Biology subjects

An, Z.

Publications and source records attributed to An, Z..

At least 19 recordsLinked to original sources

bioRxiv2026

A kinetic-aware approach to infer metabolic variations and flux using transcriptomics and metabolomics data

Assessing metabolic variations and flux quantities enable systematic understandings of metabolic shifts, reprogramming, adaptation and interactions in human diseases. However, omics-based estimation of metabolic flux and its variation remains challenging due to several fundamental limitations: the need for disease and tissue context specific metabolic model; nonlinear enzyme kinetic model that links enzyme and substrate changes to reaction flux; partial, unpaired, and snap-shot measurements across omics modalities; and uncertainty in computational prediction. Here, we present Michaelis-Menten model-based Flux Estimation Analysis (mmFEA), a Monte Carlo framework for estimating condition-specific flux changes by integrating paired or unpaired metabolomics and transcriptomics (or proteomics) data. mmFEA separates each reaction-rate change into enzyme- and substrate-associated components and assess reaction rate using Michaelis-Menten kinetics equation. A baseline metabolite saturation rate is introduced by integrating protein language model predicted kinetic parameters and human baseline level metabolic concentration to enable kinetic-aware integration of unpaired substrate and enzyme level measurements. Distribution of metabolic flux and variations between conditions are further computed using MCMC sampling by treating Michaelis-Menten-derived marginal flux distribution as prior and coherency in flux balance as likelihood. To benchmark mmFEA, we generated an in-house multi-omics data set including transcriptomics, metabolomics, metabolic activity functional assay, and CRISPR screening data using pancreatic cancer cell line system treated by APEX1 inhibitors. We demonstrated that mmFEA could accurately capture experimentally observed metabolic changes and achieved a better performance than all baseline methods. Our analysis revealed the necessity in using both substrate and enzyme modality and kinetic aware model in metabolic flux assessment. Further analysis using independent pancreatic cancer cohorts further validated the robustness of mmFEA, supporting integration of condition-linked unpaired data. Pan-cancer and spatial multi-omic applications demonstrated the use of mmFEA for resolving context-dependent metabolic variation when direct flux measurements are unavailable. Together, mmFEA provides a mechanistically grounded framework for estimating relative metabolic flux changes and their uncertainty from heterogeneous omics data.

systems biology↗

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

OCOO-T : A SIMPLE AND SCALABLE VIRTUAL CELL MODEL FOR TRANSCRIPTIONAL PERTURBATION RESPONSE PREDICTION

Predicting single-cell transcriptional responses to genetic, chemical and cytokine perturbations is a fundamental challenge in computational biology and AI Virtual Cell (AIVC) modeling, with direct implications for drug discovery and the elucidation of gene regulatory networks. Existing approaches often rely on auxiliary cell-state encoders, hierarchical variational autoencoders, dedicated Transformer encoder-decoder modules, or gene-interaction priors to compress high-dimensional expression profiles into latent representations. While effective, these designs increase architectural complexity and may limit scalability and generalizability. This paper introduces OCOO-T 1, a minimalist flow-matching-based AIVC model for transcriptional perturbation response prediction. OCOO-T utilizes a vanilla Transformer stack that operates directly on continuous gene expression profiles and formulates perturbation response prediction as a continuous-time denoising process. Perturbation embeddings, dosage information, and cell-line/cell-type specificity are integrated through adaptive layer normalization and in-context tokens. Comprehensive evaluations on Tahoe100M, Replogle, and PBMC benchmarks demonstrate that OCOO-T achieves state-of-the-art performance across diverse perturbations and cell types while effectively scaling to long transcriptional profiles through patching and depatching of cellular contexts. By leveraging the simplicity of Transformer-based denoising for single-cell omics, OCOO-T provides an effective and scalable framework for in-silico cellular simulation.

