Rapid discovery of novel prophages using biological feature engineering and machine learning

Genomic characterization and probiotic assessment of Bifidobacterium breve JKL2022 with strain-specific CLA-converting properties

Bifidobacterium breve is a well-recognized probiotic species. B. breve JKL2022, a strain isolated from the feces of healthy infants that exhibits superior conjugated linoleic acid (CLA)-converting activity, was functionally characterized for probiotic safety and applicability through genomic and in vitro analyses. The JKL2022 genome comprises a 2,313,948 bp sequence assembled into a single contig, encoding a total of 1,998 genes. In silico predictive analyses confirmed the absence of virulence factors and acquired resistance genes while verifying its intrinsic antimicrobial resistance profile. Several CAZymes were identified, consistent with the strain's fermentation profile. Additionally, the gene encoding the key enzyme for CLA conversion was identified as a 993-bp lai gene, underscoring the species-level differences in microbial CLA metabolism. The functionality, stress tolerance, and safety of JKL2022 were further confirmed through experimental assessments. JKL2022 exhibited tolerance to acid and bile salts, auto-aggregation, and cell surface hydrophobicity, indicating its potential to survive gastrointestinal transit. Furthermore, JKL2022 exhibited α-glucosidase inhibitory activity and tested negative for starch hydrolysis, hemolysis, and gelatinase activity. The inherent probiotic properties of Bifidobacterium, combined with the strain-specific CLA conversion using growing cells and postbiotic preparations, contribute to the potential health benefits of B. breve JKL2022, as verified in this study.

TIGR-Tas: A family of modular RNA-guided DNA-targeting systems in prokaryotes and their viruses

RNA-guided systems provide remarkable versatility, enabling diverse biological functions. Through iterative structural and sequence homology-based mining starting with a guide RNA-interaction domain of Cas9, we identified a family of RNA-guided DNA-targeting proteins in phage and parasitic bacteria. Each system consists of a tandem interspaced guide RNA (TIGR) array and a TIGR-associated (Tas) protein containing a nucleolar protein (Nop) domain, sometimes fused to HNH (TasH)- or RuvC (TasR)-nuclease domains. We show that TIGR arrays are processed into 36-nucleotide RNAs (tigRNAs) that direct sequence-specific DNA binding through a tandem-spacer targeting mechanism. TasR can be reprogrammed for precise DNA cleavage, including in human cells. The structure of TasR reveals striking similarities to box C/D small nucleolar ribonucleoproteins and IS110 RNA-guided transposases, providing insights into the evolution of diverse RNA-guided systems.

Fully resolved genome assembly of a Macrococcus bovicus isolated from a human skin infection

The complete circular genome of Macrococcus bovicus LI0213 isolated from a human skin lesion was obtained using a hybrid assembly of Nanopore and Illumina reads. The genome consisting of a 2,082,488-bp chromosome and three plasmids, contains phage-related sequences and represents the first fully resolved genome of M. bovicus from human origin.

Development of Chimera AMP-Endolysin with Wider Spectra Against Gram-Negative Bacteria Using High-Throughput Assay

Bacteriophage-derived endolysins are being developed as an alternative to antimicrobials. The development of endolysins against Gram-negative bacteria requires the discovery of effective endolysins against the target species and the capability to penetrate the outer membrane of bacteria by endolysin. Here, we propose an efficient endolysin development approach that combines a data-driven endolysin search utilizing bacterial genomes with high-throughput laboratory assays. As a proof of concept, we analyzed endolysin genes detected in 273 bacterial genomes of Acinetobacter, Pseudomonas, and Escherichia. Firstly, we conducted assays of 192 recombinants of endolysin genes obtained through in silico search from bacterial genomes and identified natural endolysins degrading peptidoglycan of Acinetobacter baumannii. Then, we performed high-throughput screening against Gram-negative bacteria for hundreds of chimera AMP-endolysins, natural endolysin conjugated with antimicrobial peptide. As a result, we obtained four chimera AMP-endolysins against A. baumannii, which demonstrated the minimum inhibitory concentration ranging from 4 to 8 μg/mL. Moreover, we assessed the antimicrobial spectra of these chimera AMP-endolysins, validating that two endolysins exhibited antimicrobial efficacy against Pseudomonas aeruginosa and Escherichia coli with <32 μg/mL of concentration. This endolysin development approach can be applied to other Gram-negative bacterial targets and is expected to facilitate the acquisition of effective novel endolysins.

Engineering Phages to Fight Multidrug-Resistant Bacteria

Facing the global "superbug" crisis due to the emergence and selection for antibiotic resistance, phages are among the most promising solutions. Fighting multidrug-resistant bacteria requires precise diagnosis of bacterial pathogens and specific cell-killing. Phages have several potential advantages over conventional antibacterial agents such as host specificity, self-amplification, easy production, low toxicity as well as biofilm degradation. However, the narrow host range, uncharacterized properties, as well as potential risks from exponential replication and evolution of natural phages, currently limit their applications. Engineering phages can not only enhance the host bacteria range and improve phage efficacy, but also confer new functions. This review first summarizes major phage engineering techniques including both chemical modification and genetic engineering. Subsequent sections discuss the applications of engineered phages for bacterial pathogen detection and ablation through interdisciplinary approaches of synthetic biology and nanotechnology. We discuss future directions and persistent challenges in the ongoing exploration of phage engineering for pathogen control.

