Biology subjects

Tang, J.

Publications and source records attributed to Tang, J..

At least 91 records · Page 5Linked to original sources

bioRxiv2025

Accurate Somatic SV detection via sequence graph model-based local pan-genome optimization

Somatic structural variations (SVs) are critical genomic alterations in cancer genomes. Long-read sequencing (LRS) is theoretically optimal for detecting somatic SVs. However, influenced by reads-to-reference alignment errors, particularly in low-complexity or highly repetitive genomic intervals, current LRS-based somatic SV callers fail to accurately detect SVs. Moreover, the lack of comprehensive ground-truth datasets hinders accurate evaluation. Here, we introduce SVscope, a novel algorithm that fundamentally addresses these challenges by leveraging full-length sequence information from span-reads and integrating local graph-genome optimization with a random forest strategy. SVscope outperforms state-of-the-art methods on six paired long-read whole-genome sequencing (WGS) benchmark cell lines, achieving a maximum F1-score improvement of 16.7%. In simulated datasets, SVscope demonstrates superior performance in both somatic SV detection and read phasing tasks. Based on the findings from SVscope, we validated 47 somatic SVs, thereby significantly expanding the existing experimentally validated ground-truth somatic SV dataset by 69.1%.

bioinformatics↗

bioRxiv2025

Self-iterative multiple instance learning enables the prediction of CD4+ T cell immunogenic epitopes

Accurately predicting the antigen presentation to CD4+ T cells and subsequent induction of immune response is fundamentally important for vaccine development, autoimmune disease treatments, and cancer neoepitope identification. In immunopeptidomics, single-allelic data are highly specific but limited in allele scope, while multi-allelic data contain broader coverage at the cost of weakly labeling. Existing computational approaches either overlook the massive multi-allelic data or introduce label ambiguity due to inadequate modeling strategies. Here, we introduce ImmuScope, a weakly supervised deep-learning framework integrating precise MHC-II antigen presentation, CD4+ T cell epitopes, and immunogenicity predictions. ImmuScope leverages self-iterative multiple-instance learning with positive-anchor triplet loss to explore peptide-MHC-II (pMHC-II) binding from weakly labeled multi-allelic data and single-allelic data, comprising over 600,000 ligands across 142 alleles. Moreover, ImmuScope can also interpret the MHC-II binding specificity and motif deconvolution of immunopeptidomics data. We successfully applied ImmuScope to discover melanoma neoantigens, revealing variations in pMHC-II binding and immunogenicity upon epitope mutations. We further employed ImmuScope to assess the effects of SARS-CoV-2 epitope mutations on immune escape, with its predictions aligned well with experimentally determined immune escape dynamics. Overall, ImmuScope provides a comprehensive solution for CD4+ T cell antigen recognition and immunogenicity assessment, with broad potential for advancing vaccine design and personalized immunotherapy.

bioinformatics↗

bioRxiv2025

A genealogy-based approach for revealing ancestry-specific structures in admixed populations

Elucidating ancestry-specific structures in admixed populations is crucial for comprehending population history and mitigating confounding effects in genome-wide association studies. Existing methods for elucidating the ancestry-specific structures generally rely on frequency-based estimates of genetic relationship matrix (GRM) among admixed individuals after masking segments from ancestry components not being targeted for investigation. However, these approaches disregard linkage information between markers, potentially limiting their resolution in revealing structure within an ancestry component. We introduce ancestry-specific expected GRM (as-eGRM), a novel framework for elucidating the relatedness within ancestry components between admixed individuals. The key design of as-eGRM consists of defining ancestry-specific pairwise relatedness between individuals based on genealogical trees encoded in the Ancestral Recombination Graph (ARG) and local ancestry calls and computing the expectation of the ancestry-specific relatedness across the genome. Comprehensive evaluations using both simulated stepping-stone models of population structure and empirical datasets based on three-way admixed Latino cohorts showed that analysis based on as-eGRM robustly outperforms existing methods in revealing the structure in admixed populations with diverse demographic histories. Taken together, as-eGRM has the promise to better reveal the fine-scale structure within an ancestry component of admixed individuals, which can help improve the robustness and interpretation of findings from association studies of disease or complex traits for these understudied populations.

