Biology subjects

Peng, Z.

Publications and source records attributed to Peng, Z..

At least 19 recordsLinked to original sources

bioRxiv2026

RiboDFM: Robust RNA sequence design using backbone-conditioned discrete flow matching

Designing RNA sequences that fold into predefined 3D structures is central to RNA engineering, yet it remains challenging due to the scarcity of experimentally determined structures and the complex sequence-structure relationships. Here, we present RiboDFM, a backbone-conditioned discrete flow-matching framework for RNA sequence design. RiboDFM progressively resolves fully masked sequences by integrating a geometric backbone encoder, an auxiliary base-pairing prediction head, and language model representations into a discrete flow sampler. To expand our training data, we supplemented experimental structures with a large-scale collection of trRosettaRNA2-predicted models. Across diverse benchmarks, RiboDFM achieves high sequence recovery and generates sequences with greater tertiary-structure self-consistency than existing methods. RiboDFM also performs strongly on aptamer-focused design tasks. Finally, beyond sequence generation, the nucleotide probabilities derived from RiboDFM' s flow trajectory serve as an intrinsic measure of sequence-structure compatibility. This enables reliable confidence estimation for designed sequences, as well as reference-free, zero-shot assessment of predicted structure models. Together, these results establish RiboDFM as an effective framework for structure-based RNA sequence design.

bioinformatics↗

bioRxiv2026

A public-data reproducibility benchmark of a 2.5-D U-Net with a slice-level topology term for type-B aortic dissection segmentation

Background: Type-B aortic dissection requires separation of the true and false lumina in heterogeneous, anisotropic CTA. A topology term may reduce fragmentation, but its reproducibility must be tested under a fixed case split and independent native-grid evaluation. Methods: We used the public Graz type-B aortic dissection CTA collection (40 cases; CC BY 4.0) and froze a 28/6/6 case split for training, development, and held-out testing. A compact 2.5-D U-Net trained with cross-entropy plus multiclass Dice was compared with the identical model augmented by a fixed-weight (0.5) differentiable slice-level soft clDice term. Two training seeds were run for each arm. All 40 cases passed the NRRD audit; predictions for six held-out cases were restored to their native grids and scored with case-level 3-D voxel-grid Dice, voxel-grid morphological clDice, 6-connected component counts, physical-volume error, and 2,000-resample case-level paired bootstrap intervals. Results: With seed 20261007, topology-minus-baseline differences were -0.028 (95% CI -0.053 to -0.004) for true-lumen Dice, +0.032 (-0.006 to +0.096) for true-lumen voxel-grid clDice, +0.071 (+0.037 to +0.103) for false-lumen Dice, and +0.073 (+0.048 to +0.103) for false-lumen voxel-grid clDice. With seed 20261008, the corresponding differences were -0.040 (-0.065 to -0.018), +0.013 (-0.034 to +0.071), -0.099 (-0.187 to -0.004), and -0.058 (-0.124 to -0.001). Descriptive pooled means over 12 case rows per arm were true-lumen Dice 0.577 for topology versus 0.611 for baseline and false-lumen Dice 0.409 versus 0.422; these pooled rows are not independent cases. Conclusions: Under this fixed six-case test split, the corrected slice-level topology term did not produce a consistent benefit. True-lumen Dice was lower in both seeds, while false-lumen effects reversed direction between seeds; topology-clDice changes were small and/or interval-crossing. The benchmark supports an auditable failure-analysis workflow but does not support a general topology benefit, clinical use, external generalization, or a panvascular decision system.

bioinformatics↗

bioRxiv2026

trFlow: ultrafast all-atom protein conformational ensemble generation via conditional flow matching

Generating diverse, all-atom protein conformations remains challenging, especially for recovering alternative functional states and modeling conformational ensembles. We present trFlow, an ultrafast approach for generating all-atom conformations using geometry-guided flow matching. trFlow leverages three core innovations: (1) a hybrid single- and multi-step sampling scheme that balances accuracy and speed; (2) a decoupled representation-generation architecture that maximizes computational efficiency; and (3) fine-tuning on NMR ensembles to better capture conformational heterogeneity. Extensive benchmarking across dual-conformation proteins, NMR ensembles, and a CASP16 dynamic target shows that trFlow outperforms competing methods in recovering challenging states and reproducing experimental heterogeneity, while being an order of magnitude faster than existing all-atom approaches. These results establish trFlow as a practical and scalable framework for large-scale generation of all-atom conformational ensembles.

