociated to spatial cell-type transfer, spatial imputation, gene panel optimization, and mapping new technologies on already existing atlases. Our results highlight contrastive pretraining as a powerful alternative to reconstruction-based strategies for single-cell modeling, providing a path toward general-purpose, technology-agnostic cell representations.

bioinformatics↗