Biology subjects

Duan, H.

Publications and source records attributed to Duan, H..

At least 37 records · Page 2Linked to original sources

bioRxiv2025

Vaccine Elicitation of HIV-1 Neutralizing Antibodies Against Both V2 Apex and Fusion Peptide in Rhesus Macaques

Broadly neutralizing antibodies targeting multiple sites of HIV-1 Env vulnerability can be induced by infection, but simultaneous elicitation of neutralizing antibodies (NAbs) against multiple epitopes has not yet been achieved by vaccination. In this study, we designed a dual-epitope vaccine targeting both fusion peptide (FP) and V2 apex and evaluated its capacity to induce NAbs against both epitopes in rhesus macaques. This vaccine combined an FP conjugate with a cocktail of engineered Env trimers with enhanced V2 apex recognition and increased antigen retention in lymph nodes. Immunization of macaques with the dual-epitope vaccine elicited >1000-fold higher autologous tier 2-neutralization titers than the wildtype Env trimer and enhanced heterologous NAb breadth. Both FP and V2-apex monoclonal antibodies (mAb) were isolated from immunized macaques and showed heterologous neutralization with genetic and structural signatures that were similar to well-characterized FP and V2 apex bNAbs, although the V2 apex mAbs showed incomplete maturation. These results demonstrate proof-of-concept for simultaneous vaccine elicitation of NAbs against multiple sites of Env vulnerability, which will likely be critical for an effective HIV-1 vaccine. HIGHLIGHTSO_LIDesigned a dual-epitope vaccine targeting both fusion peptide (FP) and V2 apex C_LIO_LIV2-SET Env trimer conferred higher binding of V2 apex bNAbs, longer retention in draining lymph nodes, and >1000-fold higher induction of autologous neutralization titers compared with wildtype Env trimer C_LIO_LIDual-epitope vaccine enhanced serum tier 2 neutralization breadth C_LIO_LIDual-epitope vaccine elicited FP- and V2 apex-specific neutralizing mAbs with genetic signatures similar to well-characterized FP and V2 apex bNAbs C_LI

immunology↗

bioRxiv2025

AI-Guided Design of Cyclic Peptide Binders Targeting TREM2 Using CycleRFdiffusion and Experimental Validation

Triggering receptor expressed on myeloid cells 2 (TREM2) plays a central role in regulating microglial function in the central nervous system and has emerged as a promising therapeutic target for Alzheimers disease. Despite advances in antibody-based therapeutics, small molecules and peptides capable of modulating TREM2 remain limited. Here, we present a cyclic peptide design pipeline that integrates CycleRFdiffusion, ProteinMPNN for sequence design, and HighFold for structural prediction and screening. Using the TREM2 structure as input, we generated and screened 1,500 peptide-target complexes, prioritizing four candidates that met structural and energetic criteria. Subsequent biophysical evaluation identified TP4 as a TREM2 binder, demonstrating consistent binding in spectral shift, microscale thermophoresis, and surface plasmon resonance. Pharmacokinetic profiling indicated that TP4 possesses favorable plasma stability and moderate metabolic stability, supporting its tractability for further optimization. This study establishes a generalizable framework for AI-driven cyclic peptide discovery and provides the first proof-of-concept demonstration of TREM2-targeted cyclic peptide binders.

pharmacology and toxicology↗

bioRxiv2025

Preparation of Scutellariae Radix and Corn-Soy Peptides Complex and Its Hepatoprotective Mechanisms