Dimer-monomer transition defines a hyper-thermostable peptidoglycan hydrolase mined from bacterial proteome by lysin-derived antimicrobial peptide-primed screening

Phage-derived peptidoglycan hydrolases (i.e. lysins) are considered promising alternatives to conventional antibiotics due to their direct peptidoglycan degradation activity and low risk of resistance development. The discovery of these enzymes is often hampered by the limited availability of phage genomes. Herein, we report a new strategy to mine active peptidoglycan hydrolases from bacterial proteomes by lysin-derived antimicrobial peptide-primed screening. As a proof-of-concept, five peptidoglycan hydrolases from the Acinetobacter baumannii proteome (PHAb7-PHAb11) were identified using PlyF307 lysin-derived peptide as a template. Among them, PHAb10 and PHAb11 showed potent bactericidal activity against multiple pathogens even after treatment at 100°C for 1 hr, while the other three were thermosensitive. We solved the crystal structures of PHAb8, PHAb10, and PHAb11 and unveiled that hyper-thermostable PHAb10 underwent a unique folding-refolding thermodynamic scheme mediated by a dimer-monomer transition, while thermosensitive PHAb8 formed a monomer. Two mouse models of bacterial infection further demonstrated the safety and efficacy of PHAb10. In conclusion, our antimicrobial peptide-primed strategy provides new clues for the discovery of promising antimicrobial drugs.

A single amplified genome catalog reveals the dynamics of mobilome and resistome in the human microbiome

BACKGROUND

The increase in metagenome-assembled genomes (MAGs) has advanced our understanding of the functional characterization and taxonomic assignment within the human microbiome. However, MAGs, as population consensus genomes, often aggregate heterogeneity among species and strains, thereby obfuscating the precise relationships between microbial hosts and mobile genetic elements (MGEs). In contrast, single amplified genomes (SAGs) derived via single-cell genome sequencing can capture individual genomic content, including MGEs.

RESULTS

We introduce the first substantial SAG dataset (bbsag20) from the human oral and gut microbiome, comprising 17,202 SAGs above medium-quality without co-assembly. This collection unveils a diversity of bacterial lineages across 312 oral and 647 gut species, demonstrating different taxonomic compositions from MAGs. Moreover, the SAGs showed cellular-level evidence of the translocation of oral bacteria to the gut. We also identified broad-host-range MGEs harboring antibiotic resistance genes (ARGs), which were not detected in the MAGs.

CONCLUSIONS

The difference in taxonomic composition between SAGs and MAGs indicates that combining both methods would be effective in expanding the genome catalog. By connecting mobilomes and resistomes in individual samples, SAGs could meticulously chart a dynamic network of ARGs on MGEs, pinpointing potential ARG reservoirs and their spreading patterns in the microbial community. Video Abstract.

Protein Set Transformer: A protein-based genome language model to power high diversity viromics

Exponential increases in microbial and viral genomic data demand transformational advances in scalable, generalizable frameworks for their interpretation. Standard homology-based functional analyses are hindered by the rapid divergence of microbial and especially viral genomes and proteins that significantly decreases the volume of usable data. Here, we present Protein Set Transformer (PST), a protein-based genome language model that models genomes as sets of proteins without considering sparsely available functional labels. Trained on >100k viruses, PST outperformed other homology- and language model-based approaches for relating viral genomes based on shared protein content. Further, PST demonstrated protein structural and functional awareness by clustering capsid-fold-containing proteins with known capsid proteins and uniquely clustering late gene proteins within related viruses. Our data establish PST as a valuable method for diverse viral genomics, ecology, and evolutionary applications. We posit that the PST framework can be a foundation model for microbial genomics when trained on suitable data.

Uncovering Endolysins against Methicillin-Resistant Staphylococcus aureus Using a Microbial Single-Cell Genome Database

Endolysins, peptidoglycan hydrolases derived from bacteriophages (phages), are being developed as a promising alternative to conventional antibiotics. To obtain highly active endolysins, a diverse library of these endolysins is vital. We propose here microbial single-cell genome sequencing as an efficient tool to discover dozens of previously unknown endolysins, owing to its culture-independent sequencing method. As a proof of concept, we analyzed and recovered endolysin genes within prophage regions of Staphylococcus single-amplified genomes in human skin microbiome samples. We constructed a library of chimeric endolysins by shuffling domains of the natural endolysins and performed high-throughput screening against Staphylococcus aureus. One of the lead endolysins, bbst1027, exhibited desirable antimicrobial properties, such as rapid bactericidal activity, no detectable resistance development, and in vivo efficacy. We foresee that this endolysin discovery pipeline is in principle applicable to any bacterial target and boost the development of novel antimicrobial agents.

Protein Set Transformer: A protein-based genome language model to power high diversity viromics

Exponential increases in microbial and viral genomic data demand transformational advances in scalable, generalizable frameworks for their interpretation. Standard homology-based functional analyses are hindered by the rapid divergence of microbial and especially viral genomes and proteins that significantly decreases the volume of usable data. Here, we present Protein Set Transformer (PST), a protein-based genome language model that models genomes as sets of proteins without considering sparsely available functional labels. Trained on >100k viruses, PST outperformed other homology- and language model-based approaches for relating viral genomes based on shared protein content. Further, PST demonstrated protein structural and functional awareness by clustering capsid-fold-containing proteins with known capsid proteins and uniquely clustering late gene proteins within related viruses. Our data establish PST as a valuable method for diverse viral genomics, ecology, and evolutionary applications. We posit that the PST framework can be a foundation model for microbial genomics when trained on suitable data.

