Biology subjects

Wu, W.

Publications and source records attributed to Wu, W..

At least 127 records · Page 7Linked to original sources

bioRxiv2024

scGNN+: Adapting ChatGPT for Seamless Tutorial and Code Optimization

Foundation models have transformed AI by leveraging large-scale data to efficiently perform diverse tasks, and their applications in bioinformatics are primarily focused on data-centric tasks like cell type annotation and gene expression analysis. However, their potential extends beyond data analysis, offering significant opportunities in software development and optimization, such as code refinement, tutorial generation, and advanced visualization. For example, models like OpenAI Codex can suggest optimized code snippets and generate well-documented, reproducible workflows, enhancing accessibility and reducing computational complexity. Despite these advantages, the use of foundation models for improving computational tool engineering in single-cell research remains underutilized. To address this gap, we developed scGNN+, a web-based platform that combines the power of graph neural networks with the capabilities of ChatGPT to enhance reproducibility, code optimization, and visualization. scGNN+ further simplifies the process for users by generating standardized, well-annotated code, making complex procedures more accessible to non-programmers. Additionally, ChatGPT integration allows users to create high-quality, customizable visualizations through natural language prompts, improving data interpretation and presentation. Ultimately, scGNN+ offers a user-friendly, reproducible, and optimized solution for single-cell research, leveraging the full potential of foundation models in bioinformatics. scGNN+ is publicly available at https://bmblx.bmi.osumc.edu/scgnn+.

bioinformatics↗

bioRxiv2024

SC-MAMBA2: Leveraging State-Space Models for Efficient Single-Cell Ultra-Long Transcriptome Modeling

AO_SCPLOWBSTRACTC_SCPLOWThe rapid advancement of single-cell sequencing technology has significantly deepened our understanding of cellular heterogeneity, yet it concurrently presents substantial challenges for the unified modeling of single-cell data. Simultaneously, pre-trained foundation models have achieved notable success in domains such as natural language processing and image analysis. However, extending these models to accommodate ultra-long single-cell transcriptome sequences, characterized by an extensive number of genes, remains a formidable task. In this study, we introduce SC-MAMBA2, based on the MAMBA2 architecture, meticulously designed with a bidirectional modeling approach tailored for single-cell transcriptomics data. As the first single-cell foundation model to integrate state-space models (SSMs) underlying MAMBA2 architecture, SC-MAMBA2 features over 625 million parameters, covers more than 60,000 genes, and was pre-trained on a dataset of over 57 million cells, making it the most comprehensive solution for processing ultra-long transcriptome sequences. Extensive bench-marking across a diverse array of downstream tasks consistently demonstrates that SC-MAMBA2 surpasses state-of-the-art models, delivering superior accuracy and enhanced computational efficiency.

cell biology↗

bioRxiv2024

Bimodal cell mass distribution separates CD8+ T cells into two distinct types with divergentdifferentiation dynamics

T cells are central to immune defense, yet existing molecular and phenotypic assays do not fully capture a cells intrinsic immune potential. Here we show that a single physical property, buoyant mass, reveals hidden heterogeneity within phenotypically similar, resting CD8+ T cells. Using suspended microchannel resonator measurements, we identify two distinct populations: "light" cells, enriched for mitochondrial content but prone to delayed activation and exhaustion, and "heavy" cells, biosynthetically poised for proliferation and memory formation. In patients with melanoma receiving immune checkpoint blockade, pre-treatment buoyant mass profiling of circulating T cells predicted therapeutic response with an accuracy comparable with standard tumor-derived biomarkers. Our findings establish buoyant mass as a label-free, stimulation-independent measure of systemic T cell fitness, providing a rapid and broadly applicable framework for immune profiling and response prediction in cancer and beyond.

biophysics↗

bioRxiv2024

Transcription elongation factor ELOF1 is required for efficient somatic hypermutation and class switch recombination