bioinformatics↗

bioRxiv2026

PanVasc Research for AI assisted evidence analysis in panvascular intervention

Panvascular intervention research requires evidence workflows that preserve source identity, outcome definitions and observation windows. We developed PanVasc Research, an executable research framework, and evaluated a fixed local Qwen3-4B model using two complementary tasks. Fifty ClinicalTrials.gov records from five vascular query strata generated 200 source-fidelity tests with intact evidence or controlled removal of the requested source, primary outcome or timeframe, plus 50 clean controls. A separate 100-statement sample from the official NLI4CT test set assessed clinical-trial entailment and evidence selection against existing expert labels. Generic and checklist prompts used identical evidence and maximum generation budgets; each output was also evaluated with an input-only deterministic contract. Registry exact accuracy was 51/200 (25.5%) with the generic prompt and 77/200 (38.5%) with the checklist; intact-case agreement was 50/50 and 48/50. The rule baseline recovered 200/200 tuples. NLI label accuracy was 48/100 (48.0%) and 50/100 (50.0%), respectively. The source contract retained 35 and 31 incorrect NLI labels in the two arms. Registry references establish fidelity to a registration snapshot, while NLI4CT concerns breast-cancer trials and does not validate vascular expertise. The framework also retains provenance-recorded literature retrieval, structured research planning and local numerical analysis. These experiments support a bounded assessment of source handling and semantic failure, rather than a new foundation model or autonomous scientific discovery. A proposed endpoint representation identifies the additional domain annotation and independent validation required for a panvascular research model.

bioinformatics↗

bioRxiv2026

Distinct mechanisms of neutralization by antibodies targeting a conserved pneumovirus F epitope

Pneumoviruses cause seasonal outbreaks leading to hospitalizations of vulnerable populations such as infants and the elderly. Cross-neutralizing antibodies targeting viral fusion have been isolated from infected individuals, but their elicitation and mechanisms of action remain understudied. Here, we describe two vaccine-elicited antibody classes, LOR24 and LOR69, that bind an overlapping epitope, and identify the somatic mutations that endow their breadth and potency, respectively. Cryo-electron microscopy structures of both antibodies bound to the HRSV fusion (F) protein show binding modes distinct from each other and from the previously described cross-neutralizing antibody MPE8, yet they all use similar motifs for binding. Complementary in vitro and electron microscopy experiments show that these antibodies either lock prefusion F as a trimer, arrest F in a monomeric or intermediate state, or promote the transition to the postfusion conformation. This work sheds light on mechanisms of pneumovirus neutralization and the elicitation of cross-neutralizing antibody responses.

immunology↗

bioRxiv2026

Profiling Siglec-7 and Siglec-9 ligands across the LuCaP PDX series: Implications for glyco-immune checkpoint inhibition in advanced prostate cancer