Precise microbiome engineering using natural and synthetic bacteriophages targeting an artificial bacterial consortium

In natural microbiomes, microorganisms interact with each other and exhibit diverse functions. Microbiome engineering, which enables bacterial knockdown, is a promising method to elucidate the functions of targeted bacteria in microbiomes. However, few methods to selectively kill target microorganisms in the microbiome without affecting the growth of nontarget microorganisms are available. In this study, we focused on the host-specific lytic ability of virulent phages and validated their potency for precise microbiome engineering. In an artificial microbiome consisting of Escherichia coli, Pseudomonas putida, Bacillus subtilis, and Lactiplantibacillus plantarum, the addition of bacteriophages infecting their respective host strains specifically reduced the number of these bacteria more than 102 orders. Remarkably, the reduction in target bacteria did not affect the growth of nontarget bacteria, indicating that bacteriophages were effective tools for precise microbiome engineering. Moreover, a virulent derivative of the λ phage was synthesized from prophage DNA in the genome of λ lysogen by in vivo DNA assembly and phage-rebooting techniques, and E. coli-targeted microbiome engineering was achieved. These results propose a novel approach for precise microbiome engineering using bacteriophages, in which virulent phages are synthesized from prophage DNA in lysogenic strains without isolating phages from environmental samples.

Antimicrobial resistance crisis: could artificial intelligence be the solution?

Antimicrobial resistance is a global public health threat, and the World Health Organization (WHO) has announced a priority list of the most threatening pathogens against which novel antibiotics need to be developed. The discovery and introduction of novel antibiotics are time-consuming and expensive. According to WHO's report of antibacterial agents in clinical development, only 18 novel antibiotics have been approved since 2014. Therefore, novel antibiotics are critically needed. Artificial intelligence (AI) has been rapidly applied to drug development since its recent technical breakthrough and has dramatically improved the efficiency of the discovery of novel antibiotics. Here, we first summarized recently marketed novel antibiotics, and antibiotic candidates in clinical development. In addition, we systematically reviewed the involvement of AI in antibacterial drug development and utilization, including small molecules, antimicrobial peptides, phage therapy, essential oils, as well as resistance mechanism prediction, and antibiotic stewardship.

Activity of novel virus families infecting soil nitrifiers is concomitant with host niche differentiation

Chemolithoautotrophic nitrifiers are model groups for linking phylogeny, evolution, and ecophysiology. Ammonia-oxidizing bacteria (AOB) typically dominate the first step of ammonia oxidation at high ammonium supply rates, ammonia-oxidizing archaea (AOA) and complete ammonia-oxidizing Nitrospira (comammox) are often active at lower supply rates or during AOB inactivity, and nitrite-oxidizing bacteria (NOB) complete canonical nitrification. Soil virus communities are dynamic but contributions to functional processes are largely undetermined. In addition, characterizing viruses infecting hosts with low relative abundance, such as nitrifiers, may be constrained by vast viral diversity, partial genome recovery, and difficulties in host linkage. Here, we describe a targeted incubation study that aimed to determine whether growth of different nitrifier groups in soil is associated with active virus populations and if process-focused analyses facilitate characterization of high-quality virus genomes. dsDNA viruses infecting different nitrifier groups were enriched in situ via differential host inhibition. Growth of each nitrifier group was consistent with predicted inhibition profiles and concomitant with the abundance of their viruses. These included 61 high-quality/complete virus genomes 35-173 kb in length with minimal similarity to validated families. AOA viruses lacked ammonia monooxygenase sub-unit C (amoC) genes found in marine AOA viruses but some encoded AOA-specific multicopper oxidase type 1 (MCO1), previously implicated in copper acquisition, and suggesting a role in supporting energy metabolism of soil AOA. Findings demonstrate focused incubation studies facilitate characterization of active host-virus interactions associated with specific processes and viruses of soil AOA, AOB, and NOB are dynamic and potentially influence nitrogen cycling processes.

Viral diversity and host associations in microbial electrolysis cells

In microbial electrolysis cells (MECs), microbial communities catalyze conversions between dissolved organic compounds, electrical energy, and energy carriers such as hydrogen and methane. Bacteria and archaea, which catalyze reactions on the anode and cathode of MECs, interact with phages; however, phage communities have previously not been examined in MECs. In this study, we used metagenomic sequencing to study prokaryotes and phages in nine MECs. A total of 852 prokaryotic draft genomes representing 278 species, and 1476 phage contigs representing 873 phage species were assembled. Among high quality prokaryotic genomes (>95% completion), 55% carried a prophage, and the three Desulfobacterota spp. that dominated the anode communities all carried prophages. Geobacter anodireducens, one of the bacteria dominating the anode communities, carried a CRISPR spacer showing evidence of a previous infection by a Peduoviridae phage present in the liquid of some MECs. Methanobacteriaceae spp. and an Acetobacterium sp., which dominated the cathodes, had several associations with Straboviridae spp. The results of this study show that phage communities in MECs are diverse and interact with functional microorganisms on both the anode and cathode.