genetics↗

bioRxiv2025

Toward a privacy-preserving predictive foundation model of single-cell transcriptomics with federated learning and tabular modeling

The ability to pre-train on vast amounts of data to build foundation models (FMs) has achieved remarkable success in numerous domains, including natural language processing, computer vision, and, more recently, single-cell genomics--epitomized by GeneFormer, scGPT, and scFoundation. However, as single-cell FMs begin to train on increasingly large corpora, significant privacy and ethical concerns arise. Moreover, unlike text data, single-cell data is unordered and exhibits a unique tabular structure that most existing single-cell FMs overlook. In this study, we propose Tabula, a privacy-preserving and tabular-structure aware FM designed with federated learning (FL) and tabular modeling. Tabula combines the advantages of FMs and FL, enabling collaborative model training across multiple clients without compromising data privacy. In contrast to earlier single-cell FMs--which treat single-cell data like natural language (viewing cells as "words" defined by genes)--Tabula introduces a novel pretraining strategy that explicitly models the tabular structure of single-cell data. Extensive experimental results show that Tabula outperforms state-of-the-art methods in various downstream tasks (including cell type annotation, gene imputation, gene perturbation, multi-batch integration, and multi-omics integration) while requiring only half the data for pretraining and preserving data privacy. Furthermore, Tabula accurately reveals pairwise and even combinatorial regulatory logic across diverse biological systems, including hematopoiesis, pancreatic endogenesis, neurogenesis, and cardiogenesis. Thus, Tabula provides a new foundation model that explicitly incorporates the tabular nature of single-cell data alongside FL, paving the way for creating a "virtual cell" for human health under critical privacy preservation.

bioinformatics↗

bioRxiv2024

Developmentally Incomplete Barb Rami Increased the Morphological Diversity of Early Feathers

Recent studies have significantly advanced our understanding of the evolutionary processes and developmental mechanisms behind feather branching patterns. However, little attention has been given to the tissue differentiation of branches associate with this process. Here, we describe eight feathers preserved in Burmese ambers, characterized by ribbon-like barb rami composed solely of cortex, lacking a medulla. These feathers are categorized into three morphotypes based on the details of their barb rami morphologies. For each morphotype, potential developmental scenarios were proposed through inference of barb development in chickens, suggesting limited tissue differentiation of barb rami in these early feathers. Functional simulations indicate that the configuration of modern barb rami is aerodynamically more stable than those of early feathers, indicating that flexural stiffness may have been a crucial factor driving the evolution of feather branches toward modern configurations. The discovery of developmentally incomplete barb rami in Burmite feathers implies that, while the three-level hierarchical branching pattern of modern feathers had evolved by the Jurassic period, tissue differentiation patterns of feather branches may not have fully stabilized by the Late Cretaceous. This instability likely contributes to the morphological diversity of early feathers, resulting in forms not seen in modern birds.

evolutionary biology↗

bioRxiv2024

Green Elegance T2T: An upgraded telomere-to-telomere genome assembly and annotation of looseleaf lettuce (Lactuca sativa var. crispa)

Lettuce (Lactuca sativa L.), a globally important vegetable crop, is cultivated in various horticultural varieties. Looseleaf lettuce is one of the main varieties that are consumed. Here, we report a telomere-to-telomere (T2T) genome assembly of looseleaf lettuce (L. sativa var. crispa cv. Green Elegance). By combining the sequencing data of previously reported Green Elegance v1.2, the 42.20 Gb clean ultra-long reads produced by Oxford Nanopore Technology were used to close the gaps and upgrade the genome assembly. After filling 11 gaps in Green Elegance v1.2, the final gapless telomere-to-telomere genome assembly is 2.58 Gb in length with a contig N50 of 282.47 Mb, containing nine centromeres and 18 telomeres. The genome contained 41,375 protein-coding genes, of which 99.10% were functionally annotated. This T2T genome of looseleaf lettuce will be valuable for the identification of genetic variation and the advancement of lettuce breeding.

genomics↗

bioRxiv2024

Interplay between autophagy, the unfolded protein response and the cytoplasmic heat stress response during heat stress in maize

