bioRxiv Science⌕ Search

Biology subjects

Wu, L.

Publications and source records attributed to Wu, L..

At least 73 records · Page 4Linked to original sources
bioRxiv2025

Closing the Sim-to-Real Gap: An End-to-End Robotic Ultrasound System Leveraging In Vivo Reinforcement Learning and 3D-Prior Guided Hybrid Control

Abdominal ultrasound is a crucial first-line diagnostic tool, yet its efficacy is inherently constrained by a strong dependency on operator skill, leading to significant inter-operator variability and limiting its widespread adoption. While robotic ultrasound systems with artificial intelligence have emerged to mitigate this, current approaches face critical limitations: they often rely on simulated or phantom data for training, which hampers generalizability, and employ control strategies that lack the flexibility to adapt to complex human anatomy. To address these challenges, this paper introduces a novel, end-to-end robotic system for autonomous abdominal scanning. Our methodology is systematic: it begins with coarse scan planning via RGB and 3D point cloud-based localization of key anatomical landmarks (e.g., xiphoid process). The core of our approach involves training a reinforcement learning scanning policy directly on live human volunteers, enabling the development of a strategy that is robust to anatomical diversity. This high-level policy is executed by a sophisticated hybrid force-position controller, enhanced with real-time normal vector calibration and 3D point cloud-based respiratory phase detection to dynamically adjust contact force, particularly during inspiration. This innovation proves critical for improving scan quality in subjects with higher Body Mass Index (BMI). Extensive validation against expert sonographers demonstrates that our system achieves high precision in standard plane recognition and superior scanning efficiency. Furthermore, to holistically assess performance where standard planes are difficult to acquire, we introduce a 3D reconstruction-based coverage metric. Results confirm that our system delivers strong adaptability and significantly improved consistency, marking a substantial step toward clinically viable, operator-independent ultrasound robotics.

bioengineering↗
bioRxiv2025

Torsins Organize CLCC1 Assembly to Safeguard ER Bilayer and Lipid Homeostasis

The TMEM41B scramblase and its regulatory partner CLCC1 initiate lipid flux by equilibrating newly-synthesized phospholipids across the endoplasmic reticulum (ER) bilayer, a fundamental process required for diverse events ranging from membrane biogenesis to bulk lipid supply. Loss of CLCC1/TMEM41B causes ER bilayer imbalance, inducing giant ER-enclosed lipid droplets (geLDs) and driving rapid progression into severe MASH. Here we identify CLCC1 as the long-missing client of the lumenal Torsin ATPases, which selectively engage oligomerized CLCC1 at sites of ER bilayer imbalance. Hepatic TorsinA inactivation triggers geLD formation amid disrupted lipoprotein biogenesis and severe MASH, closely phenocopying CLCC1/TMEM41B deficiency. Mechanistically, Torsins act as foldases that drive CLCC1 oligomerization for its recruitment to imbalanced bilayers. Remarkably, ectopic CLCC1 expression reverses cellular and systemic lipid disorders arising from hepatic TorsinA deficiency. Hence, Torsin ATPases emerge as fundamental regulators that organizes CLCC1 and the downstream TMEM41B scramblase to govern lipid partitioning and membrane homeostasis.

cell biology↗
bioRxiv2025

Super-CUT&Tag: A Sensitive, Spatially Resolved Approach for Epigenomic Profiling in Tissue Sections