Advanced prostate cancer exhibits profound cellular and molecular heterogeneity, frequently becoming resistant to androgen receptor (AR) targeting through lineage plasticity and neuroendocrine differentiation. Immunotherapies have shown limited efficacy in prostate cancer, largely due to its immunosuppressive tumour microenvironment. Hypersialylation contributes to immune evasion by engaging sialic acid-binding immunoglobulin-like lectins (Siglecs) on immune cells, forming glyco-immune checkpoints. Although this pathway represents a promising therapeutic target, the distribution of Siglec ligands across diverse prostate cancer phenotypes and their response to standard-of-care hormone therapy remain poorly understood. Here, we utilised high-affinity engineered sialoglycan-binding reagents (HYDRA) to perform comprehensive immunohistochemical profiling of Siglec-7 and Siglec-9 ligands across a panel of 40 Washington Carcinoma of the Prostate (LuCaP) patient-derived xenograft (PDX) models. Ligand expression was evaluated in relation to AR status and neuroendocrine phenotype. To determine the impact of androgen deprivation on the tumour glycome, ligand expression was compared between matched PDX lines grown in intact and castrated mice. Our findings reveal widespread but heterogeneous expression of Siglec-7 and Siglec-9 ligands across the LuCaP cohort. Expression levels were comparable between AR-positive adenocarcinoma models and AR-negative neuroendocrine variants, demonstrating that this glyco-immune checkpoint is maintained across distinct prostate cancer lineages. Under castration conditions, glycan remodelling occurred in a model-dependent manner. A subset of PDX models exhibited reduced Siglec ligand expression following castration, suggesting partial AR dependence. In contrast, other models displayed increased ligand expression, consistent with adaptive immune evasion in response to therapeutic stress, while a third group remained largely unchanged. Collectively, our study demonstrates that the Siglec-7/9 glyco-immune checkpoint axis is broadly maintained across the spectrum of prostate cancer lineage plasticity but is dynamically remodelled by androgen deprivation in a patient-specific manner. These findings support the sialoglycan-Siglec axis as a lineage-independent immunotherapeutic target and suggest that strategies aimed at disrupting Siglec-mediated immune suppression, such as tumour desialylation, may be most effective when combined with androgen deprivation therapy to enhance anti-tumour immunity.

cancer biology↗

bioRxiv2026

Linoleic Acid-Lyso PG Axis promoting lipid droplet-mitochondria tethering by stabilizing Noncanonically Mitochondrial PPAR β/δ to Ameliorate Microglial Dysfunction in subarachnoid hemorrhage

Background Microglial lipid handling and mitochondrial failure contribute to brain injury after subarachnoid hemorrhage (SAH), but the lipid signals coupling these processes remain unclear. We investigated whether linoleic acid (LA) restores microglial homeostasis through lysophosphatidylglycerol 16:0 (LPG[16:0]) and peroxisome proliferator-activated receptor-{delta} (PPAR{delta}). Methods Cerebrospinal fluid metabolomics included 30 patients with aneurysmal SAH and 10 control participants. Mechanisms were examined in a blood-injection mouse model and hemoglobin-exposed primary mouse microglia using targeted lipidomics, RNA sequencing, mitochondrial and phagocytosis assays, pharmacological perturbation, fractionation, coimmunoprecipitation, thermal shift analysis, and structural modeling. Behavioral outcomes were evaluated by open-field, Y-maze, and Morris water-maze testing. Results; CSF LA was higher in SAH and discriminated the groups within this cohort (area under the curve, 0.9967 [95% CI, 0.9859-1.000]; P<0.001). LA attenuated inflammatory activation and restored phagocytosis, mitochondrial membrane potential, respiration, and ATP production in hemoglobin-exposed microglia. LA restored PLA2G15-associated LPG(16:0), which phenocopied these effects. Transcriptomic and inhibitor analyses identified PPAR{delta} as a downstream effector. LPG(16:0) increased PPAR{delta} stability, and fractionation and protease protection identified a PPAR{delta} pool on the cytosolic face of the outer mitochondrial membrane. PPAR{delta} associated with PLIN2 and CPT1A, promoted lipid droplet-mitochondria apposition, and supported fatty acid oxidation. In mice, LA reduced neuroinflammatory injury and partially improved anxiety-related behavior and spatial memory.

neuroscience↗

bioRxiv2026

Plasma titration provides a physical ruler for cross-platform proteomics

Plasma proteomics is expanding across platforms and cohorts, and integrating these data for AI demands comparability at the protein level, not merely concordant associations1,2. Affinity and MS platforms use distinct probes (antibodies, aptamers, or peptides) and signal readouts, yielding contradictory cross-platform results3,4. Without a known quantitative truth, we cannot distinguish biology from measurement distortion, leaving no gold standard for integration. Here we introduce Plasmix--a plasma reference suite with predefined male:female ratios (M, 1:0; Y, 3:1; P, 1:1; X, 1:3; F, 0:1)--and show that preserving this quantitative titration gradient, not just technical repeatability, predicts cross-platform concordance and identifies protein measurements suitable for integration. Profiling Plasmix across five platforms (Olink, SomaScan, NULISA, AAgAtlas, and MS-DIA) and 12 protocols across 17 batches, we found discordance is dominated by signal generation, not sample identity, and platforms distort signals in a protein-specific manner. Crucially, proteins retaining the titration response showed stronger agreement in an independent cohort; anchoring to the Plasmix midpoint (P) via sample-to-reference ratios reduced distortions, extending harmonizable coverage by 10-20%. Plasmix thus provides a physical ruler to benchmark accuracy, identifying genuinely integrable measurements before pooling datasets or training AI models--a critical bottleneck for plasma proteomics.

