Biology subjects

Dong, N.

Publications and source records attributed to Dong, N..

14 recordsLinked to original sources

bioRxiv2026

DeepD enables efficient genome modeling and controllable long-context sequence generation across domains of life

Genome foundation models hold great potential for advancing biological research, but modeling diverse complex genomes and generating controllable genome-scale sequences remain challenging. We present DeepD, a phylogeny-informed sparse mixture-of-experts model operating at single-nucleotide resolution over a 1-million-token context while reducing active computation. Pretrained on unselected whole-genome sequences across all domains of life, DeepD learns genic and intergenic sequence features and supports broad sequence-to-function prediction across species. DeepD performs well in human protein fitness and disease variants effect prediction, such as BRCA1 and BRCA2, with particular strength in regulatory regions, such as promoter and transcription-factor binding sites. Expert routing captures exon-intron boundaries, transposable elements, and distal cis-regulatory elements. DeepD also generates experimentally validated core promoters with specified transcriptional strengths. Finally, the Context-Anchor-Body strategy enables generation of complete yeast chromosome sequences and megabase-scale Arabidopsis segments, preserving gene structure and order while introducing variations within natural ranges. Together, DeepD is an efficient genome foundation model that unifies accurate sequence-to-function prediction with controllable genome-scale generation, advancing predictive and generative design of complex genomes.

bioinformatics↗

bioRxiv2026

Circulating apoB-containing lipoproteins constituted blood-borne innate immune barrier limits ZIKV dissemination and congenital infection risk

Following establishment of primary viremia, flaviviruses require prompt penetration through vascular endothelium to achieve systemic dissemination, yet most infections remain asymptomatic or self-limited. Canonical defenses, including complement and phagocyte clearance, cannot fully explain this containment, suggesting an additional blood-borne antiviral barrier. Here, we identify apoB containing lipoproteins as potent restriction factors against Zika virus (ZIKV), a devastating danger of global pregnancy. By displaying surface accessible phosphatidylserine (PS), they competitively occupy host PS receptors and block entry by viruses exploiting apoptotic mimicry, including flaviviruses, Ebola, and chikungunya virus. In vivo, increasing apoB lipoproteins enhances serum antiviral activity and limits systemic ZIKV dissemination. This barrier extends to the maternal fetal interface, where elevated maternal LDL attenuates vertical transmission and fetal abnormalities. Its efficacy is modulated by gestational fluctuations in maternal LDL level and low abundance in fetal circulating LDL. Collectively, apoB containing lipoproteins form a constitutive blood-borne innate immune barrier that restrains apoptotic-mimicry viruses during the intravascular phase, limits viral dissemination and reduces congenital infection risk. This study provides a rationale for developing diet augmentation strategy to strengthen this innate barrier in at-risk individuals.

microbiology↗

bioRxiv2026

Functional and evolutionary insights into the emerging tet(X4)-carrying non-O1/O139 Vibrio cholerae from retail pork

The tigecycline resistance gene tet(X4) is prevalent in Enterobacteriaceae, particularly in Escherichia coli. To our knowledge, no study has reported the dissemination dynamics of tet(X4) in Vibrio spp. Herein, we isolated and characterized a first tet(X4)-positive non-O1/O139 Vibrio cholerae isolate from retail pork. Genomic sequencing identified a novel tet(X4) variant in the V. cholerae chromosome, harboring a G568A nucleotide substitution that resulted in an Ala190Thr (A190T) amino acid substitution in Tet(X4). While this Tet(X4)-A190T variant conferred lower phenotypic resistance to tetracyclines (including tigecycline) than the wild-type Tet(X4), its overall catalytic efficiency against these antibiotics was paradoxically enhanced despite a reduced substrate affinity. Genomic comparisons revealed that two copies of ISCR2 flanked the variant gene, and the structure was ISCR2-hp-hp-abh-tet(X4)G568A -ISCR2, which is highly homologous to the reported E. coli plasmids carrying tet(X4). In addition, it confirmed the presence of an ISCR2-mediated circular intermediate, proving this modules capacity for horizontal transfer of the tet(X4)G568A variant. Furthermore, the ISCR2-tet(X4) genetic structure carrying the G568A substitution was integrated within a chimeric SXT/R391-like integrative and conjugative element (ICE), which is also serving as a vehicle for genetic dissemination. As per our knowledge, this is the first report on the emergence of SXT/R391-like ICE carrying tet(X4) in Vibrio strains. Our finding demonstrates that the clinically relevant tigecycline resistance gene tet(X4), previously confined mainly to Enterobacterales from humans and livestock, is now actively spreading into environmental Vibrio populations. This cross-species transfer highlights a previously underappreciated ecological and public health concern in aquatic ecosystems.