bioinformatics↗

bioRxiv2026

Accurate de novo design of peptides from programming biophysical landscape

Peptides regulate virtually every cellular process and emerge as transformative modalities in bioengineering and therapeutics. However, the de novo design of functional peptides remains challenging, as peptide-protein interactions are governed by delicate and dynamic biophysics that are difficult to model, and experimental datasets are scarce. Here we present NeoPep, a generative deep-learning framework that encodes biophysical principles to accurately design functional peptides de novo. By integrating over 5 million peptide-protein complexes spanning experimentally determined, sequence mimics, and structure ensembles, NeoPep learns the complex biophysical landscape governing peptide function and enables its programmable control. In prospective application across 10 diverse and challenging targets, NeoPep generated potent peptide binders, agonists, and antagonists with hit rates of 12.5-66.7%, even in the absence of defined binding sites or structure information. Beyond de novo co-design, NeoPep supports standalone structure or sequence redesign, readily discriminating subtle context differences. In the structure redesign mode, it generates highly selective peptides with atomic-level conformational accuracy (C RMSD < 2.0 [A] to our solved cryo-EM structure). This structural precision circumvents the need for experimental structure determination, further accelerating iterative sequence redesign to yield a 43.3-fold improvement in potency. These findings establish a general framework for translating biophysical principles into actionable peptide functions, with broad implications for basic research, bioengineering, and medicine.

molecular biology↗

bioRxiv2026

A genetic tool targeting brain GPR75

G-protein-coupled receptor 75 (GPR75) has emerged as an important mediator in diet-induced obesity (DIO) and a promising therapeutic target for anti-obesity drugs. However, the anatomical location of GPR75 in the brain remains unclear, hindering the understanding of GPR75 biology in DIO. Here, we generated a new GPR75-GFP-Ires-Cre knockin mouse strain, in which the Cre expression is driven by the endogenous GPR75 promoter and the GFP is fused with the C-terminal of the GPR75 protein. Both Cre and GFP were confirmed to be colocalized with the endogenous GPR75 expression. In addition, the GPR75-GFP fusion protein remains functionally normal with unaltered susceptibility to DIO. Moreover, using this mouse strain, we found that GPR75 is broadly expressed throughout the brain and mainly localized to the cytoplasm of brain neurons. This new genetic tool can therefore be used to study the neural basis for GPR75 in mediating DIO.

neuroscience↗

bioRxiv2026

Novel Muscle-Tropic AAV Capsids with Dramatically Enhanced Transduction and Safety Profiles in Non-Human Primates

Adeno-associated virus (AAV) gene therapy holds immense promise for treating muscular dystrophies, yet its efficacy and safety are constrained by the suboptimal tissue tropism of natural serotypes. Here, we employed the REACH platform, which combines rational design and directed evolution, to engineer muscle targeting vectors. Systemic administration in non-human primates (NHPs) revealed that lead candidate M1 mediates a >10-fold increase in skeletal muscle transduction compared to the AAV9 and 2-3 fold higher than MyoAAV, while concurrently achieving a remarkable 183-fold reduction in liver distribution. Furthermore, M1 exhibited significant de-targeting from key off-target tissues, including dorsal root ganglia (11 fold), lung (27 fold), spleen (2 fold), and kidney (2 fold). These findings demonstrate that the REACH platform can generate AAV capsids with simultaneously enhanced muscle tropism and favorable safety profiles, addressing a critical bottleneck in muscle-directed gene therapy.

bioengineering↗

bioRxiv2026

Identification of AAV Capsids with Enhanced Intravitreal Transduction and Favorable Safety Profile in Non-Human Primates