Phages are unrecognized players in the ecology of the oral pathogen Porphyromonas gingivalis

BACKGROUND

Porphyromonas gingivalis (hereafter "Pg") is an oral pathogen that has been hypothesized to act as a keystone driver of inflammation and periodontal disease. Although Pg is most readily recovered from individuals with actively progressing periodontal disease, healthy individuals and those with stable non-progressing disease are also colonized by Pg. Insights into the factors shaping the striking strain-level variation in Pg, and its variable associations with disease, are needed to achieve a more mechanistic understanding of periodontal disease and its progression. One of the key forces often shaping strain-level diversity in microbial communities is infection of bacteria by their viral (phage) predators and symbionts. Surprisingly, although Pg has been the subject of study for over 40 years, essentially nothing is known of its phages, and the prevailing paradigm is that phages are not important in the ecology of Pg.

RESULTS

Here we systematically addressed the question of whether Pg are infected by phages-and we found that they are. We found that prophages are common in Pg, they are genomically diverse, and they encode genes that have the potential to alter Pg physiology and interactions. We found that phages represent unrecognized targets of the prevalent CRISPR-Cas defense systems in Pg, and that Pg strains encode numerous additional mechanistically diverse candidate anti-phage defense systems. We also found that phages and candidate anti-phage defense system elements together are major contributors to strain-level diversity and the species pangenome of this oral pathogen. Finally, we demonstrate that prophages harbored by a model Pg strain are active in culture, producing extracellular viral particles in broth cultures.

CONCLUSION

This work definitively establishes that phages are a major unrecognized force shaping the ecology and intra-species strain-level diversity of the well-studied oral pathogen Pg. The foundational phage sequence datasets and model systems that we establish here add to the rich context of all that is already known about Pg, and point to numerous avenues of future inquiry that promise to shed new light on fundamental features of phage impacts on human health and disease broadly. Video Abstract.

Target enrichment of uncultured human oral bacteria with phage-derived molecules found by single-cell genomics

Advances in culture-independent microbial analysis, such as metagenomics and single-cell genomics, have significantly increased our understanding of microbial lineages. While these methods have uncovered a large number of novel microbial taxa, many remain uncultured, and their function and mode of existence in the environment are still unknown. This study aims to explore the use of bacteriophage-derived molecules as probes for detecting and isolating uncultured bacteria. Here, we proposed multiplex single-cell sequencing to obtain massive uncultured oral bacterial genomes and searched prophage sequences from over 450 obtained human oral bacterial single-amplified genomes (SAGs). The focus was on the cell wall binding domain (CBD) in phage endolysin, and fluorescent protein-fused CBDs were generated based on several CBD gene sequences predicted from Streptococcus SAGs. The ability of the Streptococcus prophage-derived CBDs to detect and enrich specific Streptococcus species from human saliva while maintaining cell viability was confirmed by magnetic separation and flow cytometry. The approach to phage-derived molecule generation based on uncultured bacterial SAG is expected to improve the process of designing molecules that selectively capture or detect specific bacteria, notably from uncultured gram-positive bacteria, and will have applications in isolation and in situ detection of beneficial or pathogenic bacteria.

Sequencing the genomes of LPP-1, the first isolated cyanophage, and its relative LPP-2 reveal different integration mechanisms in closely related phages

In the early 1960s, the first cyanophage was isolated. The description of this phage, named LPP-1, led to the extensive investigation of various cyanophages and to the study of their interactions with their cyanobacterial hosts towards controlling blooms. Here, the genomes of LPP-1 and its putative relative, LPP-2 were sequenced. Sequencing these genomes revealed that LPP-1 and LPP-2 are members of a group of short-tailed cyanophages, which are distantly related to the T7-like cyanophages. Most of the phages in this group have the ability to lysogenize their hosts. Their ability to switch between lytic and lysogenic infection may explain the formation of cyanobacterial blooms despite the persistence of their phages. This lysogenic capacity of the LPP-1-like phages occurs despite the lack of an obvious integrase gene within their genomes. Interestingly, we show that LPP-2 integrates into the host genome through an integration site in high proximity to a recombination endonuclease that may have integrase activity. Further understanding of cyanobacterial-phage relationships may provide insight into their population dynamics and suggest novel approaches for control of destructive cyanobacterial blooms.

Gauge your phage: benchmarking of bacteriophage identification tools in metagenomic sequencing data

BACKGROUND

The prediction of bacteriophage sequences in metagenomic datasets has become a topic of considerable interest, leading to the development of many novel bioinformatic tools. A comparative analysis of ten state-of-the-art phage identification tools was performed to inform their usage in microbiome research.

METHODS

Artificial contigs generated from complete RefSeq genomes representing phages, plasmids, and chromosomes, and a previously sequenced mock community containing four phage species, were used to evaluate the precision, recall, and F1 scores of the tools. We also generated a dataset of randomly shuffled sequences to quantify false-positive calls. In addition, a set of previously simulated viromes was used to assess diversity bias in each tool's output.