molecular biology↗

bioRxiv2026

Siglec-15 is a glyco-immune checkpoint in prostate cancer regulating immune evasion and metastasis

Prostate cancer is a leading cause of cancer-related mortality in men, and effective treatment options are limited for advanced and metastatic disease. The sialoglycan immune checkpoint Siglec-15 has emerged as a key mediator of tumour-associated immune suppression in several malignancies; however, its expression and functional role in prostate cancer remain poorly defined. Here, using dual immunofluorescence and immunohistochemistry, we demonstrate that Siglec-15 is expressed by prostate tumour epithelial cells, immunosuppressive macrophage phenotypes, and bone-resorbing osteoclasts within the tumour microenvironment. Mechanistically, we show that direct Siglec-15 receptor crosslinking, either by antibodies or tumour cell-derived conditioned medium, promotes monocyte-to-macrophage differentiation, generating macrophages with immunosuppressive and pathogenic phenotypes. Using therapeutic antibodies, we show that Siglec-15 blockade suppresses supernatant-induced monocyte to macrophage differentiation, allowing for the recovery of CD8 T-cell activation. Furthermore, we reveal that macrophage colony-stimulating factor (M-CSF) driven monocyte-derived macrophage differentiation is partially dependent on Siglec-15 signalling, with Siglec-15 blockade enhancing CD8 T-cell responses. In addition, anti-Siglec-15 treatment suppressed osteoclast differentiation, highlighting a dual role for Siglec-15 in prostate cancer immune suppression and bone remodelling. Consistent with these in vitro findings, therapeutic Siglec-15 blockade significantly reduced subcutaneous tumour growth in a CD8 T-cell-dependent manner and prolonged survival in a mouse model of prostate cancer metastasis. Together, these findings identify Siglec-15 as a central regulator of the prostate cancer glyco-immune axis, linking tumour-associated macrophage immune suppression with osteoclast-mediated bone remodelling, providing a compelling rationale for the clinical development of Siglec-15-targeted therapies for patients with advanced disease.

cancer biology↗

bioRxiv2026

Targeted proteomics of postmortem human brain reveals neurobiologic heterogeneity in Alzheimer's disease using NULISA technology

BackgroundAlzheimers disease (AD) is clinicopathologically heterogeneous. A proportion of patients living with AD present clinically at a younger onset of cognitive symptoms before 65 years old and/or non-amnestic clinical syndromes. Neuropathologically, corticolimbic distribution of neurofibrillary tangle pathology occurs on a continuum with some cases having greater cortical tau pathology relative to limbic regions and others with relatively restricted accumulation in limbic structures. These patterns of corticolimbic tangle distribution are associated with clinical presentation and age at onset. This study sought to examine protein expression differences across the spectrum of clinicopathologic heterogeneity using the NULISA targeted proteomics platform. MethodsA series of thirteen neuropathologically diagnosed AD cases from Mayo Clinic prospectively followed research studies were selected to reflect heterogeneity of clinical syndromes and corticolimbic distribution of tangle pathology. Frozen postmortem brain tissue samples were isolated from inferior parietal cortex and homogenized in RIPA buffer for analysis using Alamar Biosciences NULISA CNS disease 120 panel. Applying a conservative detection threshold of 75% level of detection for the novel application of NULISA in human brain, we evaluated levels of 69 of 129 protein targets across samples. We examined associations between age at onset cognitive symptoms and corticolimbic distribution of tangles (CLix) separately with individual protein targets using linear regression analysis. ResultsAD cases with a younger age at onset had higher measured levels of ubiquitin, while older age was associated with higher levels of total tau, CRH, and NPTX2. Investigations of corticolimbic heterogeneity revealed AD cases with lower CLix score (i.e., cortical predominant distribution of tau) had higher measured p-tau181, p-tau231, ubiquitin, and p62. AD cases with higher CLix (i.e., relative cortical sparing) had higher levels of total tau, CRH, NPTX2, MDH1, and HBA1. Brain-derived total tau consistently showed a stronger association in both models. ConclusionThis work demonstrates the utility of postmortem proteomics for investigating biomarkers associated with AD clinicopathologic heterogeneity. We observed proteomic differences in synapse integrity, tau post-translational modification, and ubiquitination associated with age at symptomatic onset and corticolimbic distribution of tangle pathology.