Spatial profiling of protein-DNA interactions is essential for understanding gene regulatory processes, but current methods suffer from low sensitivity, limited resolution, and complex workflows. Here we introduce Super-CUT&Tag (Solid-phase interface embedded with an Ultra-dense DNA barcode array to Profile Epigenomic landscapes for spatially Resolved CUT&Tag), a robust spatial epigenomic method that directly transfers in situ protein-DNA interactions from tissue sections onto a pre-fabricated spatially barcoded array. Incorporating 3D polyamidoamine (PAMAM) dendrimers markedly improves chromatin capture efficiency, achieving several-fold higher sensitivity compared to current spatial CUT&Tag approaches. Using Super-CUT&Tag, we profile H3K27ac spatial landscape and uncover previously hidden active-chromatin heterogeneity across mouse embryonic tissues and stages. Integration with spatial transcriptomics further reveals spatiotemporal enhancer regulation dynamics of Neurod2 during corticogenesis. By overcoming key bottlenecks in sensitivity, resolution and usability, Super-CUT&Tag establishes a powerful platform for spatial epigenomic profiling in complex tissues.

genetics↗
bioRxiv2025

RBM15 promotes lung adenocarcinoma progression and palbociclib sensitivity via m6A-mediated STIL activation and downstream cyclin D1/CDK4 signaling

RNA-binding motif protein 15 (RBM15), an N6-methyladenosine (m6A) methyltransferase, plays a crucial role in the progression of lung adenocarcinoma (LUAD); however, the underlying mechanisms remain insufficiently defined. In this study, we investigated the contribution of RBM15-mediated m6A modification to LUAD pathogenesis. Our data showed that RBM15 was highly expressed in LUAD tissues and cell lines. Notably, its expression levels were positively correlated with tumor TNM staging and negatively correlated with differentiation status. The overexpression of RBM15 enhanced cell proliferation, migration, invasion, and cell cycle progression, while simultaneously inhibiting apoptosis. Conversely, RBM15 knockdown reversed these phenotypes and induced G2/M-phase arrest. In vivo experiments further confirmed that RBM15 knockdown led to the suppression of tumor growth. RNA sequencing, MeRIP sequencing, and bioinformatics analyses identified STIL as a gene directly regulated by RBM15 through m6A modification, a finding subsequently validated by luciferase reporter assays. Western blot and rescue experiments demonstrated that RBM15 and STIL cooperatively activate the cyclin D1/CDK4 signaling pathway, thereby promoting LUAD progression. However, this effect was markedly attenuated by CDK4/6 inhibitor application. Collectively, these findings indicated that the RBM15-STIL-cyclin D1/CDK4 axis represents a promising therapeutic target for LUAD.

cancer biology↗
bioRxiv2025

A stress-dependent TRIM28-ALKBH2 feedback loop modulates chemotherapy response

Chemoresistance to DNA-damaging agents, including platinum and alkylating compounds, limits treatment efficacy in non-small cell lung cancer (NSCLC) and is frequently associated with elevated DNA repair activity. Here, we identify a stress-responsive feedback loop between the E3 ligase TRIM28 and the demethylase ALKBH2. TRIM28 binds ALKBH2 and promotes its K48-linked polyubiquitination and proteasomal degradation, whereas ALKBH2 enhances TRIM28 transcription and protein levels, forming a feedback loop. Notably, alkylation stress induces a biphasic response: acute MMS exposure enhances TRIM28-ALKBH2 association and accelerates ALKBH2 degradation, whereas prolonged exposure promotes TRIM28 degradation, leading to ALKBH2 stabilization and transcriptional upregulation. Clinically, ALKBH2 is frequently elevated in lung adenocarcinoma and is associated with worse survival, whereas TRIM28 exhibits prognostic value specifically in chemotherapy-treated patients. Functionally, MMS-cisplatin co-treatment increased DNA damage and reduced clonogenic survival by counteracting ALKBH2-dependent alkylation tolerance, linking the TRIM28-ALKBH2 loop to chemoresistance in NSCLC.