RESULTS

VIBRANT and VirSorter2 achieved the highest F1 scores (0.93) in the RefSeq artificial contigs dataset, with several other tools also performing well. Kraken2 had the highest F1 score (0.86) in the mock community benchmark by a large margin (0.3 higher than DeepVirFinder in second place), mainly due to its high precision (0.96). Generally, k-mer-based tools performed better than reference similarity tools and gene-based methods. Several tools, most notably PPR-Meta, called a high number of false positives in the randomly shuffled sequences. When analysing the diversity of the genomes that each tool predicted from a virome set, most tools produced a viral genome set that had similar alpha- and beta-diversity patterns to the original population, with Seeker being a notable exception.

CONCLUSIONS

This study provides key metrics used to assess performance of phage detection tools, offers a framework for further comparison of additional viral discovery tools, and discusses optimal strategies for using these tools. We highlight that the choice of tool for identification of phages in metagenomic datasets, as well as their parameters, can bias the results and provide pointers for different use case scenarios. We have also made our benchmarking dataset available for download in order to facilitate future comparisons of phage identification tools. Video Abstract.

A model industrial workhorse: Bacillus subtilis strain 168 and its genome after a quarter of a century

The vast majority of genomic sequences are automatically annotated using various software programs. The accuracy of these annotations depends heavily on the very few manual annotation efforts that combine verified experimental data with genomic sequences from model organisms. Here, we summarize the updated functional annotation of Bacillus subtilis strain 168, a quarter century after its genome sequence was first made public. Since the last such effort 5 years ago, 1168 genetic functions have been updated, allowing the construction of a new metabolic model of this organism of environmental and industrial interest. The emphasis in this review is on new metabolic insights, the role of metals in metabolism and macromolecule biosynthesis, functions involved in biofilm formation, features controlling cell growth, and finally, protein agents that allow class discrimination, thus allowing maintenance management, and accuracy of all cell processes. New 'genomic objects' and an extensive updated literature review have been included for the sequence, now available at the International Nucleotide Sequence Database Collaboration (INSDC: AccNum AL009126.4).

Population-level impacts of antibiotic usage on the human gut microbiome

The widespread usage of antimicrobials has driven the evolution of resistance in pathogenic microbes, both increased prevalence of antimicrobial resistance genes (ARGs) and their spread across species by horizontal gene transfer (HGT). However, the impact on the wider community of commensal microbes associated with the human body, the microbiome, is less well understood. Small-scale studies have determined the transient impacts of antibiotic consumption but we conduct an extensive survey of ARGs in 8972 metagenomes to determine the population-level impacts. Focusing on 3096 gut microbiomes from healthy individuals not taking antibiotics we demonstrate highly significant correlations between both the total ARG abundance and diversity and per capita antibiotic usage rates across ten countries spanning three continents. Samples from China were notable outliers. We use a collection of 154,723 human-associated metagenome assembled genomes (MAGs) to link these ARGs to taxa and detect HGT. This reveals that the correlations in ARG abundance are driven by multi-species mobile ARGs shared between pathogens and commensals, within a highly connected central component of the network of MAGs and ARGs. We also observe that individual human gut ARG profiles cluster into two types or resistotypes. The less frequent resistotype has higher overall ARG abundance, is associated with certain classes of resistance, and is linked to species-specific genes in the Proteobacteria on the periphery of the ARG network.

PhaGAA: an integrated web server platform for phage genome annotation and analysis

MOTIVATION

Phage genome annotation plays a key role in the design of phage therapy. To date, there have been various genome annotation tools for phages, but most of these tools focus on mono-functional annotation and have complex operational processes. Accordingly, comprehensive and user-friendly platforms for phage genome annotation are needed.

RESULTS

Here, we propose PhaGAA, an online integrated platform for phage genome annotation and analysis. By incorporating several annotation tools, PhaGAA is constructed to annotate the prophage genome at DNA and protein levels and provide the analytical results. Furthermore, PhaGAA could mine and annotate phage genomes from bacterial genome or metagenome. In summary, PhaGAA will be a useful resource for experimental biologists and help advance the phage synthetic biology in basic and application research.

AVAILABILITY AND IMPLEMENTATION

PhaGAA is freely available at http://phage.xialab.info/.

Genetic determinants of host tropism in Klebsiella phages

Bacteriophages play key roles in bacterial ecology and evolution and are potential antimicrobials. However, the determinants of phage-host specificity remain elusive. Here, we isolate 46 phages to challenge 138 representative clinical isolates of Klebsiella pneumoniae, a widespread opportunistic pathogen. Spot tests show a narrow host range for most phages, with <2% of 6,319 phage-host combinations tested yielding detectable interactions. Bacterial capsule diversity is the main factor restricting phage host range. Consequently, phage-encoded depolymerases are key determinants of host tropism, and depolymerase sequence types are associated with the ability to infect specific capsular types across phage families. However, all phages with a broader host range found do not encode canonical depolymerases, suggesting alternative modes of entry. These findings expand our knowledge of the complex interactions between bacteria and their viruses and point out the feasibility of predicting the first steps of phage infection using bacterial and phage genome sequences.