We report the discovery of novel adeno-associated virus (AAV) capsid variants engineered for superior intravitreal (IVT) gene delivery to the primate retina. Utilizing the REACH platform, we constructed a diverse AAV variant library and employed a multi-stage screening strategy involving in vitro selection on human retinal pigment cells followed by direct in vivo screening in non-human primates (NHPs). Following IVT administration in NHPS of a barcoded variant pool, next-generation sequencing analysis of retinal tissues identified lead candidates (e.g., E52, E54, and E57) that achieved transduction levels in the neural retina and RPE 5-10 fold higher than the benchmark R100. Concurrently, these high-potency variants exhibited an exceptional ocular confinement profile, with minimal to undetectable vector genome distribution in systemic organs. This combination of markedly enhanced retinal transduction and stringent local tropism establishes these engineered capsids as promising next-generation vectors for the treatment of inherited and acquired retinal diseases via a minimally invasive IVT route.

bioengineering↗

bioRxiv2026

Novel Engineered AAV Variants Demonstrate Superior Blood-Brain Barrier Penetration and Safety in Non-Human Primates

Systemic delivery of adeno-associated virus (AAV) for gene therapy of central nervous system (CNS) disorders is limited by inefficient blood-brain barrier (BBB) penetration and dose-limiting toxicity in peripheral organs, notably the liver and dorsal root ganglia (DRG)1-5. Here, we report the development of novel AAV variants via a proprietary capsid engineering platform (REACH). In non-human primates (NHPs), intravenous administration of lead variants resulted in transgene expression levels in the brain that were 600-2000 fold higher than AAV9 at the RNA level, concomitant with a 10-50 fold reduction in liver tropism and minimal off-target exposure in the heart and DRG. These engineered capsids achieve unprecedented, pan-CNS transduction with a markedly improved safety profile, representing a transformative platform for treating a broad spectrum of neurological diseases.

neuroscience↗

bioRxiv2026

LRP6-Guided Engineering of AAV9 Variants with Enhanced Blood-Brain Barrier Penetration and Reduced Liver Tropism in Non-Human Primates

The blood-brain barrier (BBB) severely restricts the delivery of systemically administered adeno-associated virus (AAV) vectors for central nervous system (CNS) gene therapy. To overcome this limitation, we engineered a library of AAV9 capsid variants through rational design focused on the low-density lipoprotein receptor-related protein 6 (LRP6), a conserved mediator of transcytosis. A multi-tiered screening strategy, encompassing human BBB endothelial cells followed by neuronal and glial target cells in vitro, identified three lead variants (QL9-21, QL9-22, and QL9-25) with markedly enhanced transduction potential. In mice, these variants achieved a 5-28 fold increase in brain-wide gene delivery compared to AAV9, without elevating hepatic tropism. Crucially, evaluation in non-human primates (NHPs) revealed that the lead variant, QL9-21, mediated a striking 3-40 fold enhancement in viral genome delivery across all examined brain regions versus AAV9, while concurrently reducing liver accumulation by 2.6 fold. Our study establishes an LRP6-guided engineering platform that yields novel AAV9 vectors capable of efficient, species-conserved BBB penetration coupled with a favorable safety profile, representing a significant advance toward clinically translatable CNS gene therapies.

neuroscience↗

bioRxiv2026

Somatic mutation in human cerebellum illustrates neuron type-specific patterns of age-related mutation

Human neurodegenerative disorders are characterized by exquisite specificity for neuronal types, but the basis of this is unknown. Here, we show that cerebellar granule neurons (GN)--the most abundant neuronal type in the human brain--accumulate somatic mutations in patterns highly distinct from cerebral cortical neurons, and more closely resembling oligodendroglia and other dividing cells. We find shared mutational signatures between normal aging GNs and medulloblastoma subtypes, suggesting the GN lineage as a tumor cell of origin. Whole-genome sequence of multiple single GNs from the same donor allowed analysis of specific times of neurogenesis, revealing a rich lineage tree that includes GNs that become postmitotic 2 years or more after birth, yet migrating postnatally to populate both the cerebellar vermis and the distant cerebellar hemisphere. Our results show that neuronal type-specific somatic mutation patterns enlighten normal development, cancer origins and potentially the cell type-specificity of neurodegeneration.

neuroscience↗

bioRxiv2026

Large-scale quantum computing framework enhances drug discovery in multiple stages

