Biology subjects

Zhai, S.

Publications and source records attributed to Zhai, S..

At least 19 recordsLinked to original sources

bioRxiv2026

Sustained Volumetric Compression Induces Cell Jamming and Primes Breast Cancer Cells for Enhanced Post-Compression Migration and Invasion

During tumor growth and progression, cancer cells are exposed to sustained physical confinement and volumetric compression that can alter cell volume, cytoskeletal organization, mechanotransduction, and invasive behavior. However, whether breast cancer cells retain a compression-induced mechanical memory after release from sustained volumetric compression, and how this memory influences subsequent migration and invasion, remains poorly understood. Here, by controlling cell volume using PEG - mediated volumetric compression, we investigated the compression and post-compression recovery responses of MCF-7 breast cancer cells. Cells were compressed for four days, followed by four days of recovery after PEG removal, and analyzed using daily morphological tracking, single-cell time-lapse imaging, F-actin and YAP staining, wound healing assays, and 3D spheroid invasion assays. We show that sustained volumetric compression shifts MCF-7 cells into a compact, jammed-like, low-motility state characterized by reduced morphodynamic remodeling, suppressed collective migration, and limited spheroid invasion. In contrast, post-compression recovery induces a distinct mechanobiological state marked by increased cell area and perimeter, altered single-cell trajectories, heterogeneous F-actin remodeling, enhanced YAP nuclear localization in enlarged recovered cells, accelerated wound closure, and increased spheroid invasion and cell dissemination. These findings suggest that prior volumetric compression can prime breast cancer cells for enhanced migration and invasion after stress release, supporting post-compression recovery as a form of mechanical memory that may contribute to tumor dissemination.

bioengineering↗

bioRxiv2026

MAERM: Predicting Enzyme-Reaction Matching Relationships with a Mixed-Attention Model

Harnessing enzyme specificity requires a thorough understanding of enzyme promiscuity, which determines enzymes catalytic scope; however, measuring this scope still relies heavily on labor-intensive analytical approaches. While data-driven approaches have emerged to predict the catalytic scope of enzymes, these methods continue to face challenges such as restricted datasets and insufficient integration of enzyme structural information and reaction transformations. Here, we introduce MAERM, an innovative mixed-attention model designed to predict enzyme-reaction matching relationships. Built on our MAERM-DB, a dataset with broad coverage of validated and chemoenzymatic catalysis data, MAERM utilizes a local-global attention module to integrate multimodal enzyme information with fine-grained reaction representations, thereby predicting enzyme-reaction matching probabilities. Results show that MAERM consistently outperforms all baselines, with an average F1-score of 0.984. Notably, on challenging test samples with less than 40% sequence identity to the training set, MAERM outperforms the second-ranked model by 5.9% in F1-score. In addition, MAERM achieves the highest top-10 success rate of 51.7% on Enzyme-405 and the highest balanced accuracy of 0.697 on BioCat-547, further supporting its generalizability in enzyme screening and chemoenzymatic catalysis. Finally, MAERM can serve as an efficient scoring module. When integrated with ProteinMPNN, MAERM has successfully guided novel enzyme design for two carbonyl reduction reactions, resulting in enhanced catalytic potential for the native substrate and demonstrating broad compatibility. Overall, MAERM has the potential to reduce the experimental cost of measuring enzymes catalytic scope, facilitate enzyme design, and ultimately accelerate the design-build-test-learn cycle in enzyme engineering.

bioinformatics↗

bioRxiv2026

Surface-specific film assembly of a Vibrio cholerae adhesin peptide modulated by environmental salts