High temperatures can substantially reduce plant survival and reproduction, and therefore decrease crop productivity. Plants activate several pathways in response to high temperatures, including the endoplasmic reticulum (ER) unfolded protein response (UPR), the cytoplasmic heat stress response, and autophagy, which together aid in stress tolerance by refolding or degrading misfolded and unfolded proteins. The relationship between the UPR and autophagy is known, as ER stress activates the UPR, and a key activator of the UPR, Inositol-requiring enzyme type 1 (IRE1B), upregulates autophagy. To assess the relationship between the distinct responses to heat stress in maize, we analyzed the effect of disruption of the UPR, via a bzip60 mutant, on autophagy. We found that genes related to autophagy are upregulated in the bzip60 mutant, and that autophagy is activated in this mutant even in the absence of stress. We in turn analyzed the effect of loss of autophagy on the UPR. The bZIP60 mRNA spliced form, a marker for the UPR, was increased in mutants in ATG10, a core component of the autophagy machinery, during ER stress. By contrast, bZIP60 splicing was unaffected in atg10 mutants in response to heat, whereas cytoplasmic heat stress components were increased. Loss of autophagy therefore differentially affects other heat stress response pathways, depending on the specific stress conditions leading to the activation of the pathways.

plant biology↗

bioRxiv2024

Machine learning engineered PoLixNano nanoparticles overcome delivery barriers for nebulized mRNA therapeutics

There continues to be a dearth of competent inhalable mRNA delivery although it holds great potential for addressing a wide variety of refractory diseases. The huge advances seen with parenteral-administered lipid nanoparticle (LNP) have not been translated into nebulized mRNA delivery due to the aggressive nebulization process and insurmountable barriers inherent to respiratory mucosa. Here, we show amphiphilic block copolymers revealed by machine learning (ML) can spontaneously form stabilized nanoparticles (PoLixNano) with the lipids components of LNP and simultaneously impart the PoLixNano with "shield" (shear force-resistant) and "spear" (pulmonary barriers-penetrating abilities) capabilities. We present a ML approach that leverages physicochemical properties and inhaled mRNA transfection profiles of a chemically diverse library of polymeric components to validate the integration of "shield" and "spear" properties as highly predictive indicators of transfection efficiency. This quantitative structure-mRNA transfection prediction (QSMTP) model identifies top-performing amphiphilic-copolymers from more than 10000 candidates and suggests their mucus-penetrating ability outweights the shear force-resistant property in contributing to efficient mRNA transfection. The optimized PoLixNano substantially outperforms the LNP counterpart and mediates up to 1114-times higher levels of mRNA transfection in animal models with negligible toxicities. The PoLixNano promotes overwhelming SARS-CoV-2 antigen-specific sIgA antibody secretion and expansion of TRM cells which collectively confers 100% protection in mice against lethal SARS-CoV-2 challenges and blocks the transmission of Omicron variant between hamsters. PoLixNano also displays versatile therapeutic potential in lung carcinoma and cystic fibrosis models. Our study provides new insights for designing delivery platforms of aerosol-inhaled mRNA therapeutics with clinical translation potential.

synthetic biology↗

bioRxiv2024

Uncovering pre-sensitizing agents to FLT3 inhibitors in acute myeloid leukemia with ReSisTrace lineage tracing

While FLT3 inhibitors have significantly improved the treatment of aggressive FLT3-mutated acute myeloid leukemia (AML), the emergence of resistance remains as a major challenge. Here, we applied our recently developed single-cell lineage-tracing method ReSisTrace to identify cells that are pre-resistant or pre-sensitive to FLT3 inhibitors midostaurin and quizartinib in FLT3-ITD-positive AML. By comparing the gene expression profiles of these cells, we unraveled the transcriptional pre-resistance signatures, including G1 to S phase transition 1 (GSPT1) gene. Targeting GSPT1 with the small molecule CC-90009 exhibited strong synergistic effect when combined with FLT3 inhibitors in the FLT3-ITD-mutated MOLM-13 and MV4-11 cell lines and primary AML patient samples. Further, we identified novel compounds that induced transcriptomic changes opposite to the pre-resistance signatures, thereby driving cells to FLT3 inhibitor-sensitive states. Vistusertib (mTOR inhibitor), linsitinib (IGF1R and insulin receptor inhibitor), and meisoindigo (IGF1R and Src family kinase inhibitor), all inhibiting pathways parallel to or downstream of oncogenic FLT3 signaling, were predicted and validated to pre-sensitize the FLT3-ITD-mutated cell lines and primary cells to FLT3 inhibitors. Collectively, these findings demonstrate the validity of our lineage-tracing method in unveiling pre-existing transcriptional features of treatment vulnerability in hematological cancers, and elucidate novel strategies for enhancing FLT3 inhibitor treatment efficacy in FLT3-ITD-positive AML by preventing the emergence of treatment resistance.