neuroscience↗

bioRxiv2026

Buffered sweat microfluidics with AI-enabled translation for clinically actionable blood urea estimation and renal risk stratification

Kidney-function assessment relies on blood urea as a clinically informative metabolic marker; however, its dependence on venipuncture and centralised laboratory testing limits high-frequency monitoring and delays timely clinical intervention. Here, we report an integrated platform combining a wearable buffered microfluidic patch with a physiology-informed, data-driven calibration framework for real-time, non-invasive estimation of blood urea from microlitre-scale sweat volumes (4.79 L). By precisely regulating the release kinetics of internal buffer salts, the device stabilises the local reaction microenvironment, mitigating variability in sweat pH and flow to ensure reproducible measurement. The resulting signals are processed through an artificial intelligence (AI)-enabled analysis pipeline that integrates sweat urea with patient-specific physiological information to generate clinically interpretable outputs. In multicentre studies, sweat urea shows a strong association with blood urea across diverse cohorts, but with nonlinear and time-lagged relationships that limit direct use. The AI-enabled calibration model compensates for these effects, enabling high-fidelity estimation of blood urea (r = 0.945 versus gold-standard measurements) at clinically relevant concordance levels. The platform further identifies kidney injury with 89.1% accuracy and stratifies disease severity with 83.2% accuracy. Notably, these results demonstrate that the integration of physicochemical stabilisation and AI-enabled data-driven translation establishes sweat as a clinically actionable surrogate for renal monitoring, supporting population-level estimation and highlighting the potential for personalised longitudinal assessment, and enabling a scalable, non-invasive strategy for high-frequency kidney disease management.

biophysics↗

bioRxiv2026

EnzOracle: Mechanism-aware prediction of enzyme environmental adaptation via a classification-guided mixture-of-experts framework

Industrial biocatalysis increasingly requires enzymes capable of operating under extreme physicochemical conditions, yet most natural sequence data reflect adaptation to mild environments, leading conventional predictive models to suffer from regression-to-the-mean effects in extremophilic regimes. Here we present EnzOracle, a classification-guided mixture-of-experts framework that enables distribution-aware prediction of enzyme melting temperature (Tm), optimal catalytic temperature (Topt), and optimal pH (pHopt) directly from sequence. EnzOracle demonstrated robust predictive accuracy across diverse benchmarks, achieving RMSE of 5.245{degrees}C for Tm, 11.458{degrees}C for Topt, and 0.781 for pHopt. Beyond predictive accuracy, we introduce a trait-resolved molecular simulation strategy to evaluate whether EnzOracle-derived attribution patterns correspond to independent physical mechanisms. Across representative systems, attention hotspots mapped onto rigidity-conferring interaction networks for Tm, dynamically preorganized active-site ensembles for Topt, and pH-dependent electrostatic and hydration networks for pHopt. These orthogonal validations indicate that EnzOracle captures transferable biophysical principles of enzyme environmental adaptation rather than merely exploiting dataset-specific correlations, positioning sequence-based learning as a mechanism-aware framework for discovering stability and activity determinants across diverse catalytic landscapes.

bioinformatics↗

bioRxiv2026

Expansion Revealing of Pathology Resolves Nanostructures Associated with Inflammatory Phenotypes in COVID-19 Decedent Human Brain Tissue