Underwater adhesion research increasingly draws on bioinspired systems to uncover the molecular mechanisms that enable strong interfacial binding in aqueous environments. The biofilm adhesin Bap1 from Vibrio cholerae contains a short peptide motif, SYWFFGWHTK (CP), which exhibits exceptional adhesive performance, surpassing mussel foot protein mfp5 under comparable conditions. Despite its promise, the roles of ionic environments and aggregation behavior in governing CP adhesion remain unclear. In this study, we investigate how ion identity influences CP aggregation, film formation, and interfacial properties. Using dynamic light scattering, we identify the formation of micron-scale assemblies of aggregated molecular clusters (AAMCs), with size distributions modulated by salt type. Quartz crystal microbalance with dissipation and liquid atomic force microscopy reveal that CP film formation is both surface- and ion-dependent. On gold substrates, AAMCs preferentially adsorb and collapse into rigid, smooth nanofilms, consistent with hydrophobic-driven compaction. In contrast, silicate surfaces inhibit such collapse, yielding distinct morphologies and interfacial energetics. These findings demonstrate that surface chemistry and ionic conditions jointly regulate peptide aggregation and adhesion. This work provides mechanistic insight into hydrophobic-rich peptide systems and informs the rational design of next-generation wet adhesives, with broader implications for biomaterials and peptide-based formulations. GRAPHICAL ABSTRACT O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=129 SRC="FIGDIR/small/733527v2_ufig1.gif" ALT="Figure 1"> View larger version (39K): org.highwire.dtl.DTLVardef@1f3f3cborg.highwire.dtl.DTLVardef@10aae0corg.highwire.dtl.DTLVardef@289970org.highwire.dtl.DTLVardef@171b026_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗

bioRxiv2026

High-PepBinder: A pLM-Guided Latent Diffusion Framework for Affinity-Aware Target-Specific Peptide Design

Peptides, as therapeutic molecules, offer unique advantages in targeting complex protein surfaces, yet their rational design remains limited by the vastness of the sequence space and the constraints of traditional approaches. Here, we propose High-PepBinder, a sequence-only conditional diffusion framework for target-specific peptide generation. Guided by the target protein sequence, High-PepBinder adopts a dual encoder architecture that integrates protein language models (pLMs) with the diffusion model. This approach cascades the peptide generation model with an affinity classifier and enables the generation process to capture affinity-related features of the peptides through lightweight joint optimization. Due to the scarcity of protein-peptide affinity data, we constructed PepPBA, to our knowledge the most comprehensive dataset to date, and established a structure- and physics-based screening pipeline to prioritize top candidates. Results show that High-PepBinder demonstrates competitive performance across multiple peptide generation and affinity-related tasks. For representative targets, including KEAP1, XIAP, and EGFR, the generated peptides preserve key binding geometries and interface patterns of reference peptides in predicted complexes, while maintaining sequence diversity and favorable predicted properties. Overall, High-PepBinder contributes toward a general and sequence-only strategy for peptide design, offering a computational framework for expanding peptide discovery against challenging targets.

bioinformatics↗

bioRxiv2026

Protein-peptide Interaction Representation Learning with Pretrained Language Models

Protein-peptide Interactions (PpIs) paly essential roles in diverse cellular processes, yet their systematic identification remains challenging due to the limited availability of experimentally annotated protein-peptide interaction data. To address this challenge, we present PepInter, a sequence-based Deep Learning (DL) framework that leverages large-scale pretraining on structurally derived pseudo protein-peptide pairs to learn interaction-relevant representations. Specifically, energy-dominant peptide fragments are extracted from protein complexes curated from non-redundant Protein Data Bank (PDB) structures, enabling the construction of pseudo protein-peptide interaction pairs that capture interface interaction patterns shared with canonical protein-protein interactions. This strategy allows the model to acquire interaction-aware priors in the absence of large-scale annotated protein-peptide complex datasets. Built upon the ESM-Cambrian (ESMC) architecture, PepInter adopts a two-stage pretraining strategy. In the first stage, masked language modeling is used to learn general protein sequence representations. In the second stage, the model is further trained to predict Rosetta-derived energetic scores, explicitly incorporating structural interaction signals into the learned embeddings. Following pretraining, PepInter is fine-tuned for both protein-peptide interaction classification and peptide bioactivity regression tasks. Across multiple benchmark datasets, including protein-peptide binding affinity prediction, PepInter consistently outperforms existing baseline methods and demonstrates strong generalization in identifying biologically meaningful PpIs. Case studies further highlight its ability to recover known interaction patterns and predict novel protein-peptide interactions. Together, these results establish PepInter as a scalable and effective framework for protein-peptide interaction prediction, with strong potential to accelerate peptide-based drug discovery.

bioinformatics↗

bioRxiv2025

Vibrio cholerae adhesin-derived peptide mediates strong pull-off forces in aqueous high ionic strength environments