microbiology↗

bioRxiv2026

EssTFNet: Integration of Adaptive Time - Frequency and DNA Language Models for Interpretable Human Essential Gene Prediction

Essential genes are defined as those that are indispensable for an organisms survival. The loss of function of these genes results in cell death or an inability to complete the normal life cycle. Research on essential genes is pivotal in elucidating the origin and evolution of life, as well as in identifying potential therapeutic targets. Therefore, accurately predicting essential genes is of great scientific importance and has many applications in basic research and the biomedical field. In this study, we propose EssTFNet, a novel and interpretable deep learning framework that combines adaptive time-frequency analysis with a DNA language model to achieve accurate prediction of human essential genes while enabling mechanistic biological interpretation. Specifically, EssTFNet leverages the architecture of ATFNet, which innovatively maps DNA and protein sequences into equivalent time-series signals to extract periodic and non-stationary features, thereby enhancing the models capacity to capture complex sequence patterns. Through effective feature selection and architectural optimization, EssTFNet strikes a favorable balance among prediction accuracy, model interpretability, and cross-tissues generalization capability, significantly outperforming current mainstream sequence-based deep learning methods with an AUC of 0.8336 and an AUPR of 0.8212. Additionally, the DeepLIFT attribution method was employed to identify functional motifs closely associated with gene essentiality, offering valuable insights for experimental validation. For the convenience of researchers, we have developed an easy-to-use web server and made it along with the source code in a GitHub repository: https://github.com/QIANJINYDX/EssTFNet. Overall, this study presents a potentially useful methodological framework for human essential genes prediction, which could provide valuable insights for future research and applications in this field.

bioinformatics↗

bioRxiv2025

MassNet: billion-scale AI-friendly mass spectral corpus enables robust de novo peptide sequencing

Breakthroughs in artificial intelligence (AI) for natural language processing and computer vision have been largely driven by high-quality, large-scale datasets such as OpenWebText and ImageNet. Inspired by this, we present MassNet, a foundational resource for proteomics designed to accelerate deep learning applications. MassNet is the largest known corpus of data-dependent acquisition (DDA) mass spectrometry (MS) data, derived from ~30 TB of raw files and comprising 1.54 billion MS/MS spectra, resulting in 558 million peptide-spectrum matches (PSMs) across 35 species, including animals, plants, and microbes. Within the human subset, MassNet includes more than 1.7 million precursors and 19,966 proteins, covering 98% of annotated human proteins. To enable efficient AI training, we developed the Mass Spectrometry Data Tensor (MSDT), a structured format based on Parquet that enables standardized, high-performance batch access and seamless integration with GPU and TPU platforms for distributed training. We further extended MassNet to support de novo peptide sequencing, which infers peptide sequences directly from MS/MS spectra without reference databases, and is critical for discovering novel proteins, characterizing non-model organisms, and identifying post-translational modifications (PTMs). We introduce XuanjiNovo, a non-autoregressive Transformer model that leverages a curriculum learning strategy to enhance training stability. By dynamically adjusting learning difficulty based on model performance, XuanjiNovo achieves smooth convergence on complex, multi-distributional data without manual hyperparameter tuning. Trained on 100 million PSMs from the MassNet, it consistently outperforms state-of-the-art methods across diverse benchmarking tasks. Peptide recall exceeds 0.8 on the Bacteroides thetaiotaomicron and Zea mays datasets. On human data acquired using the Orbitrap Astral platform, XuanjiNovo achieves achieves 38.8% to 144.3% improvement over existing models. MassNet represents the first large-scale, standardized foundational dataset in proteomics, marking a critical milestone in the integration of artificial intelligence into proteomics research.