biochemistry↗
bioRxiv2025

A unified genetic perturbation language for human cellular programming

Evolution simultaneously and combinatorially explores complex genetic changes across perturbation classes, including gene knockouts, knockdowns, overexpression, and the creation of new genes from existing domains. Separate technologies are capable of genetic perturbations at scale in human cells, but these methods are largely mutually incompatible. Here we present CRISPR-All, a unified genetic perturbation language for programming of any major type of genetic perturbation simultaneously, in any combination, at genome scale, in primary human cells. This is enabled by a standardized molecular architecture for each major perturbation class, development of a functional syntax for combining arbitrary numbers of elements across classes, and linkage to unique single cell compatible barcodes. To facilitate use, CRISPR-All converts high level descriptions of desired complex genetic changes into a single DNA sequence that can rewire genomic programs within a cell. Using the CRISPR-All language allowed for head-to-head functional comparisons across perturbation types in a comprehensive analysis of all previously identified genetic enhancements of human CAR-T cells. Combining CRISPR-All programs with single cell RNA sequencing revealed a greater diversity of phenotypic states, including improved functional performance, only accessible through distinct perturbation classes. Finally, CRISPR-All combinatorial genome scale screening of up to four distinct perturbations simultaneously revealed additive functional improvements in human T cells accessible only through iterative multiplexing of modifications across perturbation classes. CRISPR-All enables exploration of a combinatorial genetic perturbation space, which may be impactful for biological and clinical applications.

genetics↗
bioRxiv2025

Single-cell Spatial Transcriptional Profiling Uncovers Heterogeneous Cellular Responses to Pathogenic Tau in a Mouse Model of Neurodegeneration

Intracellular tau accumulation contributes to neurodegeneration and systemic dysfunction. Although tau primarily aggregates in neurons as neurofibrillary tangles (NFTs), it also affects other cell types through poorly understood mechanisms. To define the molecular characteristics of tangle-bearing neurons and their influence on surrounding cells, we used CosMx SMI to evaluate the expression of 950 genes in more than 265,000 cells from rTg(tauP301L)4510 tauopathy and control mouse brains. In the cerebral cortex, tau pathology disrupted the excitatory-inhibitory neuron balance and altered pathways critical for myelination, similar myelin changes were observed in postmortem brain tissue from patients with progressive supranuclear palsy. In the hypothalamus, tau accumulation was associated with transcriptional changes linked to pathways involved in body weight and metabolic regulation. These findings reveal molecular mechanisms underlying both cell-autonomous and non-autonomous effects of tauopathy and identify targets for further investigation into tau pathogenesis and its systemic impact.

neuroscience↗
bioRxiv2025

Reconstructing the emergence of the human chorion via HIPPO-mediated trophoblast induction

The first lineage decision in the mammalian blastocyst commits outer cells to the trophectoderm and initiates the trajectory that gives rise to the placental chorion. The molecular sequence that unfolds downstream of HIPPO pathway inactivation, linking human trophectoderm specification to the early organization of the chorion, has remained unknown. Here, we establish a developmentally informed model that leverages HIPPO pathway modulation to induce the native trophectoderm trajectory in the absence of exogenous BMP or WNT signaling. We first transiently reset primed human pluripotent stem cells into a trophectoderm-competent ground state, followed by LATS kinase inhibition to set the trajectory in motion. To benchmark fidelity, we built an embryo-chorion single-cell reference integrating published early human and placental transcriptomes and applied a computational stage-matching tool to align our cultures to natural development. Stage matching revealed an ordered progression along the trophectoderm trajectory from early TE to post-implantation trophoblast. With extended culture, all major cell types of the nascent chorion emerged, encompassing both trophoblast and chorionic mesoderm lineages. Within the trophoblast, we identified proliferative and non-cycling villous cytotrophoblast, a columnar population connecting villous and extravillous domains, as well as syncytia and extravillous subtypes. When cultured in suspension, these lineages self-organized into three-dimensional organoids that recapitulated the stromal-epithelial architecture and proliferative-syncytial polarity of the emergent chorion. We identified CLDN6 as a defining surface marker of columnar trophoblast, the population that bridges villous and extravillous compartments. Prospective isolation of living CLDN6+ trophoblast revealed their capacity to reacquire a proliferative villous state and, under directed cues, generate both syncytial and extravillous fates, confirming their proposed dual developmental potential within the chorion. Together, these findings establish a developmentally informed framework that connects human trophectoderm specification to the emergent chorion and provides a dynamic platform for investigating the earliest steps of placental specification and the origins of implantation disorders.