Somatic hypermutation (SHM) and class switch recombination (CSR) diversify immunoglobulin (Ig) genes and are initiated by the activation induced deaminase (AID), a single-stranded DNA cytidine deaminase that is thought to engage its substrate in the context of RNA polymerase II (RNAPII) transcription. Through a loss of function genetic screen, we identified numerous potential factors involved in SHM including ELOF1, a component of the RNAPII elongation complex that has been shown to function in DNA repair and transcription elongation. Loss of ELOF1 strongly compromises SHM, CSR, and AID targeting and alters RNAPII transcription by reducing RNAPII pausing downstream of transcription start sites and levels of serine 5 but not serine 2 phosphorylated RNAPII throughout transcribed genes. ELOF1 must bind to RNAPII to be a proximity partner for AID and to function in SHM and CSR. We propose that ELOF1 helps create the appropriate stalled RNAPII substrate on which AID acts. HighlightsO_LIA CRISPR knockout screen has identified numerous potential SHM factors. C_LIO_LISHM, CSR, and AID targeting are strongly compromised in the absence of ELOF1. C_LIO_LIELOF1 must interact with RNAPII to be an AID proximity partner and support AID targeting. C_LIO_LIELOF1 supports RNAPII pausing and generation of the substrate for AID action. C_LI

immunology↗

bioRxiv2024

Stronger premicrosaccadic sensitivity enhancement for dark contrasts in the primate superior colliculus

Microsaccades are associated with enhanced visual perception and neural sensitivity right before their onset, and this has implications for interpreting experiments involving the covert allocation of peripheral spatial attention. However, the detailed properties of premicrosaccadic enhancement are not fully known. Here we investigated how such enhancement in the superior colliculus depends on luminance polarity. Rhesus macaque monkeys fixated a small spot while we presented either dark or bright image patches of different contrasts within the recorded neurons response fields. Besides replicating premicrosaccadic enhancement of visual sensitivity, we observed stronger enhancement for dark contrasts. This was especially true at moderate contrast levels (such as 10-20%), and it occurred independent of an individual neurons preference for either darks or brights. On the other hand, postmicrosaccadic visual sensitivity suppression was similar for either luminance polarity. Our results reveal an intriguing asymmetry in the properties of perimicrosaccadic modulations of superior colliculus visual neural sensitivity.

neuroscience↗

bioRxiv2024

High-quality peptide evidence for annotating non-canonical open reading frames as human proteins

A major scientific drive is to characterize the protein-coding genome as it provides the primary basis for the study of human health. But the fundamental question remains: what has been missed in prior genomic analyses? Over the past decade, the translation of non-canonical open reading frames (ncORFs) has been observed across human cell types and disease states, with major implications for proteomics, genomics, and clinical science. However, the impact of ncORFs has been limited by the absence of a large-scale understanding of their contribution to the human proteome. Here, we report the collaborative efforts of stakeholders in proteomics, immunopeptidomics, Ribo-seq ORF discovery, and gene annotation, to produce a consensus landscape of protein-level evidence for ncORFs. We show that at least 25% of a set of 7,264 ncORFs give rise to translated gene products, yielding over 3,000 peptides in a pan-proteome analysis encompassing 3.8 billion mass spectra from 95,520 experiments. With these data, we developed an annotation framework for ncORFs and created public tools for researchers through GENCODE and PeptideAtlas. This work will provide a platform to advance ncORF-derived proteins in biomedical discovery and, beyond humans, diverse animals and plants where ncORFs are similarly observed.

molecular biology↗

bioRxiv2024

Tree reconstruction guarantees from CRISPR-Cas9 lineage tracing data using Neighbor-Joining