Advances in the field of phage-based therapy with special emphasis on computational resources

In the current era, one of the major challenges is to manage the treatment of drug/antibiotic-resistant strains of bacteria. Phage therapy, a century-old technique, may serve as an alternative to antibiotics in treating bacterial infections caused by drug-resistant strains of bacteria. In this review, a systematic attempt has been made to summarize phage-based therapy in depth. This review has been divided into the following two sections: general information and computer-aided phage therapy (CAPT). In the case of general information, we cover the history of phage therapy, the mechanism of action, the status of phage-based products (approved and clinical trials) and the challenges. This review emphasizes CAPT, where we have covered primary phage-associated resources, phage prediction methods and pipelines. This review covers a wide range of databases and resources, including viral genomes and proteins, phage receptors, host genomes of phages, phage-host interactions and lytic proteins. In the post-genomic era, identifying the most suitable phage for lysing a drug-resistant strain of bacterium is crucial for developing alternate treatments for drug-resistant bacteria and this remains a challenging problem. Thus, we compile all phage-associated prediction methods that include the prediction of phages for a bacterial strain, the host for a phage and the identification of interacting phage-host pairs. Most of these methods have been developed using machine learning and deep learning techniques. This review also discussed recent advances in the field of CAPT, where we briefly describe computational tools available for predicting phage virions, the life cycle of phages and prophage identification. Finally, we describe phage-based therapy's advantages, challenges and opportunities.

Bacteriophages playing nice: Lysogenic bacteriophage replication stable in the human gut microbiota

Bacteriophages, viruses specific to bacteria, coexist with their bacterial hosts with limited diversity fluctuations in the guts of healthy individuals where they replicate mostly via lysogenic replication. This favors 'piggy-back-the-winner' over 'kill-the-winner' dynamics which are driven by lytic bacteriophage replication. Revisiting the deep-viral sequencing data of a healthy individual sampled over 2.4 years, we explore how these dynamics occur. Prophages found in assembled bacterial metagenomes were also found extra-cellularly, as induced phage particles (iPPs), likely derived from prophage activation. These iPPs were diverse and continually present in low abundance, relative to the highly abundant but less diverse lytic phage population. The continuous detection of low levels of iPPs suggests that spontaneous induction regularly occurs in this healthy individual, possibly allowing prophages to maintain their ability to replicate and avoiding degradation and loss from the gut microbiota.

Prophage-Derived Regions in Curtobacterium Genomes: Good Things, Small Packages

Curtobacterium is a genus of Gram-positive bacteria within the order Actinomycetales. Some Curtobacterium species (C. flaccumfaciens, C. plantarum) are harmful pathogens of agricultural crops such as soybean, dry beans, peas, sugar beet and beetroot, which occur throughout the world. Bacteriophages (bacterial viruses) are considered to be potential curative agents to control the spread of harmful bacteria. Temperate bacteriophages integrate their genomes into bacterial chromosomes (prophages), sometimes substantially influencing bacterial lifestyle and pathogenicity. About 200 publicly available genomes of Curtobacterium species, including environmental metagenomic sequences, were inspected for the presence of sequences of possible prophage origin using bioinformatic methods. The comparison of the search results with several ubiquitous bacterial groups showed the relatively low level of the presence of prophage traces in Curtobacterium genomes. Genomic and phylogenetic analyses were undertaken for the evaluation of the evolutionary and taxonomic positioning of predicted prophages. The analyses indicated the relatedness of Curtobacterium prophage-derived sequences with temperate actinophages of siphoviral morphology. In most cases, the predicted prophages can represent novel phage taxa not described previously. One of the predicted temperate phages was induced from the Curtobacterium genome. Bioinformatic analysis of the modelled proteins encoded in prophage-derived regions led to the discovery of some 100 putative glycopolymer-degrading enzymes that contained enzymatic domains with predicted cell-wall- and cell-envelope-degrading activity; these included glycosidases and peptidases. These proteins can be considered for the experimental design of new antibacterials against Curtobacterium phytopathogens.

Characterization of a P1-bacteriophage-like plasmid (phage-plasmid) harbouring bla CTX-M-15 in Salmonella enterica serovar Typhi

Antimicrobial-resistance (AMR) genes can be transferred between microbial cells via horizontal gene transfer (HGT), which involves mobile and integrative elements such as plasmids, bacteriophages, transposons, integrons and pathogenicity islands. Bacteriophages are found in abundance in the microbial world, but their role in virulence and AMR has not fully been elucidated in the Enterobacterales. With short-read sequencing paving the way to systematic high-throughput AMR gene detection, long-read sequencing technologies now enable us to establish how such genes are structurally connected into meaningful genomic units, raising questions about how they might cooperate to achieve their biological function. Here, we describe a novel ~98 kbp circular P1-bacteriophage-like plasmid termed ph681355 isolated from a clinical Salmonella enterica serovar Typhi isolate. It carries bla _CTX-M-15, an IncY plasmid replicon (repY gene) and the ISEcP1 mobile element and is, to our knowledge, the first reported P1-bacteriophage-like plasmid (phage-plasmid) in S. enterica Typhi. We compared ph681355 to two previously described phage-plasmids, pSJ46 from S. enterica serovar Indiana and pMCR-1-P3 from Escherichia coli, and found high nucleotide similarity across the backbone. However, we saw low ph681355 backbone similarity to plasmid p60006 associated with the extensively drug-resistant S. enterica Typhi outbreak isolate in Pakistan, providing evidence of an alternative route for bla _CTX-M-15 transmission. Our discovery highlights the importance of utilizing long-read sequencing in interrogating bacterial genomic architecture to fully understand AMR mechanisms and their clinical relevance. It also raises questions regarding how widespread bacteriophage-mediated HGT might be, suggesting that the resulting genomic plasticity might be higher than previously thought.