Amidst the growing global burden of alcohol-associated liver disease (ALD) and the limited clinical interventions, this study innovatively developed a complex of Scutellariae Radix and corn-soy composite peptides (SR-CP). SR-CP exhibited significantly enhanced bioaccessible antioxidant capacity, demonstrating a 0.84-2.89-fold increase in DPPH, ABTS, {middle dot}OH, O2-{middle dot}, and FRAP value versus individual components post-gastrointestinal digestion. In ethanol-injured BRL-3A hepatocytes, SR-CP (60 g/mL) significantly ameliorated cellular viability by 38.28% (vs. ethanol model), normalized ALT/AST leakage by 19.92%/87.35%, restored SOD/GSH antioxidant reserves to 24.74 {+/-} 1.71 U/mg prot and 55.35 {+/-} 1.64 mol/g prot respectively, while suppressing lipid peroxidation through 65.93% reduction in MDA production (vs. ethanol model). Mechanistic profiling revealed dual regulatory effects: 1) Inhibiting CYP2E1/NADPH-OX ({downarrow} 51.97%/80.52%) to prevent excessive ROS production and activating the AMPK/SIRT1 axis ({uparrow} 30.70%/6.99%); 2) Potentiation of Nrf2 nuclear translocation ({uparrow} 0.48-fold) and HO-1 induction ({uparrow} 1.63-fold) through AMPK-mediated antioxidant signaling. The superior hepatoprotection of SR-CP over single constituents establishes a novel paradigm for functional food development, bridging traditional herbal wisdom with molecular-targeted ALD intervention via the AMPK/Nrf2/HO-1 axis.

pharmacology and toxicology↗

bioRxiv2025

Multi-omics analyses reveal the effect of DNA methylation on iridoid glycosides biosynthesis in Rehmannia glutinosa

Rehmannia glutinosa roots produce a group of lipophilic bioactive components known as iridoid glycosides. However, the molecular mechanisms by which DNA methylation regulates the biosynthesis of iridoid glycosides in R. glutinosa remain unknown. Herein, the development of R. glutinosa roots and the content of iridoid glycosides in the Wenxian region were significantly higher than those in Xinxiang. Low methylation level contributed to the accumulation of iridoid glycosides and the expression of related enzyme genes. Demethylation promoted both roots growth and development, as well as the accumulation of iridoid glycosides. Up-regulated RgALDH13, RgHDR1, RgG10H4, RgDXR1, RgG10H3, and RgUPD1, along with transcription factors (TFs), form the regulatory network for the biosynthesis of iridoid glycosides. Furthermore, the primary active region of the RgG10H4 promoter is located in the -164 bp region, where the RgMYB2 protein specifically binds to the TAACCA motif in the RgG10H4 promoter. Collectively, low levels of DNA methylation enhance the expression of core genes, followed by inducing the accumulation of iridoid glycosides, which suggests that RgMYB2-RgG10H4 plays a positive role in this process. These findings will contribute to a deeper understanding of the role of DNA methylation in the accumulation of iridoid glycosides. HighlightLow levels of DNA methylation contribute to the accumulation of iridoid glycosides and the expression of key enzyme genes in R. glutinosa.

molecular biology↗

bioRxiv2025

PPIKB: A Comprehensive Knowledge Base and Analysis Platform for Protein Peptide Interactions Based on Literature and Patents

Protein-protein interactions (PPIs) underpin a multitude of fundamental biological processes and represent critical therapeutic targets. Peptide compounds--owing to their structural flexibility, high target specificity, and the combined advantages of small molecules and biologics--have emerged as powerful modulators of these interactions. However, the exponential growth of scientific publications and patent filings poses a significant challenge for the systematic integration of protein-peptide interaction (PPepI) data. To address this, we implemented a high-throughput extraction pipeline powered by large language models (LLMs) to harvest affinity data from both research articles and patents, followed by rigorous manual curation. Applying this approach to 4,241 research papers and 483 patents published between 1972 and 2025, we compiled 21,760 manually validated quantitative affinity entries encompassing 2,013 unique protein targets derived from human, mouse, rat, viral, bacterial, and other species. In parallel, we consolidated 18,005 structural entries of protein-peptide complexes from the Protein Data Bank and associated literature. Leveraging these comprehensive datasets, we developed PPIKB, a user-friendly database and analysis platform that enables efficient querying, visualization, and in-depth analysis of protein-peptide interactions. PPIKB also provides a suite of tools tailored to facilitate peptide-based drug discovery, protein engineering, and systems-level biological research. By combining state-of-the-art text-mining technologies with expert validation, PPIKB sets a new standard for accessibility and utility in the field of protein-peptide interaction data resources. Access PPIKB at: https://ppikb.duanlab.ac

bioinformatics↗

bioRxiv2025

AlphaFold3 for Non-canonical Cyclic Peptide Modeling: Hierarchical Benchmarking Reveals Accuracy and Practical Guidelines