cancer biology↗

bioRxiv2024

Interpretable high-resolution dimension reduction of spatial transcriptomics data by DeepFuseNMF

Spatial transcriptomics (ST) technologies have revolutionized tissue architecture studies by capturing gene expression with spatial context. However, high-dimensional ST data often have limited spatial resolution and exhibit considerable noise and sparsity, posing significant challenges in deciphering subtle spatial structures and underlying biological activities. Here, we introduce SpaHDmap, an interpretable dimension reduction framework that enhances spatial resolution by integrating ST gene expression with high-resolution histology images. SpaHDmap incorporates non-negative matrix factorization into a multimodal fusion encoder-decoder architecture, enabling the identification of interpretable, high-resolution embeddings. Furthermore, SpaHDmap can simultaneously analyze multiple samples and is compatible with various types of histology images. Extensive evaluations on synthetic and real ST datasets from various technologies and tissue types demonstrate that SpaHDmap can effectively produce highly interpretable, high-resolution embeddings, and detects refined spatial structures. SpaHDmap represents a powerful approach for integrating ST data and histology images, offering deeper insights into complex tissue structures and functions.

bioinformatics↗

bioRxiv2024

Novel Pannexin 1 isoform is increased in cancer

Alternative translation initiation (ATI) is a process of increasing protein diversity from one transcript, allowing cells to rapidly respond to signals, which is particularly important in cancer cells. Here, we report potential internal translation start sites exist in PANX1 which have implications in trafficking and channel function. Using mouse (mPANX1) constructs for each internal methionine, we saw that these PANX1 isoforms were N-glycosylated, could traffic to the cell surface and mPANX1-M37 formed functional channels activated by C-terminus cleavage or 1-adrenoceptor stimulation. We also identified a ~25 kDa isoform of mPANX1 (mPANX1-25K) endogenously expressed in mouse melanoma cell lines that could be confirmed with a cognate peptide. mPANX1-25K lacks the mPANX1 N-terminus and most likely corresponds to the M210 internal translation start site since we could not identify any alternative transcripts that would produce this ATI product. When we expressed the human equivalent of M210 in Hs578T PANX1 KO cells with and without wildtype human PANX1, we determined M211 exhibits a predominantly intracellular localization, is N-glycosylated and can interact with full-length human PANX1. Collectively, these findings indicate species specific differences in the abundance of PANX1 ATI isoforms which could act independently or in conjunction with the canonical full-length protein in melanoma.

cell biology↗

bioRxiv2024

Selective disruption of Traf1/cIAP2 interaction attenuates inflammatory responses and limits sepsis and rheumatoid arthritis

Tumor necrosis factor receptor-associated factor 1 (TRAF1) is an immune signaling adapter protein linked to increased susceptibility to rheumatoid arthritis (RA). TRAF1 has dual roles in regulating NF-{kappa}B and MAPK signaling: it promotes signaling through its association with cellular inhibitor of apoptosis 2 (cIAP2) downstream of certain tumor necrosis factor receptor (TNFR) family members but inhibits Toll-like receptor (TLR) signaling by limiting linear ubiquitination of key signaling proteins. Here, we identify a critical mutation in TRAF1 (V203A in humans, V196A in mice) that disrupts its interaction with cIAP2, leading to a significant reduction in TLR signaling and downstream inflammation in human and murine macrophages. We demonstrate that TRAF1 is recruited to the TLR4 complex and is indispensable for the recruitment of cIAP2, facilitating TAK1 phosphorylation and the activation of NF-{kappa}B and MAPK signaling pathways. Remarkably, mice harboring the TRAF1 V196A mutation are protected from LPS-induced septic shock and exhibit markedly reduced joint inflammation and disease severity in a collagen antibody-induced arthritis (CAIA) model of RA. These findings reveal a previously unrecognized and crucial role for the TRAF1/cIAP2 axis in promoting inflammation and offer a promising foundation for the development of novel therapeutic strategies for inflammatory conditions, such as sepsis and RA.