Eleven quick tips for data cleaning and feature engineering

Applying computational statistics or machine learning methods to data is a key component of many scientific studies, in any field, but alone might not be sufficient to generate robust and reliable outcomes and results. Before applying any discovery method, preprocessing steps are necessary to prepare the data to the computational analysis. In this framework, data cleaning and feature engineering are key pillars of any scientific study involving data analysis and that should be adequately designed and performed since the first phases of the project. We call "feature" a variable describing a particular trait of a person or an observation, recorded usually as a column in a dataset. Even if pivotal, these data cleaning and feature engineering steps sometimes are done poorly or inefficiently, especially by beginners and unexperienced researchers. For this reason, we propose here our quick tips for data cleaning and feature engineering on how to carry out these important preprocessing steps correctly avoiding common mistakes and pitfalls. Although we designed these guidelines with bioinformatics and health informatics scenarios in mind, we believe they can more in general be applied to any scientific area. We therefore target these guidelines to any researcher or practitioners wanting to perform data cleaning or feature engineering. We believe our simple recommendations can help researchers and scholars perform better computational analyses that can lead, in turn, to more solid outcomes and more reliable discoveries.

3CAC: improving the classification of phages and plasmids in metagenomic assemblies using assembly graphs

MOTIVATION

Bacteriophages and plasmids usually coexist with their host bacteria in microbial communities and play important roles in microbial evolution. Accurately identifying sequence contigs as phages, plasmids and bacterial chromosomes in mixed metagenomic assemblies is critical for further unraveling their functions. Many classification tools have been developed for identifying either phages or plasmids in metagenomic assemblies. However, only two classifiers, PPR-Meta and viralVerify, were proposed to simultaneously identify phages and plasmids in mixed metagenomic assemblies. Due to the very high fraction of chromosome contigs in the assemblies, both tools achieve high precision in the classification of chromosomes but perform poorly in classifying phages and plasmids. Short contigs in these assemblies are often wrongly classified or classified as uncertain.

RESULTS

Here we present 3CAC, a new three-class classifier that improves the precision of phage and plasmid classification. 3CAC starts with an initial three-class classification generated by existing classifiers and improves the classification of short contigs and contigs with low confidence classification by using proximity in the assembly graph. Evaluation on simulated metagenomes and on real human gut microbiome samples showed that 3CAC outperformed PPR-Meta and viralVerify in both precision and recall, and increased F1-score by 10-60 percentage points.

AVAILABILITY AND IMPLEMENTATION

The 3CAC software is available on https://github.com/Shamir-Lab/3CAC.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Evaluating Plant Gene Models Using Machine Learning

Gene models are regions of the genome that can be transcribed into RNA and translated to proteins, or belong to a class of non-coding RNA genes. The prediction of gene models is a complex process that can be unreliable, leading to false positive annotations. To help support the calling of confident conserved gene models and minimize false positives arising during gene model prediction we have developed Truegene, a machine learning approach to classify potential low confidence gene models using 14 gene and 41 protein-based characteristics. Amino acid and nucleotide sequence-based features were calculated for conserved (high confidence) and non-conserved (low confidence) annotated genes from the published Pisum sativum Cameor genome. These features were used to train eXtreme Gradient Boost (XGBoost) classifier models to predict whether a gene model is likely to be real. The optimized models demonstrated a prediction accuracy ranging from 87% to 90% and an F-1 score of 0.91–0.94. We used SHapley Additive exPlanations (SHAP) and feature importance plots to identify the features that contribute to the model predictions, and we show that protein and gene-based features can be used to build accurate models for gene prediction that have applications in supporting future gene annotation processes.

Phenotypic characterization and analysis of complete genomes of two distinct strains of the proposed species "L. swaminathanii"

Recently, a new Listeria species, “Listeria swaminathanii”, was proposed. Here, we phenotypically and genotypically characterize two additional strains that were previously obtained from soil samples and compare the results to the type strain. Complete genomes for both strains were assembled from hybrid Illumina and Nanopore sequencing reads and annotated. Further genomic analysis including average nucleotide identity (ANI) and detection of mobile genetic elements and genes of interest (e.g., virulence-associated) were conducted. The strains showed 98.7–98.8% ANI with the type strain. The UTK C1-0015 genome contained a partial monocin locus and a plasmid, while the UTK C1-0024 genome contained a full monocin locus and a prophage. Phenotypic characterization consistent with those performed on the proposed type strain was conducted to assess consistency of phenotypes across a greater diversity of the proposed species (n = 3 instead of n = 1). Only a few findings were notably different from those of the type strain, such as catalase activity, glycerol metabolism, starch metabolism, and growth at 41 °C. This study further expands our understanding of this newly proposed sensu stricto Listeria species.

Unifying the known and unknown microbial coding sequence space

Genes of unknown function are among the biggest challenges in molecular biology, especially in microbial systems, where 40-60% of the predicted genes are unknown. Despite previous attempts, systematic approaches to include the unknown fraction into analytical workflows are still lacking. Here, we present a conceptual framework, its translation into the computational workflow AGNOSTOS and a demonstration on how we can bridge the known-unknown gap in genomes and metagenomes. By analyzing 415,971,742 genes predicted from 1749 metagenomes and 28,941 bacterial and archaeal genomes, we quantify the extent of the unknown fraction, its diversity, and its relevance across multiple organisms and environments. The unknown sequence space is exceptionally diverse, phylogenetically more conserved than the known fraction and predominantly taxonomically restricted at the species level. From the 71 M genes identified to be of unknown function, we compiled a collection of 283,874 lineage-specific genes of unknown function for Cand. Patescibacteria (also known as Candidate Phyla Radiation, CPR), which provides a significant resource to expand our understanding of their unusual biology. Finally, by identifying a target gene of unknown function for antibiotic resistance, we demonstrate how we can enable the generation of hypotheses that can be used to augment experimental data.