Despite the revolutionary impact of AlphaFold3 on structural biology, this models capability in predicting non-canonical cyclic peptides remains unexplored. Given the clinical significance of cyclic peptides containing unnatural residues as a therapeutic modality, we present the first systematic evaluation of AlphaFold3 for this class of molecules. To facilitate benchmarking, we developed an automated input pipeline to streamline AlphaFold3 predictions for cyclic peptides. Our study aims to: (1) quantify the hierarchical accuracy (all-atoms, C atoms, and atoms of unnatural residue levels) of AlphaFold3 in predicting both non-canonical cyclic peptide monomers and complexes; (2) assess the reliability of AlphaFold3s confidence metrics; (3) evaluate the influence of multiple sequence alignment (MSA) and structural templates; and (4) identify systematic biases in AlphaFold3s predictions. Based on these analyses, we provide practical guidelines for applying AlphaFold3 in cyclic peptide structure prediction for facilitating the related research of bioactive cyclic peptides.

bioinformatics↗

bioRxiv2025

MeDCycFold: A Rosetta Distillation Model to Accelerate Structure Prediction of Cyclic Peptides with Backbone N-methylation and D-amino Acids

Cyclic peptides with backbone N-methylated amino acids(BNMeAAs) and D-amino acids(D-AAs) have gained attention for their stability, membrane permeability, and other therapeutic potentials. Currently, Rosetta can predict their structures using energy calculations, but this method is heavily time-consuming. Moreover, structural data for cyclic peptides containing BNMeAAs and D-AAs are extremely insufficient to build a data-driven structure prediction model. To address these problems, we propose MeDCycFold, a deep learning-based Rosetta distillation model by fine-tuning the AlphaFold model. First, a cyclic peptide structure dataset is constructed using Rosetta by sampling massive conformations for cyclic peptides with BNMeAAs and D-AAs and evaluating their energy scores. Then, the AlphaFold model is fine-tuned with the extended 56 BNMeAAs and D-AAs. Besides, a relative position cyclic matrix is introduced for head-to-tail cyclization in the cyclic peptides. Finally, a force field is employed to reduce clashes in the predicted structures. Empirical experiments show that our proposed MeDCycFold speeds up structure prediction by 49 times while maintaining the prediction accuracy comparable to Rosetta, which can greatly accelerate the development of cyclic peptide drugs.

bioinformatics↗

bioRxiv2025

Early-life gut microbiome and stress-axis perturbations dysregulate systemic, mucosal, and brain immunity

BackgroundEarly-life disruptions to the gut microbiome and stress-axis significantly influence the development of immune, neuroendocrine, and other physiological systems. However, the precise microbial species and pathways mediating these effects remain poorly characterized. Using a murine model, we investigated the individual and combined effects of early-life antibiotic exposure and chronic stress on gut microbiota composition, short-chain fatty acid (SCFA) production, hypothalamic-pituitary-adrenal (HPA) axis activity, and systemic, mucosal, and neuroimmune responses. ResultsBroad-spectrum antibiotic treatments severely reduced microbial diversity and SCFA concentrations, with changes persisting into adulthood. Chronic early-life stress exerted more modest but notable effects, reducing key SCFA-producing taxa and impacting microbiome metabolic output. Combined disruptions led to altered microglial active phenotype and cytokine profiles, impaired immune cell populations, and suppressed HPA axis activity. Multi-omic correlational analyses revealed strong associations between SCFAs, specific gut microbes, and immune responses, implicating SCFAs as critical mediators of gut-brain communication. Notably, antibiotic exposure exacerbated susceptibility to allergic airway inflammation, highlighting the systemic consequences of early-life microbiome disturbances. ConclusionsThese findings demonstrate that early microbial perturbations impair neuroimmune maturation, HPA axis regulation, and host resilience to inflammatory diseases. Our study underscores the importance of preserving the early-life microbiome to support long-term immune and neurodevelopmental health, offering insights into potential therapeutic interventions for mitigating the impact of early-life microbiota disruptions.