CRISPR-Cas9 based lineage tracing technologies have enabled the reconstruction of single-cell phylogenies from transcriptional readouts. However, developing tree-reconstruction algorithms with theoretical guarantees in this setting is challenging. In this work, we derive a reconstruction algorithm with theoretical guarantees using Neighbor-Joining (NJ) on distances that are moment-matched to estimate the true tree distances. We develop a series of tools to analyze this algorithm and prove its theoretical guarantees. When the parameters of the data generating process are known and there is no missing data, our results align with established results from common evolutionary models, such as Cavender-Farris-Neyman and Jukes-Cantor. However, to account for the realistic case where the parameters of the data generating process are not known and there is missing data, we develop new theory that shows for the first time that it is still possible to obtain reconstruction guarantees in the CRISPR-Cas9 case and in other models of evolution. Empirically, we show on both simulated lineage tracing data and on real data from a mouse model of lung cancer the improved performance of our method as compared to the traditional use of NJ.

bioinformatics↗

bioRxiv2024

Motion correction with subspace-based self-navigation for combined angiography, perfusion and structural imaging

Motion artifacts are problematic in many MRI modalities. "Self-navigating" approaches are desirable, since no additional scan time or hardware is required. However, the generation of a navigator image, to estimate and correct motion, is difficult in cases where the tissue contrast is changing during the navigator acquisition window, such as in magnetization- prepared methods. Here we propose a subspace approach to reconstruct accurate navigators in the presence of time-varying tissue contrast and apply it to a combined angiography, perfusion and structural imaging method using a golden ratio 3D cones trajectory. This arterial spin labeling-based pulse sequence relies on subtraction of label and control images to isolate the relatively weak blood signal, making it particularly susceptible to motion corruption. An inversion pulse leads to time-varying tissue contrast across the readout train, but by reconstructing subspace coefficient maps directly, artifacts due to the varying contrast were alleviated. This resulted in high-quality navigator images that were subsequently registered to estimate and correct for motion. In addition, a split-update method was proposed to efficiently reconstruct from mismatched label/control k-space data with locally low rank regularization enforced on the difference image. The correction process was tested with numerical simulation and in vivo data from 8 healthy subjects with and without cued motion. In numerical simulation, the subspace- based navigator achieved an 84% reduction in RMSE of residual motion compared to without motion correction. In vivo, motion correction resulted in noise-like and background artifacts being greatly reduced and vessel sharpness being noticeably improved. Correlation of angiography, perfusion and structural images with motion-free reference images also increased by 159%, 53% and 12%, respectively, after motion correction. These results show that subspace-based navigators can effectively improve the motion robustness of MR imaging in contrast-varying acquisitions.

neuroscience↗

bioRxiv2024

A Hippocampal-parietal Network for Reference Frame Coordination

Navigating space and forming memories based on spatial experience are crucial for survival, including storing memories in an allocentric (map-like) framework and conversion into body-centered action. The hippocampus and parietal cortex (PC) comprise a network for coordinating these reference frames, though the mechanism remains unclear. We used a task requiring remembering previous spatial locations to make correct future action and observed that hippocampus can encode the allocentric place, while PC encodes upcoming actions and relays this to hippocampus. Transformation from location to action unfolds gradually, with Came From signals diminishing and future action representations strengthening. PC sometimes encodes previous spatial locations in a route-based reference frame and conveys this to hippocampus. The signal for the future location appears first in PC, and then in hippocampus, in the form of an egocentric direction of future goal locations, suggesting egocentric encoding recently observed in hippocampus may originate in PC (or another "upstream" structure). Bidirectional signaling suggests a coordinated mechanism for integrating map-like, route-centered, and person-centered spatial reference frames at the network level during navigation.