Expansion revealing (ExR) elucidates cellular organization by separating proteins within dense nanostructures by 20x linear expansion, but requires fixation procedures incompatible with human pathology specimens. Here, we report ExR of pathology (ExRPath), which attains [~]20 nm resolution and decrowding of such tissues, through iterative 20x expansion, adapted to human brain pathology specimens. We also report a single-shot 15x expansion protocol for such tissues (15ExMPath), achieved through one-shot 15x expansion. Applying ExRPath and 15ExMPath to COVID-19-decedent brain tissue reveals periodic amyloid nanoclusters that co-localize with SARS-CoV-2 in a rare minority of patient specimens, pointing to a potential neuroinflammatory phenotype associated with COVID-19, and highlighting the power of high-throughput nanoimaging, empowered by expansion microscopy, for discovering potential novel disease mechanisms.

neuroscience↗

bioRxiv2026

Comprehensive profiling reveals Sialyl-Tn upregulation and prognostic value in prostate cancer

Prostate cancer is a common cancer in males and there is an urgent unmet clinical need to identify new therapies for advanced disease. Aberrant glycosylation is common in prostate cancer and plays a functional role in disease progression. The sialyl-Tn antigen (sTn) has been widely studied in cancer, yet its involvement in prostate cancer remains relatively unexplored. Here, we utilise a novel anti-sTn antibody (L2A5) to comprehensively monitor sTn expression levels in clinical prostate cancer tissues encompassing normal, benign, primary, metastatic castrate-resistant prostate cancer (CRPC), and patient-derived xenografts (PDXs). We show that while sTn is detected at low or negligible levels in normal prostate tissues, it is expressed in 44% of prostate tumours, and prostate cancer patients with high sTn levels have significantly poorer survival times. Analysis of metastatic therapy resistant prostate-derived tumours growing in liver and bone, shows sTn is expressed in 37.5% of cases. Furthermore, we show sTn is expressed in nearly half of PDXs tested, supporting the use of PDX models as tools for testing anti-sTn therapeutic strategies. These findings identify sTn as potential prognostic biomarker and therapeutic target in prostate cancer and lay the groundwork for the development of sTn-targeted precision therapies for advanced disease.

cancer biology↗

bioRxiv2026

Benchmarking MSA pairing for protein-protein complex structure prediction reveals a depth-over-pairing principle

AlphaFold-Multimer (AFM) and AlphaFold3 (AF3) have revolutionized protein-protein complex structure prediction. However, it remains unclear whether elaborate MSA pairing is a strict prerequisite for achieving high accuracy. Here, we systematically evaluated diverse MSA construction strategies using a rigorous benchmark of 439 heterodimers. We found that incorporating paired MSAs yields only marginal improvements in structure prediction. Crucially, experiments using shuffled pairing patterns reveal that these gains stem not from specific pairing constraints, but rather from the increased depth of input MSAs, a finding that also extends to higher-order multimers. Furthermore, we demonstrate that prioritizing the inclusion of more homologs without pairing proves to be a superior strategy, enhancing performance for both intra- and inter-species dimers as well as the more challenging antibody-antigen complexes. This observation is substantiated by the recent success of the omicMSA strategy. Mechanistic analysis indicates that this capability is driven by inter-subunit physicochemical complementarity and the intrinsic architecture of the AF3 network. Furthermore, we identify critical bottlenecks that limit prediction accuracy, including large complex dimensions, restricted interface areas, and low experimental resolution. These findings establish a "depth-over-pairing" principle, offering new insights for improving protein complex prediction.

bioinformatics↗

bioRxiv2026

Comprehensive characterization of V(D)J recombination from long-read transcriptomic data with VDJcraft