In this letter, the pull-off forces of adsorbed films of four Bap1-inspired peptides in various solvents were investigated on negatively charged mica substrates using the surface forces apparatus (SFA), complemented with dynamic light scattering (DLS) for characterizing the aggregation behavior of peptides in solution. Bap1-inspired peptides consisted of the 57 amino acid wild-type sequence (WT); a scrambled version of the WT used to investigate the impact of the primary amino acid sequence in pull-off forces (Scr); a ten amino acid sequence rich in hydrophobic content (CP) of the WT sequence, and an eight amino acid sequence (Sh1) that corresponds to the pseudo-repeating sequence in the 57 AA. SFA results showed remarkable pull-off forces for CP, particularly in the presence of salts: measured pull-off forces were 26.0 {+/-} 7.0 mN/m for no dwell-time and up to 42.0 {+/-} 8.8 mN/m when surfaces were left in contact for 30 minutes. DLS observations indicate that salts favor large peptide aggregation for all constructs (Hz > 1 {micro}m), as compared to milliQ (Hz {approx} 100-500 nm) water and DMSO (Hz {approx} 100 nm), resulting in heterogeneous peptide film thicknesses. This letter concludes with a comparison to the pull-off forces of mussel foot protein-inspired peptides reported in the literature.

biophysics↗

bioRxiv2025

Chemical activity profiling reveals how exposure to drugs or dietary compounds alters gut microbial biotransformation capacity

The gut microbiome catalyzes biotransformation reactions that influence intestinal absorption as a basis of microbiome-host interactions. A better understanding of microbiota biotransformation capacity, and its alteration in dysregulated states, would enable the prediction of individual responses to drugs and toxins and improve safety assessment. Here, we profiled chemical activities in rat gut microbiota ex vivo, and quantified biotransformation capacity changes induced by oral exposure to eight drugs and dietary compounds: tobramycin, colistin, acesulfame potassium, saccharin, bovine serum albumin (BSA), meropenem, doripenem, and daidzein. We implemented an approach involving inoculation with metabolic probes during microbiota fermentations and measured their degradation. In most exposure groups, we observed no alteration of microbiota biotransformation capacity, however, we detected significant alterations in biotransformation rates after exposure to meropenem, doripenem and tobramycin. Interestingly, common patterns of biotransformation capacity were observed in the gut microbiomes from rats exposed to carbapenems and partially shared in microbiomes exposed to tobramycin. These results aligned well with prior metagenomic and metabolomic findings. Further, correlations between microbial taxa and reaction rates were assessed to establish a link between specific bacteria and probe degradation. This functionally relevant strategy revealed alterations of microbiota biotransformation capacity, as induced by in vivo exposures.

microbiology↗

bioRxiv2025

PepPCBench is a Comprehensive Benchmark for Protein-Peptide Complex Structure Prediction with AlphaFold3

Peptides mediate up to 40% of protein-protein interactions (PPIs), offering high specificity and the ability to target binding sites inaccessible to small molecules, making them promising candidates for drug development. Accurate modeling of protein-peptide complexes is crucial for understanding fundamental biological processes and for advancing peptide-based drug design. However, due to the high conformational flexibility of peptides, predicting their interactions with proteins remains a significant challenge. Recent advancements in artificial intelligence (AI), exemplified by all-atom protein folding neural networks (PFNNs) such as AlphaFold3 (AF3), have expanded predictive capabilities beyond proteins to encompass protein-peptide complexes. Nevertheless, existing evaluations of these methods are often limited in scope and lack systematic, multi-dimensional, and fair comparisons of PFNN performance in protein-peptide complex prediction. Here, we introduce PepPCBench, a comprehensive benchmark framework specifically developed to evaluate AF3s capabilities in predicting protein-peptide complexes. This study utilizes a carefully curated dataset named PepPCSet, which is excluded from the AF3s training or validation sets. This dataset includes 261 protein-peptide complexes with peptide lengths ranging from 5 to 30 residues. Our benchmark results indicate that AF3 outperforms other PFNNs in prediction accuracy and structural validation. However, its performance remains insufficient for practical peptide drug discovery, indicating room for improvement. It is expected that PepPCBench can provide some valuable insights into the enhancement of protein-peptide complex structure prediction and the development of peptide-based therapeutics. The dataset and pipeline protocols are available at https://github.com/zhaisilong/PepPCBench.