neuroscience↗

bioRxiv2024

A single-molecule nanopore sequencing platform

Nanopore sequencing, a third-generation sequencing technology, has revolutionized the gene sequencing industry with its advantages of long reads, fast speed, real-time sequencing and analysis, and potential in detecting base modifications. This technology allows researchers to sequence longer DNA fragments in a single read, providing more comprehensive genomic information compared to previous methods. Nanopore sequencing operates on electrical signals generated by a nanopore embedded in a membrane separating two electrolyte-filled chambers. When single-stranded DNA (ssDNA) passes through the nanopore, it creates variations in the current that correspond to different DNA bases. By analyzing these current fluctuations with machine learning algorithms, the DNA sequence can be determined. In this study, we introduced several improvements to nanopore sequencing, including nanopore local chemistry sequencing, novel motor and pore proteins, chip design, and basecalling algorithms. Our new nanopore sequencing platform, CycloneSEQ, demonstrated long-duration sequencing (107 hours) on a single chip with high yield (>50 Gb). In human genomic DNA sequencing, CycloneSEQ was able to produce long reads with N50 33.6 kb and modal identity 97.0%. Preliminary findings on human whole-genome de novo assembly, variant calling, metagenomics sequencing, and single-cell RNA sequencing have further highlighted CycloneSEQs potential across different areas of genomics.

genomics↗

bioRxiv2024

Berberine is a novel mitochondrial calcium uniporter (MCU) inhibitor that disrupts MCU-EMRE assembly

The mitochondrial calcium uniporter (MCU) complex mediates Ca2+ entry into mitochondrial, which plays a crucial role in regulating cellular energy metabolism and apoptosis. Dysregulation of MCU is implicated in various diseases, such as neurodegenerative disorders, cardiac diseases and cancer. Despite its importance, developing specific and clinically viable MCU inhibitors has been challenging. Here, we identify Berberine, a well-established drug with a documented safety profile, as a potent MCU inhibitor through a virtual screening of an FDA-approved drug library. Berberine localizes within mitochondria and directly binds to the juxtamembrane loop domain of MCU. This binding disrupts the interaction of MCU with its essential regulator, EMRE, thereby inhibiting rapid Ca2+ entry into the mitochondria. Notably, Berberine pretreatment reduces mitochondrial Ca2+ overload and mitigate ischemia/reperfusion-induced myocardial injury in mice. Our findings establish Berberine as a potent MCU inhibitor, offering a safe therapeutic strategy for diseases associated with dysregulated mitochondrial calcium homeostasis.

cell biology↗

bioRxiv2024

P(all-atom) Is Unlocking New Path For Protein Design

We introduce Pallatom, an innovative protein generation model capable of producing protein structures with all-atom coordinates. Pallatom directly learns and models the joint distribution P(structure, seq) by focusing on P(all-atom), effectively addressing the interdependence between sequence and structure in protein generation. To achieve this, we propose a novel network architecture specifically designed for all-atom protein generation. Our model employs a dual-track framework that tokenizes proteins into residue-level and atomic-level representations, integrating them through a multi-layer decoding process with "traversing" representations and recycling mechanism. We also introduce the atom14 representation method, which unifies the description of unknown side-chain coordinates, ensuring high fidelity between the generated all-atom conformation and its physical structure. Experimental results demonstrate that Pallatom excels in key metrics of protein design, including designability, diversity, and novelty, showing significant improvements across the board. Our model not only enhances the accuracy of protein generation but also exhibits excellent sampling efficiency, paving the way for future applications in larger and more complex systems.

bioinformatics↗

bioRxiv2024

Loss function of tumor suppressor FRMD8 confers resistance to tamoxifen therapy via a dual mechanism

Approximately 40% ER-positive breast cancer patients suffer from therapeutic resistance to tamoxifen. Although reduced ER level is the major cause of tamoxifen resistance, the underlying mechanisms remain elusive. Here, we report that FRMD8 raises the level of ER at both transcriptional and post-translational layers. FRMD8 deficiency in MMTV-Cre+; Frmd8fl/fl; PyMT mice accelerates mammary tumor growth and loss of luminal phenotype, and confers tamoxifen resistance. Single-cell RNA profiling reveals that Frmd8 loss decreases the proportion of hormone-sensing differentiated epithelial cells and downregulates the levels of ER. Mechanically, on one hand, loss of FRMD8 inhibits ESR1 transcription via suppressing the expression of FOXO3A, a transcription factor of ESR1. On the other hand, FRMD8 interacts both with ER and UBE3A, and disrupts the interaction of UBE3A with ER, thereby blocking UBE3A-mediated ER degradation. In breast cancer patients, FRMD8 gene promoter is found hypermethylated and low level of FRMD8 predicts poor prognosis. Therefore, FRMD8 is an important regulator of ER and may control therapeutic sensitivity to tamoxifen in ER-positive breast cancer patients. Impact StatementTumor suppressive protein FRMD8 inhibits breast cancer progression by regulating the level of ER and has the possibility to be a potential target for overcoming tamoxifen treatment resistance.