cell biology↗
bioRxiv2025

Impact of N6-methyladenosine (m6A) machinery on HIV-1 replication in primary CD4+ T cells

N6-methyladenosine (m6A) is the most prevalent internal modification of cellular and viral RNA and is critical to the regulation of its localization, stability, and translation. Previous studies on the role of m6A during HIV-1 replication have produced conflicting results. Since m6A function can vary dramatically by cell type and state, here we aimed to clarify the role of the m6A machinery during HIV-1 replication in primary CD4+ T cells. Using CRISPR-Cas9 we targeted 46 cellular genes implicated in m6A or 5-methylcytosine (m5C) regulation and measured subsequent HIV-1 replication in primary CD4+ T cells. Only knockout of the m6A writer complex auxiliary proteins VIRMA and WTAP, and the m6A reader YTHDF2 were validated as significantly decreasing HIV-1 replication. In contrast, knockout of METTL3 or METTL14, which form the catalytic core of the writer complex, resulted in only marginal changes in HIV-1 infection, despite significant decreases in total cellular m6A levels. Chemical inhibition of METTL3 led to a dose-dependent decrease in HIV-1 infection, coupled with an increase in protein levels of METTL3 and other writer complex members. Expression of writer proteins was also co-dependent, revealing complex regulatory feedback mechanisms. Overall, these results clarify the role of epitranscriptomic machinery during HIV-1 replication in primary CD4+ T cells and suggest regulation by auxiliary members of the m6A writer complex is more influential than the function of the catalytic core itself on HIV-1 infection in primary CD4+ T cells. IMPORTANCEm6A is the most common chemical modification on cellular and viral RNA and regulates its stability, localization, and translation. m6A modification and its regulation varies dramatically between cell types and cell states. In this study, we investigated the role of m6A factors during HIV-1 infection of physiologically relevant primary CD4+ T cells. Using CRISPR-Cas9 to knockout 46 cellular genes implicated in RNA modification, we found only the m6A writer complex auxiliary members WTAP and VIRMA, and the reader YTHDF2, significantly affected HIV-1 replication in these cells. In contrast, knockout of METTL3 or METTL14, which form the catalytic core of the writer complex, resulted in marginal changes in HIV-1 infection, despite larger reductions in total cellular m6A levels. Our findings suggest regulation by auxiliary members of the m6A writer complex is more influential than the function of the catalytic core itself on HIV-1 infection in primary CD4+ T cells.

microbiology↗
bioRxiv2025

Oxidative stress triggers RNAPII arrest through PARylation and DNA damage

UV or gamma irradiation, as well as certain chemicals, generate DNA damage that disrupts transcription through a variety of well-characterised mechanisms. In contrast, the transcriptional response to oxidative stress remains poorly understood. Here, we describe a rapid and widespread shutdown of transcription following oxidative DNA base damage. By monitoring RNAPII occupancy and elongation dynamics, we demonstrate that oxidative stress temporarily halts RNAPII pause release and arrests the progression of elongation complexes within the gene body. We present evidence that this occurs in a unique and transient manner, characterised by abrupt arrest of elongating RNAPII dead in its tracks, followed by rapid transcriptional recovery as DNA lesions are repaired. We find that the restriction of initiation and early elongation complexes is regulated by PARylation, whereas recovery of RNAPII arrested within the gene body requires DNA repair mediated by the base excision repair (BER) and single-strand break repair (SSBR) pathways.

molecular biology↗
bioRxiv2025

Structure-Based Calculation of Excipient Effects on the Viscosity ofConcentrated Antibody Solutions