PhageLeads: Rapid Assessment of Phage Therapeutic Suitability Using an Ensemble Machine Learning Approach

The characterization of therapeutic phage genomes plays a crucial role in the success rate of phage therapies. There are three checkpoints that need to be examined for the selection of phage candidates, namely, the presence of temperate markers, antimicrobial resistance (AMR) genes, and virulence genes. However, currently, no single-step tools are available for this purpose. Hence, we have developed a tool capable of checking all three conditions required for the selection of suitable therapeutic phage candidates. This tool consists of an ensemble of machine-learning-based predictors for determining the presence of temperate markers (integrase, Cro/CI repressor, immunity repressor, DNA partitioning protein A, and antirepressor) along with the integration of the ABRicate tool to determine the presence of antibiotic resistance genes and virulence genes. Using the biological features of the temperate markers, we were able to predict the presence of the temperate markers with high MCC scores (>0.70), corresponding to the lifestyle of the phages with an accuracy of 96.5%. Additionally, the screening of 183 lytic phage genomes revealed that six phages were found to contain AMR or virulence genes, showing that not all lytic phages are suitable to be used for therapy. The suite of predictors, PhageLeads, along with the integrated ABRicate tool, can be accessed online for in silico selection of suitable therapeutic phage candidates from single genome or metagenomic contigs.

Isolation and characterization of a novel Lambda-like phage infecting the bloom-forming cyanobacteria Cylindrospermopsis raciborskii

Cylindrospermopsis raciborskii is a central bloom‐forming cyanobacteria. However, despite its ecological significance, little is known of its interactions with the phages that infect it. Currently, only a single sequenced genome of a Cylindrospermopsis‐infecting phage is publicly available. Here we describe the isolation and characterization of Cr‐LKS3, a second phage infecting Cylindrospermopsis. Cr‐LKS3 is a siphovirus with a higher genome similarity to prophages within heterotrophic bacteria genomes than to any other cyanophage/cyano‐prophage, suggesting that it represents a novel cyanophage group. The function, order and orientation of the 72 genes in the Cr‐LKS3 genome are highly similar to those of Escherichia virus Lambda (hereafter Lambda), despite the very low sequence similarity between these phages, showing high evolutionary convergence despite the substantial difference in host characteristics. Similarly to Lambda, the genome of Cr‐LKS3 contains various genes that are known to be central to lysogeny, suggesting it can enter a lysogenic cycle. Cr‐LKS3 has a unique ability to infect a host with a dramatically different GC content, without carrying any tRNA genes to compensate for this difference. This ability, together with its potential lysogenic lifestyle shed light on the complex interactions between C. raciborskii and its phages.

OUP accepted manuscript

Advances in genome sequencing have produced hundreds of thousands of bacterial genome sequences, many of which have integrated prophages derived from temperate bacteriophages. These prophages play key roles by influencing bacterial metabolism, pathogenicity, antibiotic resistance, and defense against viral attack. However, they vary considerably even among related bacterial strains, and they are challenging to identify computationally and to extract precisely for comparative genomic analyses. Here, we describe DEPhT, a multimodal tool for prophage discovery and extraction. It has three run modes that facilitate rapid screening of large numbers of bacterial genomes, precise extraction of prophage sequences, and prophage annotation. DEPhT uses genomic architectural features that discriminate between phage and bacterial sequences for efficient prophage discovery, and targeted homology searches for precise prophage extraction. DEPhT is designed for prophage discovery in Mycobacterium genomes but can be adapted broadly to other bacteria. We deploy DEPhT to demonstrate that prophages are prevalent in Mycobacterium strains but are absent not only from the few well-characterized Mycobacterium tuberculosis strains, but also are absent from all ∼30 000 sequenced M. tuberculosis strains.

From Trees to Clouds: PhageClouds for Fast Comparison of ∼640,000 Phage Genomic Sequences and Host-Centric Visualization Using Genomic Network Graphs

Background: Fast and computationally efficient strategies are required to explore genomic relationships within an increasingly large and diverse phage sequence space. Here, we present PhageClouds, a novel approach using a graph database of phage genomic sequences and their intergenomic distances to explore the phage genomic sequence space. Methods: A total of 640,000 phage genomic sequences were retrieved from a variety of databases and public virome assemblies. Intergenomic distances were calculated with dashing, an alignment-free method suitable for handling massive data sets. These data were used to build a Neo4j® graph database. Results: PhageClouds supported the search of related phages among all complete phage genomes from GenBank for a single query phage in just 10 s. Moreover, PhageClouds expanded the number of closely related phage sequences detected for both finished and draft phage genomes, in comparison with searches exclusively targeting phage entries from GenBank. Conclusions: PhageClouds is a novel resource that will facilitate the analysis of phage genomic sequences and the characterization of assembled phage genomes.