microbiology↗

bioRxiv2025

HighPlay：Cyclic Peptide Sequence Design Based on Reinforcement Learning and Protein Structure Prediction

The structural diversity and good biocompatibility of cyclic peptides has led to their emergence as potential therapeutic agents. Traditional cyclic peptide design relies on natural template modification and combinatorial chemical library screening, but suffers from bottlenecks such as limited molecular diversity, high cost, and time-consuming.AI technologies have improved design efficiency by predicting target binding patterns and generating scaffold structures, but they still require extensive laboratory screening. In this study, we propose HighPlay, which integrates reinforcement learning (Monte Carlo Tree Search) with the HighFold structure prediction model to design cyclic peptide sequences for protein targets, dynamically exploring the sequence space without the need of predefined target information. The model was applied to the design of cyclic peptide sequences for three different targets, which were screened and verified by molecular dynamics simulation, and showed good binding affinity. Specifically, the cyclic peptide sequences designed for TEAD4 target showed micromolar-level affinity in further experimental validation.

bioinformatics↗

bioRxiv2025

Accurate structure prediction of cyclic peptides containing unnatural amino acids using HighFold3

In recent years, cyclic peptides have emerged as a research hotspot in drug development due to their excellent stability, high specificity, and potential to penetrate intracellular targets. However, existing computational models still face challenges in making accurate predictions of the structures of cyclic peptides containing unnatural amino acids (unAAs), thereby limiting their application in drug design. The release of AlphaFold 3 (AF3) has significantly improved the modeling capability of biomolecular complexes and supports the definition of unAAs through residue modifications using the CCD database (Chemical Component Dictionary). Yet, its reliance on structures already present in the training library limits its ability to accurately predict cyclic peptide structures. Base on the AlphaFold 3 framework, we developed HighFold3, which comprises two sub-models: HighFold3-Linear and HighFold3-Cyclic, designed for predicting the structures of linear and cyclic peptides, respectively. Our research results show that, by incorporating a cyclic peptide positional preference matrix, HighFold3 achieves superior performance over other models (HighFold, HighFold2, and CyclicBoltz1) in predicting cyclic peptide structures, particularly excelling in complex cyclic peptides containing unAAs. HighFold3 provides an efficient tool for cyclic peptide drug design, promising to advance the application of cyclic peptides in targeted therapy, antibacterial treatments, and cellular penetration.

bioinformatics↗

bioRxiv2025

Predicting the structures of cyclic peptides containing unnatural amino acids by HighFold2

Cyclic peptides containing unnatural amino acids possess many excellent properties and have become promising candidates in drug discovery. Therefore, accurately predicting the three-dimensional structures of cyclic peptides containing unnatural residues will significantly advance the development of cyclic peptide-based therapeutics. Although deep learning-based structural prediction models have made tremendous progress, these models still cannot predict the structures of cyclic peptides containing unnatural amino acids. To address this gap, we introduce a novel model, HighFold2, built upon the AlphaFold-Multimer framework. HighFold2 first extends the pre-defined rigid groups and their initial atomic coordinates from natural amino acids to unnatural amino acids, thus enabling structural prediction for these residues. Then, it incorporates an additional neural network to characterize the atom-level features of peptides, allowing for multi-scale modeling of peptide molecules while enabling the distinction between various unnatural amino acids. Besides, HighFold2 constructs a relative position encoding matrix for cyclic peptides based on different cyclization constraints. Except for training using spatial structures with unnatural amino acids, HighFold2 also parameterizes the unnatural amino acids to relax the predicted structure by energy minimization for clash elimination. Extensive empirical experiments demonstrate that HighFold2 can accurately predict the three-dimensional structures of cyclic peptide monomers containing unnatural amino acids and their complexes with proteins, with the median RMSD for C reaching 1.891 [A]. All these results indicate the effectiveness of HighFold2, representing a significant advancement in cyclic peptide-based drug discovery.