Computational prediction of the viscosity of therapeutic monoclonal antibodies (mAbs) at high concentration is highly desirable in the early discovery and development phases where the material needed for experimental determination is typically limited. Here, we present a unique coarse-grained (CG) simulation method that enables residue-level simulation of full-length antibodies with an elastic network, under simulated shearing force, to de novo predict viscosities of solutions of two distinct mAbs (an IgG1 and an IgG4), in the absence and presence of 6 excipients. Our results suggest the method can properly distinguish the viscosity profile of the two model mAbs, and directionally forecast viscosity change in response to added excipients. Furthermore, this CG modeling approach provides detailed protein-protein interaction mapping down to residue level contacts, including contact lifetimes and nature of interactions, illuminating microscopic insights into the underlying molecular interactions. It serves as a valuable tool for viscosity prediction, mechanistic insights, and mitigation strategies.

pharmacology and toxicology↗
bioRxiv2025

SAMHD1 promotes SARS-CoV-2 infection by enhancing HNF1-dependent ACE2 expression in lung epithelial cells

Sterile alpha motif and HD domain-containing protein 1 (SAMHD1) restricts a board spectrum of viruses through multifaceted mechanisms. It also limits spontaneous- and virus-induced innate immune responses by suppressing proinflammatory cytokine and type-I interferon (IFN-I) production. Some viruses escape SAMHD1 restriction by utilizing SAMHD1-mediated innate immune suppression to establish effective infection through viral antagonism. Our previous studies showed that SAMHD1 is a proviral factor facilitating replication of severe acute respiratory syndrome coronavirus (SARS-CoV-2) in human macrophages, monocytic THP-1 and epithelial-like HEK293T cell lines by suppressing IFN responses. However, it is unclear about the function of SAMHD1 in lung epithelial cells during SARS-CoV-2 infection. Here, we report that SAMHD1 facilitates SARS-CoV-2 replication in lung epithelial Calu-3 cells by enhancing endogenous expression of the viral receptor angiotensin-converting enzyme 2 (ACE2) via hepatocyte nuclear factor 1-alpha (HNF1) and HNF1{beta}. Using pseudotyped SARS-CoV-2 and lentiviral vectors, we found that SARS-CoV-2 spike protein-mediated viral entry was suppressed in Calu-3 cells with SAMHD1 knockout (KO). SAMHD1 KO repressed ACE2 expression in Calu-3 cells at mRNA and protein levels. Functional analyses revealed that HNF1 and HNF1{beta} were crucial for the endogenous ACE2 expression in Calu-3 cells. Additionally, SAMHD1 KO led to a reduction in the expression levels and ACE2-promoting function of HNF1 and HNF1{beta}. Inhibition of IFN antiviral response by baricitinib, a Janus kinase 1 and 2 (JAK 1/2) inhibitor, did not revert the suppression of SARS-CoV-2 in SAMHD1 KO Calu-3 cells. Our findings demonstrate that SAMHD1 facilitates HNF1-mediated ACE2 expression and SARS-CoV-2 replication in Calu-3 cells via a novel mechanism beyond its IFN-suppressive function. Author SummaryDuring viral infection, SAMHD1 acts as a viral restriction factor and a suppressor of the innate immune system, controlling viral replication while also ensuring immune homeostasis. The innate immune suppressive function of SAMHD1 can be proviral for some viruses. SAMHD1 has been shown to facilitate SARS-CoV-2 infection in human macrophages, THP-1 and HEK293T cell lines by promoting IFN antagonism, but its role in lung epithelial cells is unclear. Here, we demonstrated that SAMHD1 promotes SARS-CoV-2 replication in human lung epithelial Calu-3 cells by enhancing expression of the major viral receptor ACE2. We found that SAMHD1 facilitated HNF1-medaited ACE2 expression that was required for spike protein-mediated SARS-CoV-2 entry. However, inhibiting IFN signaling in SAMHD1 KO Calu-3 cells was not sufficient to revert SARS-CoV-2 replication. Our findings shed light on the differential proviral function of SAMHD1 in ACE2 expressing cells and suggest that SAMHD1 can facilitate SARS-CoV-2 infection beyond enhancing IFN antagonism.