bioinformatics↗

bioRxiv2025

Unveiling Protein-DNA Interdependency: Harnessing Unified Multimodal Sequence Modeling, Understanding and Generation

The interdependent orchestration of proteins, coding DNA sequences (CDS), and adjacent regulatory DNA elements underpins biological functionality, where evolutionary constraints shape codon usage, transcriptional regulation, and structural stability. We present ProDMM, a multimodal framework that deciphers these systemic relationships through unified sequence modeling. Integrating a context-aware encoder with conditional generation, our architecture captures codependent patterns between protein motifs and cis-regulatory syntax. ProDMMs encoder demonstrates superior predictive capabilities in forecasting pathway-level phenotypes and protein expression profiles, outperforming conventional single-modality approaches. Its decoder generates synthetically optimized regulatory elements by resolving spatial dependencies between coding sequences and adjacent non-coding regions, enabling transcriptional fine-tuning unavailable to sequence-agnostic methods. By resolving how localized CDS-NCDS-protein interdependencies propagate to system-level phenotypes, ProDMM establishes a paradigm for holistic biological sequence engineering. This approach enables coordinated optimization of translation efficiency, enzymatic activity, and pathway flux, offering transformative potential for designing tailored biosynthesis systems and advancing sustainable bioproduction strategies.

bioinformatics↗

bioRxiv2024

Deep generative model for protein subcellular localization prediction

Protein sequence determines not only its structure but also its subcellular localization. Although a series of artificial intelligence models have been reported to predict protein subcellular localization, most of them provide only textual outputs. Here, we present deepGPS, a deep generative model for protein subcellular localization prediction. After trained with both protein primary sequences and protein subcellular localization fluorescence images, deepGPS shows the ability to predict cytoplasmic and nuclear localizations by reporting both textual labels and generative images as outputs. In addition, deepGPS shows potential to be further extended for other types of subcellular localization prediction, even with limited input data volumes for training. Finally, an openGPS website (https://bits.fudan.edu.cn/opengps) is constructed to provide a public and convenient platform for protein subcellular localization prediction with the scientific community.

bioinformatics↗

bioRxiv2024

Genomics-FM: Universal Foundation Model for Versatile andData-Efficient Functional Genomic Analysis

Artificial intelligence (AI) plays a crucial role in functional genomic analysis, offering great potential for comprehending biological phenomena such as heredity, development, diseases, and evolution. However, the development of AI models needs substantial labeled data, and these models are typically task-specific with limited generalizability to various applications. Here, we develop Genomics-FM, a genomic vocabulary driven foundation model that enables versatile and label-efficient functional genomic analysis. Specifically, Genomics-FM is first pretrained with ensemble genomic vocabulary on vast unlabelled data to learn comprehensive and generalizable representations and then finetuned with specific genomic vocabulary on limited labeled data to selectively activate and adapt the pretraining knowledge for specific tasks. We show that Genomics-FM significantly reduces the dependence on labeled data, and demonstrates the capability to outperform existing models across a comprehensive suite of tasks including genome annotation, epigenomic and expression profile prediction, and variant effect assessment. Remarkably, Genomics-FM even shows impressive zero-shot predictive capabilities across diverse species and tissues and exhibits noticeable adaptability to RNA-related tasks. With feasibility in data scarcity and even cross-domain biological scenarios, Genomics-FM will promote the broad application of AI and empower researchers to tackle previously insurmountable challenges, paving the way for groundbreaking research and discoveries.

bioinformatics↗

bioRxiv2024

Pi-PrimeNovo: An Accurate and Efficient Non-Autoregressive Deep Learning Model for De Novo Peptide Sequencing