bioinformatics↗

bioRxiv2025

The GSA Family in 2025: A Broadened Sharing Platform for Multi-Omics and Multimodal Data

The Genome Sequence Archive Family (GSA family) provides a suite of database resources for archiving, retrieving, and sharing multi-omics data for the global academic and industrial communities. It currently comprises four distinct database members: the Genome Sequence Archive (GSA, https://ngdc.cncb.ac.cn/gsa/), the Genome Sequence Archive for Human (GSA-Human, https://ngdc.cncb.ac.cn/gsa-human/), the open Archive for Miscellaneous Data (OMIX, https://ngdc.cncb.ac.cn/omix/), and the Open Biomedical Imaging Archive (OBIA, https://ngdc.cncb.ac.cn/obia). Compared to the 2021 version, the GSA family has significantly expanded by introducing a new database member, the OBIA, and through comprehensive upgrades to the existing ones. Notable enhancements to the existing members include broadening the range of accepted data types, enhancing the quality control systems, fine-tuning the data retrieval system, and refining the data-sharing management mechanisms.

genomics↗

bioRxiv2025

State-dependent modulation of spiny projection neurons controls levodopa-induced dyskinesia in a mouse model of Parkinson's disease

In the later stages of Parkinsons disease (PD), patients often manifest levodopa-induced dyskinesia (LID), compromising their quality of life. The pathophysiology underlying LID is poorly understood, and treatment options are limited. To move toward filling this treatment gap, the intrinsic and synaptic changes in striatal spiny projection neurons (SPNs) triggered by the sustained elevation of dopamine (DA) during dyskinesia were characterized using electrophysiological, pharmacological, molecular and behavioral approaches. Our studies revealed that the intrinsic excitability and functional corticostriatal connectivity of SPNs in dyskinetic mice oscillate between the on- and off-states of LID in a cell- and state-specific manner. Although triggered by levodopa, these rapid oscillations in SPN properties depended on both dopaminergic and cholinergic signaling. In a mouse PD model, disrupting M1 muscarinic receptor signaling specifically in iSPNs or deleting its downstream signaling partner CalDAG-GEFI blunted the levodopa-induced oscillation in functional connectivity, enhanced the beneficial effects of levodopa and attenuated LID severity.

neuroscience↗

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

Cell- and state-specific plasticity of striatal glutamatergic synapses is critical to the expression of levodopa-induced dyskinesia

Levodopa-induced dyskinesia (LID) is a debilitating complication of symptomatic therapy in Parkinsons disease (PD). Although there is compelling evidence that striatal pathophysiology is a major driver of LID, the circuit-specific mechanisms contributing to dysfunction remain obscure. This lack of clarity is reflected in the limited options for diminishing established LID. To address this gap, molecular, cellular, and behavioral strategies were used to interrogate striatal indirect pathway spiny projection neurons (iSPNs) in a mouse model of LID. These studies revealed that LID induction led to an up-regulation of GluN2B-containing N-methyl-d-aspartate receptors (NMDARs) specifically at iSPN glutamatergic synapses. This up-regulation was correlated with increased numbers of silent glutamatergic synapses in the hours after levodopa treatment. In this off-state, long-term potentiation (LTP) of iSPN glutamatergic synapses was readily induced and this induction was blocked by antagonists of adenosine type 2 receptors (A2aRs) or GluN2B-containing NMDARs. Systemic administration of the A2aR antagonist tozedenant at the beginning of the off-state significantly reduced the development of LID. More importantly, specifically knocking down the expression of GRIN2B mRNA in iSPNs dramatically attenuated both development and expression of LID, without compromising the beneficial effects of levodopa on movement. Taken together, these studies demonstrate that dyskinesiogenic doses of levodopa trigger cell-specific synaptic adaptations during the off-state that make an important contribution to the network pathophysiology underlying LID and suggest that targeting GluN2B-containing NMDARs in iSPNs could be therapeutically useful.