microbiology↗
bioRxiv2025

DDHD2 possesses both lipase and transacylase capacities that remodel triglyceride acyl chains

Hereditary spastic paraplegia subtype SPG54 is a genetic neurological disorder caused by mutations in the DDHD2 gene. Excessive lipid droplet accumulation is observed in the brains of SPG54 patients and DDHD2 knockout mice, consistent with DDHD2s reported neutral lipase activity. Here, we find recombinant human DDHD2 preferentially hydrolyzes diacylglycerol (DAG) over phospholipids, with a slight preference for DAG over triacylglycerol (TAG). DDHD2 also exhibits transacylase activity, which enables transfer of acyl chains from triacylglycerols to diacylglycerols and monoacylglycerols to remodel the acyl chains of triglycerides. A predicted hydrophobic amphipathic helix on DDHD2 is essential for lipid droplet binding in vitro and in cells, and its lack reduces the enzymatic activity and triglyceride acyl chain remodeling. Adipose triglyceride lipase (ATGL), but not hormone sensitive lipase (HSL), also has transacylation activity and can remodel triglyceride acyl chains, but to a lesser extent than DDHD2. Taken together, this provides evidence that DDHD2 is a neutral lipid lipase and transacylase whose broad specificity enables triglyceride acyl-chain remodeling. SIGNIFICANCE STATEMENTTriglycerides (TAGs), the primary form of long-term energy storage, have acyl chain compositions crucial for diverse cellular processes. Lipases typically hydrolyze TAGs into free fatty acids. Here, we reveal a novel function for the neutral lipid lipase DDHD2: a transacylase activity. Instead of releasing fatty acids, DDHD2 transfers them between neutral lipids, altering TAG acyl chain composition. This transacylation requires the unique oil environment of lipid droplets (LDs), which excludes water from DDHD2s lipolytic active site, favoring transacylation over hydrolysis. DDHD2s lipase and transacylase activities enable TAG acyl-chain remodeling, demonstrating the possibility that a single enzyme can catalyze TAG cycling. This finding has implications for understanding lipid metabolism, LD dynamics, and specific motor neuron diseases implicating DDHD2.

biochemistry↗
bioRxiv2025

Decoder-FFPE-seq enables sensitive, genome-wide spatial transcriptomics of archival tissues at single-cell resolution

Formalin-fixed paraffin-embedded (FFPE) tissue-compatible spatial transcriptomics (ST) is transforming biology and histopathology, but achieving high sensitivity, genome-wide coverage, single-cell resolution and researcher-accessibility remains challenging. Here, we present a dendrimeric DNA coordinate-barcoding design for FFPE-adapted spatial RNA sequencing (Decoder-FFPE-seq), by integrating RNA-to-DNA probe conversion with blotting-assisted barcoding. Decoder-FFPE-seq provides single-cell resolution with a 2.54-4.25-fold sensitivity increase in mouse brain transcriptomic profiling over similar assays with an available protocol, successfully mapping rare genes and sparse cell types. It robustly resolves tumor microenvironments (TMEs) in clinical archives preserved 6-20 years. With expanded probe libraries, Decoder-FFPE-seq captures long non-coding and viral RNAs, revealing spatial co-expression networks with mRNAs in oral squamous cell carcinomas (OSCCs). In an OSCC immunotherapy cohort, it delineates TME heterogeneity across pathology response groups and tracks spatial cellular dynamics pre- and post-immuno-chemotherapy. Decoder-FFPE-seq establishes a robust, high-resolution framework for sensitive ST profiling in fundamental and translational studies of archival tissues.

genomics↗
bioRxiv2025

CryoDataBot: a pipeline to curate cryoEM datasets for AI-driven structural biology