bioinformatics↗

bioRxiv2025

5' UTR-mediated retention of eIF3 on 80S ribosomes promotes co-translational folding of ER membrane proteins

Co-translational folding of nascent polypeptides is essential for protein function and cellular homeostasis. Ribosome-associated chaperones assist in this process, but coordination between their recruitment and translation initiation remains poorly understood. We report here that specific binding sites for eukaryotic translation initiation factor eIF3 within the 5 untranslated regions (5UTRs) of mRNAs promote its retention on 80S ribosomes during the synthesis of select endoplasmic reticulum (ER) membrane proteins. Disruption of these eIF3 binding sites leads to misfolding and sequestration of newly synthesized membrane proteins into ER whorls. Sequestration into ER whorls is exacerbated by HSP70 inhibition but can be rescued by overexpressing HSPA8 and HSPA1. Cross-linking assays reveal that 5UTR binding sites stabilize eIF3-80S interactions during early elongation, facilitating recruitment of HSPA8 to ribosomes. These findings indicate that genetic instructions within 5UTRs direct eIF3-mediated chaperone recruitment, ensuring proper co-translational folding of ER membrane proteins.

cell biology↗

bioRxiv2024

PepMSND: Integrating Multi-level Feature Engineering and Comprehensive Databases to Enhance in vivo/in vitro Peptide Blood Stability Prediction

Deep learning technology has revolutionized the field of peptides, but key questions such as how to predict the blood stability of peptides remain. While such a task can be accomplished by experiments, it requires much time and cost. Here, to address this challenge, we collect extensive experimental data on peptide stability in blood from public databases and literature and construct a database of peptide blood stability that includes 635 samples. Based on this database, we develop a novel model called PepMSND, integrating KAN, Transformer, GAT and SE(3)-Transformer to make multi-level feature engineering to make peptide stability prediction. Our model can achieve the ACC of 0.8672 and the AUC of 0.9118 on average and outperforms the baseline models. This work can facilitate the development of novel peptides with strong stability, which is crucial for their therapeutic use in clinical applications.

bioinformatics↗

bioRxiv2024

NCPepFold: Accurate Prediction of Non-canonical Cyclic Peptide Structures via Cyclization Optimization with Multigranular Representation

Artificial intelligence-based peptide structure prediction methods have revolutionized biomolecular science. However, restricting predictions to peptides composed solely of 20 natural amino acids significantly limits their practical application, as such peptides often demonstrate poor stability under physiological conditions. Here, we present NCPepFold, a computational approach that can utilize a specific cyclic position matrix to directly predict the structure of cyclic peptides with non-canonical amino acids. By integrating multi-granularity information at the residue- and atomic-level, along with fine-tuning techniques, NCPepFold significantly improves prediction accuracy, with the average peptide RMSD for cyclic peptides being 1.640 [A]. In summary, this is a novel deep learning model designed specifically for cyclic peptides with non-canonical amino acids without length restrictions, offering great potential for peptide drug design and advancing biomedical research.

bioinformatics↗

bioRxiv2024

Cycledesigner Leveraging RFdiffusion and HighFold to Design Cyclic Peptide Binders for Specific Targets

Cyclic peptides are potentially therapeutic in clinical applications, due to their great stability and activity. Yet, designing and identifying potential cyclic peptide binders targeting specific targets remains a formidable challenge, entailing significant time and resources. In this study, we modified the powerful RFdiffusion model to allow the cyclic peptide structure identification and integrated it with ProteinMPNN and HighFold to design binders for specific targets. This innovative approach, termed cycledesigner, was followed by a series of scoring functions that efficiently screen. With the combination of effective cyclic peptide design and screening, our study aims to further broaden the scope of cyclic peptide binder design.