neuroscience↗

bioRxiv2024

Hypergraph Cortical Cytoarchitectonic Parcellation with Multimodal Canine Brain Atlas

Brain atlases are vital tools in exploring the brain structure-function relationship. The burgeoning cross-species atlases have significantly accelerated our understanding of human brain development, evolution, function, and diseases. However, the existing coarse-grained macroscopic canine brain atlases greatly constrain their utility as an animal model for neurocognition research. Finer-grained brain atlas and partitions are crucial for decoding brain spatial heterogeneity and topology at different scales. Therefore, we conduct macroscopic and microscopic brain imaging to construct an interactive online dataset of multimodal canine brain atlas. Additionally, we develop a pioneering method for cortical cytoarchitectonic partitioning based on hypergraph learning. By integrating high-dimensional cytoarchitectonic features and spatial connections between cortical columns, the method leads to fine-grained partitioning patterns. This innovative approach aims to decode the biological heterogeneity of cortical microstructures, contributing to the structural annotation of canine atlas as well as public human brain atlases. The study not only offers valuable resources but also presents a novel zonation approach to investigate the cellular organization pattern and topology of the cortex.

neuroscience↗

bioRxiv2024

Ca2+-dependent phosphodiesterase 1 regulates the plasticity of striatal spiny projection neuron glutamatergic synapses

Long-term synaptic plasticity at glutamatergic synapses on striatal spiny projection neurons (SPNs) is central to learning goal-directed behaviors and habits. Although considerable attention has been paid to the mechanisms underlying synaptic strengthening and new learning, little scrutiny has been given to those involved in the attenuation of synaptic strength that attends suppression of a previously learned association. Our studies revealed a novel, non-Hebbian, long-term, postsynaptic depression of glutamatergic SPN synapses induced by interneuronal nitric oxide (NO) signaling (NO-LTD) that was preferentially engaged at quiescent synapses. This form of plasticity was gated by local Ca2+ influx through CaV1.3 Ca2+ channels and stimulation of phosphodiesterase 1 (PDE1), which degraded cyclic guanosine monophosphate (cGMP) and blunted NO signaling. Consistent with this model, mice harboring a gain-of-function mutation in the gene coding for the pore-forming subunit of CaV1.3 channels had elevated depolarization-induced dendritic Ca2+ entry and impaired NO-LTD. Extracellular uncaging of glutamate and intracellular uncaging of cGMP suggested that this Ca2+-dependent regulation of PDE1 activity allowed for local regulation of dendritic NO signaling. This inference was supported by simulation of SPN dendritic integration, which revealed that dendritic spikes engaged PDE1 in a branch-specific manner. In a mouse model of Parkinsons disease (PD), NO-LTD was absent not because of a postsynaptic deficit in NO signaling machinery, but rather due to impaired interneuronal NO release. Re-balancing intrastriatal neuromodulatory signaling in the PD model restored NO release and NO-LTD. Taken together, these studies provide novel insights into the mechanisms governing NO-LTD in SPN and its role in psychomotor disorders, like PD.

neuroscience↗

bioRxiv2023

GAPS: Geometric Attention-based Networks for Peptide Binding Sites Identification by the Transfer Learning Approach

The identification of protein-peptide binding sites significantly advances our understanding of their interaction. Recent advancements in deep learning have profoundly transformed the prediction of protein-peptide binding sites. In this work, we describe the Geometric Attention-based networks for Peptide binding Sites identification (GAPS). The GAPS constructs atom representations using geometric feature engineering and employs various attention mechanisms to update pertinent biological features. In addition, the transfer learning strategy is implemented for leveraging the pre-trained protein-protein binding sites information to enhance training of the protein-peptide binding sites recognition, taking into account the similarity of proteins and peptides. Consequently, GAPS demonstrates state-of-the-art (SOTA) performance in this task. Our model also exhibits exceptional performance across several expanded experiments including predicting the apo protein-peptide, the protein-cyclic peptide, and the predicted protein-peptide binding sites. Overall, the GAPS is a powerful, versatile, stable method suitable for diverse binding site predictions.

bioinformatics↗

bioRxiv2023

Universal protection against SARS-CoV-2 viruses by multivalent mRNA vaccine in mice