Cryogenic electron microscopy (cryoEM) has revolutionized structural biology by enabling atomic-resolution visualization of biomacromolecules. To automate atomic model building from cryoEM maps, artificial intelligence (AI) methods have emerged as powerful tools. Although high-quality, task-specific datasets play a critical role in AI-based modeling, assembling such resources often requires considerable effort and domain expertise. We present CryoDataBot, an automated pipeline that addresses this gap. It streamlines data retrieval, preprocessing, and labeling, with fine-grained quality control and flexible customization, enabling efficient generation of robust datasets. CryoDataBots effectiveness is demonstrated through improved training efficiency in U-Net models and rapid, effective retraining of CryoREAD, a widely used RNA modeling tool. By simplifying the workflow and offering customizable quality control, CryoDataBot enables researchers to easily tailor dataset construction to the specific objectives of their models, while ensuring high data quality and reducing manual workload. This flexibility supports a wide range of applications in AI-driven structural biology.

molecular biology↗
bioRxiv2025

Widespread transcriptional memory shapes heritable states and functional heterogeneity in cancer and stem cells

Recent studies show that non-genetic cellular heterogeneity, particularly through heritable cell states, fundamentally shapes cancer evolution and developmental trajectories. However, conventional single-cell transcriptomic snapshots lack temporal information needed to identify these heritable states. Here, we employ lineage-resolved single-cell transcriptomics to systematically map heritable cell states that persist across divisions, distinguishing them from transient fluctuations within a cell cycle. We uncover that heritable states are underpinned by widespread transcriptional memory, whereby gene expression is heritable, defining two classes of states in cancer and stem cells: clustered states, characterized by clustered gene expression, and latent states, marked by non-clustered gene expression. This memory shows unexpected conservation across cell types and conditions and is maintained by robust epigenetic mechanisms resistant to environmental perturbations. Functionally, memory genes predict critical behaviors including metastatic potential and lineage commitment, with latent-state genes often outperforming clustered-state genes. Our findings establish transcriptional memory as the molecular basis of heritable cellular heterogeneity, providing a framework for broadly understanding functional cellular variations across biological systems.

systems biology↗
bioRxiv2025

Cross-ancestry information transfer framework improves protein abundance prediction and protein-trait association identification

Genetics-informed proteome-wide association studies (PWAS) provide an effective way to map the complex molecular landscape of biological mechanisms for complex diseases. PWAS relies on an ancestry-matched reference panel to model protein expression using genetic variants as features and determine the proteins impact on phenotype. However, reference panels from underrepresented populations remain relatively limited. In this study, we developed an analytic framework that borrows information from potentially multiple ancestries to boost the protein abundance prediction accuracy in an underrepresented population. We illustrate the frameworks utility and reproducibility through application to PWAS in East Asians: BioBank Japan (BBJ), Korean Genome and Epidemiology Study (KoGES), and Taiwan Biobank (TWB). An ensemble of information-sharing approaches was integrated to build the Multi-Ancestry-based Best-performing Model (MABM). MABM substantially improved the prediction performance with higher performance observed in both cross-validation and an external validation dataset (Tongji-Huaxi-Shuangliu Birth Cohort). Leveraging the BBJ, we identified three times as many significant PWAS associations with MABM as with the baseline Lasso model. Notably, 47.5% of the MABM specific associations were reproduced in independent East Asian datasets with concordant effect sizes. Furthermore, MABM enhanced gene/protein prioritization for downstream functional validation by (1) confirming a greater number of well-established gene/protein-trait associations and (2) identifying previously uncharacterized trait-associated genes. The benefits of MABM were further validated in additional ancestries and demonstrated in brain tissue-based PWAS, underscoring its broad applicability. Our findings close critical gaps in multi-omics research, develop a new reference resource of genetic models of protein abundance, and facilitate trait-relevant protein discovery in underrepresented populations.

genomics↗