Peptide sequencing via tandem mass spectrometry (MS/MS) is fundamental in proteomics data analysis, playing a pivotal role in unraveling the complex world of proteins within biological systems. In contrast to conventional database searching methods, deep learning models excel in de novo sequencing peptides absent from existing databases, thereby facilitating the identification and analysis of novel peptide sequences. Current deep learning models for peptide sequencing predominantly use an autoregressive generation approach, where early errors can cascade, largely affecting overall sequence accuracy. And the usage of sequential decoding algorithms such as beam search suffers from the low inference speed. To address this, we introduce{pi} -PrimeNovo, a non-autoregressive Transformer-based deep learning model designed to perform accurate and efficient de novo peptide sequencing. With the proposed novel architecture,{pi} -PrimeNovo achieves significantly higher accuracy and up to 69x faster sequencing compared to the state-of-the-art methods. This remarkable speed makes it highly suitable for computation-extensive peptide sequencing tasks such as metaproteomic research, where{pi} -PrimeNovo efficiently identifies the microbial species-specific peptides. Moreover,{pi} -PrimeNovo has been demonstrated to have a powerful capability in accurately mining phosphopeptides in a non-enriched phosphoproteomic dataset, showing an alternative solution to detect low-abundance post-translational modifications (PTMs). We suggest that this work not only advances the development of peptide sequencing techniques but also introduces a transformative computational model with wide-range implications for biological research.

bioinformatics↗

bioRxiv2021

Small-molecule cocktails induce the differentiation of human adipose-derived mesenchymal stem cells into hepatocyte-like cells

At present, liver transplantation and hepatocyte therapy are common methods for the treatment of end-stage liver diseases, but they are restricted due to the shortage of liver donors and the safety and effectiveness of hepatocyte sources. Human adipose-derived mesenchymal stem cells (HAD-MSCs) have been applied to efficiently and stably induce phenotypic and functional liver cells or tissues in vitro due to their advantages such as wide sources and easy access to materials. In this study, the HAD-MSCs liver differentiation induction system was established and optimized based on the "cocktail method" of chemical small molecule compounds. We used HAD-MSCs as seed cells and gradually obtained mature hepatoid cells with normal phenotype and function after induction with small molecule compounds and growth factor system in vitro. The hepatoid cells induced by the two groups showed high similarity in phenotype and functional characteristics of mature hepatocytes. The differentiation system of human adipose mesenchymal stem cells into hepatocytes induced by small-molecule compounds in vitro was successfully constructed. This study will lay a foundation for the optimization of liver differentiation strategies and provide a reliable source of functional liver cells for clinical studies of liver diseases.

cell biology↗

bioRxiv2020

Generation of cardiac valve endocardial like cells from human pluripotent stem cells

The cardiac valvular endothelial cells (VECs) are an ideal cell source that could be used for the fabrication of the next generation tissue-engineered cardiac valves (TEVs). However, few studies have been focused on the derivation of this important cell type. Here we describe a chemically defined xeno-free method for generating VEC-like cells from human pluripotent stem cells (hPSCs) through an intermediate endocardium stage. HPSCs were initially specified to KDR+/ISL1+ multipotent cardiac progenitor cells (CPCs), followed by differentiation into endocardial progenitors under the combined treatment with VEGFA, BMP4 and bFGF. In the presence of VEGFA, BMP4 and TGFb, valve endocardial progenitor cells (VEPs) were efficiently derived from endocardial progenitors without a sorting step. Mechanistically, administration of TGFb and BMP4 may specify the VEP fate by facilitating the expression of key transcription factors ETV2 and NFATc1 at the immediate early stage and by activating Notch signaling at the later stage. Notch activation is likely an important part of VEP induction. HPSC-derived VEPs exhibited morphological, molecular and functional similarities to that of the primary VECs isolated from normal human aortic valves. When hPSC-derived VEPs were seeded onto the surface of the de-cellularized porcine aortic valve (DCV) matrix scaffolds, they exhibited higher proliferation and survival potential than the primary VECs. Our results suggest that hPSC-derived VEPs could serve as as a potential platform for the study of valve development, and as starting materials for the construction of the next generation TEVs. HighlightsO_LIValve endocardial progenitor cells (VEPs) could be efficiently derived from hPSCs without a sorting step C_LIO_LIThe combined treatment with TGFb and BMP4 induce VEP fate by enhancing the expression of ETV2 and NFATc1 C_LIO_LIHPSC-derived VEPs resemble the isolated primary VECs molecularly, morphologically and functionally C_LIO_LIHPSC-derived VEPs exhibit proliferative and functional potential similar to the primary VECs when seeded onto the DCVs C_LI