immunology↗

bioRxiv2024

Evaluation of a Deep Learning Based Approach to Computational Label Free Cell Viability Quantification

One of the most common techniques found in a cell biology or tissue engineering lab is the cytotoxicity assay. This can be performed using a variety of different dyes and stains and various protocols to result in a clear indication of dead and live cells within a culture to quantify the viability of a culture and monitor for sudden drops or increases in viability by a drug, material, viral vector, etc introduced into the culture. This assay helps cell biologists determine the health of their culture and what toxicity added substances may add to the culture and whether they are appropriate and safe to use with human cells. However, many of the dyes and stains used for this process are eventually toxic to cells, rendering the cells useless after testing and preventing real time monitoring of the same culture over a period of hours or days. Computation biology is moving cell biology towards novel and innovative techniques such as in silico labeling and dye free labeling using deep learning algorithms. In this work, we investigate whether it is feasible to train a Resnet CNN model to detect morphological changes in human cells that indicate cell death in order to classify cells as live or dead without utilizing a stain or dye. This work also aims to train one CNN model to count all cells regardless of viability status to get a total cell count, and then one CNN model that specifically identifies and counts all of the dead cells for an accurate dead and live cell total by utilizing both pieces of data to determine a general viability percentage for the culture. Additionally, this work explores the use of various image enhancements to understand if this process helps or impedes the deep learning models in their detection of total cells and dead cells.

cell biology↗

bioRxiv2024

Size evolution of gigantic genomes suggests stochastic outcomes of transposable element/host silencing interactions

Size evolution among gigantic genomes involves gain and loss of many gigabases of transposable elements (TEs), sequences that parasitize host genomes. Animals suppress TEs using piRNA and KRAB-ZFP pathways. TEs and hosts coevolve in an arms race, where suppression strength reflects TE fitness costs. In enormous genomes, additional TE costs become miniscule. How, then, do TEs and host suppression invoke further addition of massive DNA amounts? We analyzed TE proliferation histories, deletion rates, and community diversities in six salamander genomes (21.3 - 49.9 Gb), alongside gonadal expression of TEs and suppression pathways. TE activity is higher in testes than ovaries, attributable to lower KRAB-ZFP suppression. Unexpectedly, genome size/expansion is uncorrelated with TE deletion rate, proliferation history, expression, and host suppression. Also, TE community diversity increases with genome size, contrasting theoretical predictions. TE/host antagonism in gigantic genomes likely produces stochastic TE accumulation, determined by noisy intermolecular interactions in huge genomes/cells.

evolutionary biology↗

bioRxiv2024

Metabolomics profiling reveals association of fermented herbal feed additive, growth performance and gut microbiota in piglets

Fermented Chinese medicine (FCM), as a functional feed additive, has been widely recognized to play a significant role in protecting the intestinal health of piglets and enhancing productive performance. However, the relationship between the active components of FCM, gut microbiota, and their beneficial effects on animal performance remains unclear. In this study, metabolomics analysis revealed a significant increase in the main contents of lactic acid and propionic acid in FCM, while most glycosides and their derivatives decreased after three days of microbial fermentation. Subsequently, piglets were fed a basic diet supplemented with 1% FCM, while the control group received only the basic diet. The results indicated a significant increase in feed intake and average daily gain within 14 days (P<0.05) due to FCM supplementation. Additionally, FCM significantly improved feed efficiency from 1.76 to 1.50 (P<0.05). Meanwhile, piglets in the FCM group exhibited significantly lower frequencies of diarrhea and coughing, indicating better health. Moreover, high-throughput sequencing analysis revealed higher abundances of Lactobacillus agilis, Megasphaera elsdenii, Ligilactobacillus, and Veillonellaceae in piglets fed with FCM. In summary, FCM increased the content of active ingredients through microbial fermentation and regulated the intestinal microbiota to improve the health of piglets. FCM offers a promising potential for enhancing production performance and economic efficiency in the livestock industry.

microbiology↗

bioRxiv2024

Benchmarking algorithms for spatially variable gene identification in spatial transcriptomics