V(D)J recombination generates antigen receptor diversity in developing B and T cells. Long-read transcriptome technologies (e.g., PacBio Iso-Seq, Nanopore RNA/cDNA) capture full-length transcripts and thus resolve V(D)J events more accurately than short-read platforms. However, existing short-read tools are not applicable to or optimized for long-read data. We developed VDJcraft, the first integrated pipeline designed for V(D)J recombination analysis using long-read transcriptome sequencing data. The workflow uses a two-pass alignment strategy: global alignment to the GENCODE reference with minimap2, followed by local realignment and annotation using the international ImMunoGeneTics information system (IMGT). A customized module enhances D-gene detection sensitivity and positional precision. Sequencing errors are reduced through consensus-based correction toward the predominant subclass. Antigen-binding regions are annotated using IMGT-defined motifs to characterize CDRs and binding site composition. VDJcraft was validated on simulated and Human Genome Structural Variation Consortium (HGSVC) datasets and applied to disease datasets. It accurately recovered full-length V(D)J-C sequences and outperformed existing methods in gene detection and recombination accuracy. Long-read calls also showed significantly higher concordance with high-confidence short-read calls (Mann-Whitney U test, p = 1.55 x 10-4). Additionally, we identified 31 putative novel gene subclasses absent from the IMGT database from HGSVC datasets. Analyses of longitudinal blood samples from a COVID-19 patient revealed distinct V(D)J recombination patterns and segment enrichment, characterized by increased IGHV1-2 usage, enrichment of the IGHV3-7/IGHD6-9/IGHJ5_02 rearranged clonotype, and a transient peak in IgG2 levels at day 4 followed by a gradual return to baseline. In conclusion, VDJcraft provides a robust framework for long-read V(D)J characterization and enables the discovery of disease-associated immune signatures.

bioinformatics↗

bioRxiv2026

AetherCell: A generative engine for virtual cell perturbation and in vivo drug discovery

Virtual cell modeling is currently hindered by a "data-utility paradox": biological information is fragmented between context-rich clinical RNA-seq and perturbation-dense experimental assays, leading to poor predictive generalization in human contexts. Here, we introduce AetherCell, a generative foundation model that unifies these disparate domains into a shared, platform-aligned transcriptomic manifold. By implementing a specificity-driven learning framework, AetherCell successfully recovers low-frequency, mechanism-specific signals often obscured by systematic noise. Across extensive benchmarks, AetherCell demonstrates robust generalization, accurately predicting responses to unseen compounds and genetic perturbations. We show that the model effectively translates signals from simple cell lines to complex 3D organoids, achieving high-fidelity, whole-transcriptome prediction. Building on this foundational manifold, we demonstrate precise drug response prediction across biological scales--including patient-derived organoids and clinical cohorts. We also implement a phenotype-knowledge mixture-of-experts strategy for precision drug repurposing. This approach is validated by the in vivo discovery of teriflunomide for dry eye disease and dabigatran for ulcerative colitis. Together, AetherCell establishes a scalable, human-centric virtual cell framework for translational biology and accelerated drug discovery.

bioinformatics↗

bioRxiv2026

Role of desolvation on biomolecular liquid-liquid phase separation

Biomolecular condensates play essential roles in cellular organization and are implicated in diverse pathological processes. Their formation is driven by liquid-liquid phase separation (LLPS), a process that requires coordinated multistep desolvation of biomolecular chains and multivalent inter-chain interactions. Although coarse-grained (CG) models with implicit solvent are widely used to probe LLPS thermodynamics and kinetics, they typically neglect water-mediated desolvation effects, limiting their accuracy and mechanistic interpretability. Here, guided by all-atom simulations and experimental measurements, we develop a desolvation-aware implicit-solvent CG model by incorporating residue-level desolvation terms directly into the pairwise energy function and apply it to investigate LLPS of intrinsically disordered proteins. Incorporating these desolvation interactions reshapes the phase diagram, alleviating dense-phase overcompaction. Notably, we observe an approximately linear correlation between the temperature gap (simulation temperature relative to the critical point) and the extent of conformational expansion accompanying the dilute-to-dense phase transition, a result further supported by theoretical analysis. We also find that desolvation barriers slow early density-fluctuation growth and shorten transient kinetic arrest, whereas solvent-separated contact interactions exert the opposite effects. Both terms further modulate chain mobility within mature condensates through competing packing and energy-landscape effects. Together, this framework enables an efficient representation of desolvation in CG simulations and reveals how desolvation energetics shape both the thermodynamic landscape and kinetic properties of biomolecular LLPS.

biophysics↗