Coherent Ising machines (CIMs) excel at solving large-scale combinational optimization problems (COPs), but their insufficient long-term stability has hindered their applications in compute-intensive tasks like computer-aided drug discovery (CADD). By improving fiber vibration isolation and temperature control system, we have implemented a 2000-node CIM named QBoson-CPQC-3Gen achieving stable solutions over one hour on large-scale COPs. Graph-based encoding schemes were further introduced to realize a CIM-based CADD workflow including allosteric site detection, protein-peptide docking and intermolecular similarity calculation. CIM-based methods demonstrated superior speed and accuracy than heuristic algorithms. Especially, QBoson-CPQC-3Gen identified 2 novel druggable sites and bioactive compounds for 6 targets, which were further validated in vitro, in-cell and by crystal structures. Our contributions established a quantum-computing framework for multi-stage drug discovery, representing a significant advancement in both quantum computing applications and pharmaceutical research.

bioinformatics↗

bioRxiv2026

Molecular determinants of antibody-mediated priming to enhance detection of ctDNA

Liquid biopsies can enable cancer detection and monitoring yet remain limited by low concentrations of ctDNA. To address this limitation, we previously introduced a monoclonal antibody (mAb) "priming agent" that transiently increased the concentration of ctDNA in blood. Here, we investigated the molecular features that drive this effect. In a panel of novel mAbs that bound cfDNA, both those targeting dsDNA and those targeting mononucleosomes increased the concentration of ctDNA. mAbs with high avidity to dsDNA performed best, suggesting dsDNA as the key binding target. One agent preserved short, biologically informative molecules of cfDNA typically depleted at baseline. The Fc domain was dispensable, as F(ab)2 fragments retained priming activity. Leveraging these insights, we engineered single-chain agents using the dsDNA binding domain sso7d, expanding priming strategies beyond immunoglobulins. This study identifies the molecular features and general design principles for mAb-based priming agents to enhance recovery and detection of ctDNA. SignificanceLow concentration of ctDNA limits the sensitivity of liquid biopsies in many applications. This study shows how engineered antibodies and other dsDNA-binding molecules can inhibit clearance of ctDNA from the bloodstream and increase concentration of ctDNA in a blood draw, and identifies the key features that enable this activity. Modulation of cfDNA in the bloodstream using these agents can increase recovery of ctDNA and improve the performance of liquid biopsies for cancer detection.

bioengineering↗

bioRxiv2025

Fibulin-4 is highly expressed in metastatic breast cancer and can serve as a target of peptide-based imaging probes and experimental therapeutics

We have previously reported a cyclic peptide CRAGVGRGC (termed BLMP6) that homes to disseminating tumor cells in mouse cancer models and could be used for metastasis detection and intervention. Here, based on BLMP6 similarity to latent transforming growth factor beta binding protein 4 (LTBP4), we discovered fibulin-4 as a BLMP6 target. We show that BLMP6 mimics the LTBP4 domain binding to fibulin-4 and selectively binds to fibulin-4 in vitro. Fibulin-4 knockout in mouse 4T1 cancer cells abrogated BLMP6 homing to lung metastases. Fibulin-4 expression was found to be increased in invasive and metastatic human breast cancer. AZDye555 fluorophore-labeled BLMP6 was developed as a reagent selectively binding to invasive and metastatic human breast cancer cells in tissue sections and homing to MDA-MB-231 metastases in mice. We show that radiolabeling BLMP6 with 68Ga can be used for the detection of MDA-MB-231 metastases. We designed a peptide-drug conjugate consisting of monomethyl auristatin E (MMAE) and BLMP6 that preferentially kills aggressive cancer cells. Cytotoxicity of MMAE-BLMP6 against MDA-MB-231 tumors was confirmed in vivo. In an immunocompetent mouse model of B16F10 experimental lung metastases, treatment with MMAE-BLMP6 suppressed metastasis growth and improved survival. There was also a trend for metastasis suppression and survival improvement in the MDA-MB-231 experimental metastasis model. Our results suggest that fibulin-4 and BLMP6 may be further developed for the detection and targeting of metastatic human cancers. Statement of significanceThis study identifies fibulin-4 as a protein highly expressed in breast cancer metastasis. It evaluates the application of peptide conjugates targeting fibulin-4 in mouse models as non-invasive probes for metastasis detection and cytotoxic drug delivery.