cancer biology↗

bioRxiv2024

Structures of the multi-domain oxygen sensor DosP: remote control of a c-di-GMP phosphodiesterase by a regulatory PAS domain

The heme-based direct oxygen sensor DosP degrades c-di-GMP, a second messenger nearly unique to bacteria. In stationary phase Escherichia coli, DosP is the most abundant c-di-GMP phosphodiesterase. Ligation of O2 to a heme-binding PAS domain (hPAS) of the protein enhances the phosphodiesterase through an allosteric mechanism that has remained elusive. We determined six structures of full-length DosP in its aerobic or anaerobic conformations, with or without c-di-GMP. DosP is an elongated dimer with the regulatory heme and phosphodiesterase separated by nearly 180 [A]. In the absence of substrate, regardless of the heme status, DosP presents an equilibrium of two distinct conformations. Binding of substrate induces DosP to adopt a single, ON-state or OFF-state conformation depending on its heme status. Structural and biochemical studies of this multi-domain sensor and its mutants provide insights into signal regulation of second-messenger levels.

biochemistry↗

bioRxiv2024

High-yield production of recombinant human myelin oligodendrocyte glycoprotein in SHuffle bacteria without a refolding step

1.1Experimental autoimmune encephalomyelitis (EAE) is a model for central nervous system (CNS) autoimmune demyelinating diseases such as multiple sclerosis (MS) and MOG antibody-associated disease (MOGAD). Immunization with the extracellular domain of recombinant human MOG (rhMOG), which contains pathogenic antibody and T cell epitopes, induces B cell-dependent EAE for studies in mice. However, these studies have been hampered by rhMOG availability due to its insolubility when overexpressed in bacterial cells, and the requirement for inefficient denaturation and refolding. Here, we describe a new protocol for the high-yield production of soluble rhMOG in SHuffle cells, a commercially available E. coli strain engineered to facilitate disulfide bond formation in the cytoplasm. SHuffle cells can produce a soluble fraction of rhMOG yielding >100 mg/L. Analytical size exclusion chromatography multi-angle light scattering (SEC-MALS) and differential scanning fluorimetry of purified rhMOG reveals a homogeneous monomer with a high melting temperature, indicative of a well-folded protein. An in vitro proliferation assay establishes that purified rhMOG can be processed and recognized by T cells expressing a T cell receptor (TCR) specific for the immunodominant MOG35-55 peptide epitope. Lastly, immunization of wild-type, but not B cell deficient, mice with rhMOG resulted in robust induction of EAE, indicating a B cell-dependent induction. Our SHuffle cell method greatly simplifies rhMOG production by combining the high yield and speed of bacterial cell expression with enhanced disulfide bond formation and folding, which will enable further investigation of B cell-dependent EAE and expand human research of MOG in CNS demyelinating diseases.

biochemistry↗

bioRxiv2024

YTHDF2 upregulation and relocation dictate CD8 T cell polyfunctionality in tumor immunity