developmental biology↗

bioRxiv2020

Genomic and protein structure modelling analysis depicts the origin and infectivity of 2019-nCoV, a new coronavirus which caused a pneumonia outbreak in Wuhan, China

Detailed genomic and structure-based analysis of a new coronavirus, namely 2019-nCoV, showed that the new virus is a new type of bat coronavirus and is genetically fairly distant from the human SARS coronavirus. Structure analysis of the spike (S) protein of this new virus showed that its S protein only binds weakly to the ACE2 receptor on human cells whereas the human SARS coronavirus exhibits strongly affinity to the ACE receptor. These findings suggest that the new virus does not readily transmit between humans and should theoretically not able to cause very serious human infection. These data are important to guide design of infection control policy and inform the public on the nature of threat imposed by 2019-nCov when results of direct laboratory tests on this virus are not expected to be available in the near future.

microbiology↗

bioRxiv2019

Localizing components of shared transethnic genetic architecture of complex traits from GWAS summary data

Despite strong transethnic genetic correlations reported in the literature for many complex traits, the non-transferability of polygenic risk scores across populations suggests the presence of population-specific components of genetic architecture. We propose an approach that models GWAS summary data for one trait in two populations to estimate genome-wide proportions of population-specific/shared causal SNPs. In simulations across various genetic architectures, we show that our approach yields approximately unbiased estimates with in-sample LD and slight upward-bias with out-of-sample LD. We analyze 9 complex traits in individuals of East Asian and European ancestry, restricting to common SNPs (MAF > 5%), and find that most common causal SNPs are shared by both populations. Using the genome-wide estimates as priors in an empirical Bayes framework, we perform fine-mapping and observe that high-posterior SNPs (for both the population-specific and shared causal configurations) have highly correlated effects in East Asians and Europeans. In population-specific GWAS risk regions, we observe a 2.8x enrichment of shared high-posterior SNPs, suggesting that population-specific GWAS risk regions harbor shared causal SNPs that are undetected in the other GWAS due to differences in LD, allele frequencies, and/or sample size. Finally, we report enrichments of shared high-posterior SNPs in 53 tissue-specific functional categories and find evidence that SNP-heritability enrichments are driven largely by many low-effect common SNPs.

genomics↗

bioRxiv2018

The N-end rule E3 ligase UBR2 activates Nlrp1b inflammasomes

Innate immunity relies on the formation of different inflammasomes to initiate immune responses. The recognition of diverse infection and other danger signals by innate immune receptors trigger caspase-1 activation that induces pyroptosis. Anthrax lethal factor (LF) is a secreted bacterial protease that known to potently activate Nlrp1b inflammasomes in mouse macrophages, but the molecular mechanism underlying LF-induced Nlrp1b activation remains unknown. We here carried out both a mouse genome-wide siRNA screen and a CRISPR/Cas9 knockout screen seeking to identify genes that participate in Nlrp1b activation triggered by LF treatment. We found that the N-end rule pathway E3 ligase UBR2 is required for Nlrp1b activation and a ubiquitin conjugating E2 enzyme E2O is also involved in this process via its physically interaction with UBR2. We show that LF triggers activation of Nlrp1b by initiating the degradation of the N-terminal fragment of Nlrp1b itself that produced via an auto-cleavage process. This study deepens our understanding of innate immunity defense against bacterial infection by elucidating the functional role of UBR2-mediated N-end rule pathway in LF-induced Nlrp1b activation.

immunology↗