cancer biology↗

bioRxiv2025

Miniaturized and accessible functional ultrasound imaging system for freely moving mice

Functional ultrasound (fUS) imaging provides brain-wide activity maps with high spatiotemporal resolution and deep penetration, positioning it as a key technology for future non-invasive brain-computer interfaces (BCIs). Realizing this potential, particularly for chronic BCI applications requiring long-term monitoring in naturalistic settings, critically depends on significant system miniaturization to overcome the cost and complexity limitations of current platforms. Addressing this challenge, we present mini-fUS, a miniaturized, cost-effective fUS platform engineered for accessibility without compromising core performance for demanding neuroscience research. The system features a compact transmit-receiver, low-noise power supply, and high-speed data transfer, achieving pulse repetition frequencies up to 5 kHz with negligible jitter. Real-time GPU-accelerated beamforming and singular value decomposition (SVD) enable whole-brain activity mapping, demonstrated here in freely moving mice at up to 3.57 Hz with [~]100 {micro}m spatial resolution and 15 mm penetration depth. Validated through recordings of brain activity during sensory stimulation, anesthesia, and behavior, this design defines a practical hardware-software framework for fUS. By significantly improving accessibility and enabling robust monitoring in mobile subjects, this work advances the development of fUS for both fundamental research and future BCI technologies, while clarifying essential fUS operational principles.

neuroscience↗

bioRxiv2025

Resolving the molecular niche of pulmonary fibrosis using cryopreserved human precision cut lung slices

Background and methodsPulmonary fibrosis (PF) manifests locally and heterogeneously in the diseased lung. To uncover mechanistic underpinnings, we have utilized cryopreserved fibrotic human precision-cut lung slices (PCLS), and using single-cell spatial transcriptomics, mapped and interpreted the transcriptomic regulation of effector cell types/states. ResultsCryopreserved-thawed fibrotic human PCLS retained canonical molecular and cellular PF hallmarks including fibroblast and myofibroblast accumulation, epithelial depletion, and vascular remodeling. Notably, these cellular changes were spatially co-incident with 1) collagen-rich regions enriched for extracellular matrix (ECM)-related genes (e.g., COL1A1, SPARC), 2) expanded fibroblast-endothelial crosstalk with enhanced chemokine signaling (CXCL12- CXCR4), 3) integrin-driven mesenchymal signaling targets (ANGPTL-integrin), 4) stress-adapted and immune-interacting Alveolar Type 2 cells with impaired surfactant gene expression, and 5) endothelial reprogramming toward a contractile and profibrotic phenotype. At the transcriptomic level, the spatially-defined fibrotic niche was characterized by increased gene expression related to immune responses (IGKC and IGHG1), fibrosis (collagen, fibronectin, and DCN), mechanotransduction (MYH11, TPM2, ACTG2, ACTA2), immune cell recruitment (CCL2, IL6, ICAM1), ECM remodeling (e.g. thrombospondin 1, THBS1), and decreased gene expression related to surfactant production (SFTPB, SFTPC). At the functional level, the fibrotic niches were defined by the activation of three fibroblast mediating signaling pathways: (1) the matricellular glycoprotein, SPARC, (2) Relaxin signaling, and (3) multiple immunoglobulin transcripts, including IGHA1, IGHG1, IGHG3, and IGKC. ConclusionsWe verify model fidelity, identify both known and unanticipated pro-fibrotic genes, and advocate for the expanded use of spatial transcriptomic measurements using cryopreserved human PCLS for disease modeling and drug discovery.