Epigenetic traits impact the antitumor function of CD8 T cells, yet whether and how RNA methylation programs engage in T cell immunity is poorly understood. Here we show that the N6-methyladenosine (m6A) RNA reader YTHDF2 is highly expressed in early effector or effector-like CD8 T cells and is partially distributed in the nucleus. YTHDF2 loss in T cells exacerbates tumor progression and confers unresponsiveness to PD-1 blockade in mice and humans. In addition to initiating RNA decay for mitochondrial fitness, YTHDF2 can orchestrate chromatin regulation to promote T cell polyfunctionality. YTHDF2-mediatd preservation of gene transcription arises from the interaction of YTHDF2 with IKZF1/3. Accordingly, immunotherapy-induced efficacy could be largely restored in YTHDF2-deficient T cells through combinational use of lenalidomide. Moreover, m6A recognition is fundamental for YTHDF2 translocation to the nucleus and autoregulation at the RNA level. Thus, YTHDF2 coordinates epitranscriptional and transcriptional networks to potentiate T cell immunity. HighlightsO_LIYTHDF2 expression and distribution underpin the threshold for bona fide CD8 T cell effector response C_LIO_LICanonical YTHDF2-mRNA decay pathway alleviates mitochondrial stress and CD8 T cell exhaustion C_LIO_LINuclear YTHDF2 sequesters IKZF1/3-mediated transcriptional repression to safeguard CD8 T cell polyfunctionality C_LIO_LIThe tumoricidal activity of YTHDF2-deficient CD8 T cells could be repaired through the synergy of anti-PD-1 and lenalidomide C_LI

immunology↗

bioRxiv2024

VirID: Beyond Virus Discovery - An Integrated Platform for Comprehensive RNA Virus Characterization

RNA viruses exhibit vast phylogenetic diversity and can significantly impact public health and agriculture. However, current bioinformatics tools for viral discovery from metagenomic data frequently generate false positive virus results, overestimate viral diversity, and misclassify virus sequences. Additionally, current tools often fail to determine virus-host associations, which hampers investigation of the potential threat posed by a newly detected virus. To address these issues we developed VirID, a software tool specifically designed for the discovery and characterization of RNA viruses from metagenomic data. The basis of VirID is a comprehensive RNA-dependent RNA polymerase (RdRP) database to enhance a workflow that includes RNA virus discovery, phylogenetic analysis, and phylogeny-based virus characterization. Benchmark tests on a simulated data set demonstrated that VirID had high accuracy in profiling viruses and estimating viral richness. In evaluations with real-world samples, VirID was able to identity RNA viruses of all type, but also provided accurate estimations of viral genetic diversity and virus classification, as well as comprehensive insights into virus associations with humans, animals, and plants. VirID therefore offers a robust tool for virus discovery and serves as a valuable resource in basic virological studies, pathogen surveillance, and early warning systems for infectious disease outbreaks.

bioinformatics↗

bioRxiv2024

SOFisher: Reinforcement Learning-Guided Experiment Designs for Spatial Omics

Spatial omics technologies enable the precise detection of proteins and RNAs at high spatial resolution. Designing spatial omics experiments requires careful consideration of "what" targets to measure and "where" to position the field of views (FOVs). Current FOV sampling strategies often involve acquiring densely sampled FOVs and stitching them together, which is time-consuming, resource-intensive, and sometimes impossible. To optimize FOV sampling strategies, we developed SOFisher, a reinforcement learning-based framework that harnesses the knowledge gained from the sequence of previously sampled FOVs to guide the selection of the next FOV position, to improve the efficiency of capturing more regions of interest. We rigorously evaluated SOFishers performance using comprehensive simulations based on real spatial datasets, and our results clearly demonstrated that SOFisher consistently outperformed the conventional approach across various metrics. SOFishers robustness and generalizability were further validated through cross-domain generalization tests and its adaptability to varying FOV sizes. On a real Alzheimers Disease (AD) dataset, SOFisher successfully guided the selection of FOVs containing neurofibrillary tangles and amyloid-{beta} plaques in both single and dual target tissue landmark scenarios. Remarkably, SOFisher-guided experiment design of spatial single-omics on limited tissue areas yielded insights into AD-related cell states, subtypes, and gene programs previously obtained through extensive spatial multi-omics experiments. SOFisher has the potential to revolutionize the experiment design of spatial biology.

bioinformatics↗