The continual emergence of new severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variants challenges available SARS-CoV-2 vaccines for adequate control of outbreaks. Currently, universal vaccines capable of obviating the need for exact strain matching between mRNA vaccines and circulating viruses are absent. In this study, we designed, manufactured, and evaluated a nucleoside-modified lipid nanoparticle mRNA vaccine, aimed for offering broad-spectrum protection against recent SARS-CoV-2 variants. Additionally, the protection efficiency of monovalent, bivalent, quadrivalent, and XBB.1.5 mRNA vaccines was compared with the proposed universal vaccine. The neutralizing antibody activity against wuhan-1, BA.4/5, XBB.1.5, B.1.1.529, BQ.1.1, EG.5.1 and JN.1 was assessed using enzyme-linked immunosorbent assay, rapid fiber-optic biolayer interferometry-based biosensor, and pseudovirus neutralization test. Our results reveal that the proposed multivalent vaccine affords comprehensive protection against previously circulating, current and previously unidentified SARS-CoV-2 strains.

immunology↗

bioRxiv2022

Detection of Frog virus 3 via the system integrating RPA-CRISPR/Cas12a-SPM with deep learning

Frog virus 3 (FV3, genera Ranavirus, family Iridoviridae), a double-stranded DNA virus, results in irreparable damage to biodiversity and significant economic losses to aquaculture. Although the existing FV3 detection methods are of high sensitivity and specificity, the complex procedure and requirement of expensive instruments limit their practical implantation. Herein, we develop a fast, easy-to-implement, highly sensitive, and point-of-care (POC) detection system for FV3. Combining recombinase polymerase amplification (RPA) and CRISPR/Cas12a, we achieve a limit of detection (LoD) of 100 aM (60.2 copies/L) by optimizing RPA primers and CRISPR RNAs (crRNAs). For POC detection, we build a smartphone microscopy (SPM) and achieve an LoD of 10 aM within 40 minutes. Four positive animal-derived samples with a quantitation cycle (Cq) value of quantitative PCR (qPCR) in the range of 13 to 32 are detectable by the proposed system. In addition, we deploy deep learning models for binary classification (positive or negative samples) and multiclass classification (different concentrations of FV3 and negative samples), achieving 100% and 98.75% accuracy, respectively. Without temperature regulation and expensive equipment, RPA-CRISPR/Cas12a combined with a smartphone readout and artificial intelligence (AI) assisted classification shows great potential for FV3 detection. This integrated system holds great promise for POC detection of aquatic DNA pathogens. HighlightsO_LIAn integrated DNA detection system is developed by combining RPA, CRISPR/Cas12a, smartphone microscopy, and deep learning. C_LIO_LIThe LoD of frog virus 3 is 10 aM within 40 min. C_LIO_LIThe detection system shows good performance on animal-derived samples. C_LI

molecular biology↗

bioRxiv2021

Mining bacterial NGS data vastly expands the complete genomes of temperate phages

Temperate phages (active prophages induced from bacteria) help control pathogenicity, modulate community structure, and maintain gut homeostasis1. Complete phage genome sequences are indispensable for understanding phage biology. Traditional plaque techniques are inapplicable to temperate phages due to the lysogenicity of these phages, which curb the identification and characterization of temperate phages. Existing in silico tools for prophage prediction usually fail to detect accurate and complete temperate phage genomes2-5. In this study, by a novel computational method mining both the integrated active prophages and their spontaneously induced forms (temperate phages), we obtained 192,326 complete temperate phage genomes from bacterial next-generation sequencing (NGS) data, hence expanded the existing number of complete temperate phage genomes by more than 100-fold. The reliability of our method was validated by wet-lab experiments. The experiments demonstrated that our method can accurately determine the complete genome sequences of the temperate phages, with exact flanking sites (attP and attB sites), outperforming other state-of-the-art prophage prediction methods. Our analysis indicates that temperate phages are likely to function in the evolution of microbes by 1) cross-infecting different bacterial host species; 2) transferring antibiotic resistance and virulence genes; and 3) interacting with hosts through restriction-modification and CRISPR/anti-CRISPR systems. This work provides a comprehensive complete temperate phage genome database and relevant information, which can serve as a valuable resource for phage research.

bioinformatics↗