genomics↗

bioRxiv2025

Combining model-based and data-driven models: an application to synthetic biology resource competition

This work explores the integration of machine learning (ML) and mechanistic models (MM). While ML has demonstrated remarkable success in data-driven modeling across engineering, biology, and other scientific fields, MM remain essential for their interpretability and capacity to extrapolate beyond observed conditions based on established principles such as chemical kinetics and physiological processes. However, MM can be labor-intensive to construct and often rely on simplifying assumptions that may not fully capture real-world complexity. It is thus desirable to combine MM and ML approaches so as to enable more robust predictions, enhanced system insights, and improved handling of sparse or noisy data. A key challenge when doing so is ensuring that ML components do not disregard mechanistic information, potentially leading to overfitting or reduced interpretability. To address that challenge, this paper introduces the idea of Partially Uncertain Model Structures (PUMS) and investigates conditions that discourage the ML components from ignoring mechanistic constraints. This work also introduces the concept of embedded Physics-Informed Neural Networks (ePINNs), which consist of two loss-sharing neural networks that seamlessly blend ML and MM components. This work arose in the study of the context problem in synthetic biology. Engineered genetic circuits may exhibit unexpected behavior in living cells due to resource sharing. To illustrate the advantages of the ePINNs approach, this paper applies the framework to a gene network model subject to resource competition, demonstrating the effectiveness of this hybrid modeling approach in capturing complex system interactions while maintaining physical consistency.

systems biology↗

bioRxiv2024

Combined inhibition of S100A4 and TIGIT suppresses late-stage breast cancer metastasis to the lung by activating T and NK cells

Cancer metastasis is responsible for approximately 90% of cancer-related deaths, but very few treatment options exist currently. While the role of S100A4 in promoting metastasis has been known for decades, this knowledge has not been translated in the clinic. Here, we report that a novel monoclonal antibody against the human S100A4 protein (S1004-11) effectively suppresses breast cancer metastasis in two different mouse models. Importantly, a novel combination of ant-TIGIT and S100A4-11 can suppress lung metastases even in late-stage disease after the lung premetastatic niche (PMN) has already been established similar to a stage when many breast cancer patients are diagnosed. Mechanistically, S1004-11 mAb treatment block the formation of PMN by suppressing neutrophil infiltration and activating natural killer (NK) and T cells in the lung. Single-cell RNA-sequencing and cell:cell communication analyses indicate that TIGIT signaling suppresses NK cells, which is reversed in S100A4-11 treated PMN. In summary, this study provides compelling evidence for a novel mechanism of S100A4 function in PMN formation and the feasibility of using S100A4-11 monoclonal antibody to suppress metastases at different stages of breast cancer progression.

cancer biology↗

bioRxiv2024

Spatial organization of PI3K-PI(3,4,5)P3-AKT signaling by focal adhesions

The class I PI3K-AKT signaling pathway is the master regulator of cell survival, growth, and proliferation, and among the most frequently mutated pathways in cancer. However, where and how the PI3K-AKT signaling is spatially activated and organized in mammalian cells remain poorly understood. Here, we identified focal adhesions (FAs) as the subcellular signaling hubs organizing the activation of PI3K-PI(3,4,5)P3-AKT signaling in mammalian cells. We found that class IA PI3Ks are preferentially and dynamically recruited to FAs for activation, resulting in localized production of the critical signaling lipid PI(3,4,5)P3 around FAs. As the effector protein of PI(3,4,5)P3, AKT molecules are dynamically recruited around FAs for activation. Mechanistically, the spatial recruitment/activation of PI3K-PI(3,4,5)P3-AKT cascade are regulated by the activated FAK. Furthermore, combined inhibition of class I PI3K and FAK results in a more potent inhibitory effect on cancer cells. Thus, our results unveil a growth-factor independent, compartmentalized organization mechanism for PI3K-PI(3,4,5)P3-AKT signaling.

cell biology↗