The rapid development of spatial transcriptomics has underscored the importance of identifying spatially variable genes. As a fundamental task in spatial transcriptomic data analysis, spatially variable gene identification has been extensively studied. However, the lack of comprehensive benchmark makes it difficult to validate the effectiveness of various algorithms scattered across a large number of studies with real-world datasets. In response, this article proposes a benchmark framework to evaluate algorithms for identifying spatially variable genes through the analysis of synthesized and real-world datasets, aiming to identify the best algorithms and their corresponding application scenarios. This framework can assist medical and life scientists in selecting suitable algorithms for their research, while also aid bioinformatics scientists in developing more powerful and efficient computational methods in spatial transcriptomic research.

bioinformatics↗

bioRxiv2024

Alzheimer Disease Knowledge Graph Enhances Knowledge Discovery and Disease Prediction

BackgroundAlzheimers disease (AD), a progressive neurodegenerative disorder, continues to increase in prevalence without any effective treatments to date. In this context, knowledge graphs (KGs) have emerged as a pivotal tool in biomedical research, offering new perspectives on drug repurposing and biomarker discovery by analyzing intricate network structures. Our study seeks to build an AD-specific knowledge graph, highlighting interactions among AD, genes, variants, chemicals, drugs, and other diseases. The goal is to shed light on existing treatments, potential targets, and diagnostic methods for AD, thereby aiding in drug repurposing and the identification of biomarkers. ResultsWe annotated 800 PubMed abstracts and leveraged GPT-4 for text augmentation to enrich our training data for named entity recognition (NER) and relation classification. A comprehensive data mining model, integrating NER and relationship classification, was trained on the annotated corpus. This model was subsequently applied to extract relation triplets from unannotated abstracts. To enhance entity linking, we utilized a suite of reference biomedical databases and refine the linking accuracy through abbreviation resolution. As a result, we successfully identified 3,199,276 entity mentions and 633,733 triplets, elucidating connections between 5,000 unique entities. These connections were pivotal in constructing a comprehensive Alzheimers Disease Knowledge Graph (ADKG). We also integrated the ADKG constructed after entity linking with other biomedical databases. The ADKG served as a training ground for Knowledge Graph Embedding models with the high-ranking predicted triplets supported by evidence, underscoring the utility of ADKG in generating testable scientific hypotheses. Further application of ADKG in predictive modeling using the UK Biobank data revealed models based on ADKG outperforming others, as evidenced by higher values in the areas under the receiver operating characteristic (ROC) curves. ConclusionThe ADKG is a valuable resource for generating hypotheses and enhancing predictive models, highlighting its potential to advance ADs disease research and treatment strategies.

bioinformatics↗

bioRxiv2024

Path-based reasoning in biomedical knowledge graphs

Understanding complex interactions in biomedical networks is crucial for advancements in biomedicine, but traditional link prediction (LP) methods are limited in capturing this complexity. Representation-based learning techniques improve prediction accuracy by mapping nodes to low-dimensional embeddings, yet they often struggle with interpretability and scalability. We present BioPathNet, a novel graph neural network framework based on the Neural Bellman-Ford Network (NBFNet), addressing these limitations through path-based reasoning for LP in biomedical knowledge graphs. Unlike node-embedding frameworks, BioPathNet learns representations between node pairs by considering all relations along paths, enhancing prediction accuracy and interpretability. This allows visualization of influential paths and facilitates biological validation. BioPathNet leverages a background regulatory graph (BRG) for enhanced message passing and uses stringent negative sampling to improve precision. In evaluations across various LP tasks, such as gene function annotation, drug-disease indication, synthetic lethality, and lncRNA-mRNA interaction prediction, BioPathNet consistently outperformed shallow node embedding methods, relational graph neural networks and task-specific state-of-the-art methods, demonstrating robust performance and versatility. Our study predicts novel drug indications for diseases like acute lymphoblastic leukemia (ALL) and Alzheimers, validated by medical experts and clinical trials. We also identified new synthetic lethality gene pairs and regulatory interactions involving lncRNAs and target genes, confirmed through literature reviews. BioPathNets interpretability will enable researchers to trace prediction paths and gain molecular insights, making it a valuable tool for drug discovery, personalized medicine and biology in general.

bioinformatics↗