bioinformatics↗

bioRxiv2024

MetaLab Platform Enables Comprehensive DDA and DIA Metaproteomics Analysis

Metaproteomics studies the collective protein composition of complex microbial communities, providing insights into microbial roles in various environments. Despite its importance, metaproteomic data analysis is challenging due to the datas large and heterogeneous nature. While Data-Independent Acquisition (DIA) mode enhances proteomics sensitivity, it traditionally requires Data-Dependent Acquisition (DDA) results to build the library for peptide identification. This paper introduces an updated version of MetaLab, a software solution that streamlines metaproteomic analysis by supporting both DDA and DIA modes across various mass spectrometry (MS) platforms, including Orbitrap and timsTOF. MetaLabs key feature is its ability to perform DIA analysis without DDA results, allowing more experimental flexibility. It incorporates a deep learning strategy to train a neural network model, enhancing the accuracy and coverage of DIA results. Evaluations using diverse datasets demonstrate MetaLabs robust performance in accuracy and sensitivity. Benchmarks from large-scale human gut microbiome studies show that MetaLab increases peptide identification by 2.7 times compared to conventional methods. MetaLab is a versatile tool that facilitates comprehensive and flexible metaproteomic data analysis, aiding researchers in exploring microbial communities functionality and dynamics.

bioinformatics↗

bioRxiv2024

Molecular and physiological characterization of tillering and shade tolerance of dwarf mutants of perennial ryegrass

Tillering and shade tolerance are important traits in turfgrass, influenced by environmental factors, nutrients, and hormones. Shade stress negatively affects tillering. In this study, two dwarf mutants, shadow-1 and shadow-2, developed via Gamma-ray and fast-neutron mutagenesis, respectively, showed significantly higher tillering than the wild-type under greenhouse conditions. Both mutants demonstrated shade tolerance in plant height, grass quality, and color under 85% and 95% shade conditions, while shade-induced inhibition of tillering was observed in both the mutants and the wild-type. In comparison to wild-type plants under 95% shade conditions, we observed that the cytokinin biosynthetic gene IPT8 is upregulated, while the cytokinin inactivating gene CKX2 is downregulated in shadow-1. Similarly, the GA biosynthetic genes CPS1, GA2ox3, and GA20ox1 are upregulated, while the GA inactivating gene GA20ox8 is downregulated in the shadow-1 mutant. Furthermore, the ethylene biosynthetic genes ACS and ACO are also downregulated in the shadow-1 mutant. Consistently, we observed that wild-type plants exhibit increased GA and reduced CK levels, while shadow-1 mutant plants have reduced GA but increased CK levels. This explains the shadow-1 mutants shade tolerance in terms of plant height, grass quality, and color. Conversely, the tillering inhibitor genes CRY1, MAX2, and SnRK1 are upregulated in both wild-type and shadow-1 mutant plants. Our results provide novel insights into the mechanisms behind tillering and shade tolerance in turfgrasses under shade conditions.

plant biology↗

bioRxiv2024

Boosting the Predictive Power of Protein Representations with a Corpus of Text Annotations

Protein language models are trained to predict amino acid sequences from vast protein databases, while learning to represent proteins as feature vectors. These vector representations have enabled impressive applications, from predicting mutation effects to protein folding. One of the reasons offered for the success of these models is that conserved sequence motifs tend to be important for protein fitness. Yet, the relationship between sequence conservation and fitness can be confounded by the evolutionary and environmental context. Should we therefore look to other data sources that may contain more direct functional information? In this work, we conduct a comprehensive study examining the effects of training protein models to predict nineteen types of text annotations from UniProt. Our results show that finetuning protein models on a subset of these annotations enhances the models predictive capabilities on a variety of function prediction tasks. Notably, our model outperforms the search algorithm BLAST, which none of the pre-trained protein models accomplished in our evaluation. Our results suggest that a much wider array of data modalities, such as text annotations, may be tapped to improve protein language models. We host our model checkpoints on https://huggingface.co/h4duan.

biochemistry↗