Genetic diversity of blaKPC-gene-containing IncF plasmids from epidemiologically related and unrelated Enterobacteriaceae

AbstractBackgroundLimited information is available on whether blaKPC-containing plasmids from isolates in a hospital outbreak can be differentiated from epidemiologically unrelated blaKPC-containing plasmids based on sequence data.ObjectiveThis study aimed to evaluate the performance of three approaches to distinguish epidemiologically related from unrelated blaKPC-containing IncF plasmids.MethodEpidemiologically related isolates, were short- and long-read whole genome sequenced on an Illumina MiSeq and MinION sequencer. A hybrid assembly was performed and plasmid sequences were extracted from the assembly graph. Epidemiologically unrelated plasmid sequences were extracted from the GenBank. Pairwise comparisons were performed of epidemiologically related and unrelated plasmids based on SNP differences using snippy, phylogenetic distance using Roary and using a similarity index that penalizes size differences between plasmids (Stoesser-index). The percentage of pairwise comparisons misclassified as genetically related or as clonally unrelated was determined using different genetic thresholds for genetic relatedness for all three comparison methods.ResultsDespite the median number of SNP differences, Roary phylogenetic distance, and Stoesser-index differed between the epidemiologically related and unrelated plasmids, the range of differences overlapped between the two comparison groups for all three comparison methods. When using a genetic similarity threshold that classified 100% of epidemiologically related plasmid pairs as genetically related, the percentages of plasmids misclassified as epidemiologically related ranged from 6.7% (Roary) to 20.8% (Stoesser-index).DiscussionAlthough epidemiologically related plasmids can be distinguished from unrelated plasmids based on genetic similarity, epidemiologically related and unrelated blaKPC-containing IncF plasmids show a high degree of sequence similarity. The phylogenetic distance as determined using Roary showed the highest degree of discriminatory power between the epidemiologically related and unrelated plasmids.Impact statementAccurately distinguishing epidemiologically related from unrelated plasmids is essential to detect nosocomial plasmid transmission in outbreaks. However, limited information is available on whether blaKPC-containing plasmids from isolates in a hospital outbreak can be differentiated from epidemiologically unrelated blaKPC-containing plasmids based on sequence data. This study aimed to evaluate the performance of three approaches to distinguish epidemiologically related from unrelated blaKPC-containing IncF plasmids. Pairwise comparisons were performed of epidemiologically related and unrelated plasmids based on SNP differences using snippy, phylogenetic distance using Roary and using a similarity index that penalizes size differences between plasmids (Stoesser-index). Based on our results, epidemiologically related plasmids can be distinguished from unrelated plasmids based on genetic similarity. Despite this, epidemiologically related and unrelated blaKPC-containing IncF plasmids show a high degree of sequence similarity and judgements on the horizontal transfer of these plasmids during hospital outbreaks based on genetic identity should be made with caution. The phylogenetic distance determined using Roary showed the highest discriminatory power between the epidemiologically related and unrelated plasmids.Data summaryShort-and long-read sequence data of the epidemiologically related Enterobacteriaceae isolates included in this study are available from the publicly available European Nucleotide Archive of the European Bioinformatics Institute under study accession number: PRJEB41009. The authors confirm that all supporting data have been provided within the article and through the supplementary data files.

Download Full-text

Distinguishing blaKPC -gene-containing IncF plasmids from epidemiologically related and unrelated Enterobacteriaceae based on short- and long-read sequence data

Antimicrobial Agents and Chemotherapy ◽

10.1128/aac.00147-21 ◽

2021 ◽

Author(s):

Joep J.J.M. Stohr ◽

Marjolein F. Q. Kluytmans-van den Bergh ◽

Veronica A.T.C. Weterings ◽

John W. A. Rossen ◽

Jan A. J. W. Kluytmans

Keyword(s):

Genetic Relatedness ◽

Sequence Data ◽

Sequence Similarity ◽

Similarity Index ◽

Pairwise Comparisons ◽

Phylogenetic Distance ◽

Limited Information ◽

Hospital Outbreak ◽

Long Read ◽

High Degree

Background: Limited information is available on whether blaKPC -containing plasmids from isolates in a hospital outbreak can be differentiated from epidemiologically unrelated blaKPC-containing plasmids based on sequence data. This study aimed to evaluate the performance of three approaches to distinguish epidemiologically related from unrelated blaKPC-containing pKpQiL-like IncFII(k2)-IncFIB(pQiL) plasmids. Method: Epidemiologically related isolates, were short- and long-read whole genome sequenced. A hybrid assembly was performed and plasmid sequences were extracted from the assembly graph. Epidemiologically unrelated plasmid sequences were extracted from the GenBank. Pairwise comparisons were performed of epidemiologically related and unrelated plasmids based on SNP differences using snippy, phylogenetic distance using Roary and using a similarity index that penalizes size differences between plasmids (Stoesser-index). The percentage of pairwise comparisons misclassified as genetically related or as clonally unrelated was determined using different genetic thresholds for genetic relatedness. Results: The ranges in number of SNP differences, Roary phylogenetic distance, and Stoesser-index overlapped between the epidemiologically related and unrelated plasmids. When using a genetic similarity threshold that classified 100% of epidemiologically related plasmid pairs as genetically related, the percentages of plasmids misclassified as epidemiologically related ranged from 6.7% (Roary) to 20.8% (Stoesser-index). Discussion: Although epidemiologically related plasmids can be distinguished from unrelated plasmids based on genetic differences, blaKPC-containing pKpQiL-like IncFII(k2)-IncFIB(pQiL) plasmids show a high degree of sequence similarity. The phylogenetic distance as determined using Roary showed the highest degree of discriminatory power between the epidemiologically related and unrelated plasmids.

Download Full-text

Limited Genetic Diversity of blaCMY-2-Containing IncI1-pST12 Plasmids from Enterobacteriaceae of Human and Broiler Chicken Origin in The Netherlands

Microorganisms ◽

10.3390/microorganisms8111755 ◽

2020 ◽

Vol 8 (11) ◽

pp. 1755

Author(s):

Evert Drijver ◽

Joep Stohr ◽

Jaco Verweij ◽

Carlo Verhulst ◽

Francisca Velkers ◽

...

Keyword(s):

Escherichia Coli ◽

Sequence Data ◽

Sequence Similarity ◽

Sequencing Analysis ◽

Nucleotide Polymorphisms ◽

Pan Genome ◽

Next Generation Sequencing Analysis ◽

Long Read ◽

Short Read Sequence ◽

High Degree

Distinguishing epidemiologically related and unrelated plasmids is essential to confirm plasmid transmission. We compared IncI1–pST12 plasmids from both human and livestock origin and explored the degree of sequence similarity between plasmids from Enterobacteriaceae with different epidemiological links. Short-read sequence data of Enterobacteriaceae cultured from humans and broilers were screened for the presence of both a blaCMY-2 gene and an IncI1–pST12 replicon. Isolates were long-read sequenced on a MinION sequencer (OxfordNanopore Technologies). After plasmid reconstruction using hybrid assembly, pairwise single nucleotide polymorphisms (SNPs) were determined. The plasmids were annotated, and a pan-genome was constructed to compare genes variably present between the different plasmids. Nine Escherichia coli sequences of broiler origin, four Escherichia coli sequences, and one Salmonella enterica sequence of human origin were selected for the current analysis. A circular contig with the IncI1–pST12 replicon and blaCMY-2 gene was extracted from the assembly graph of all fourteen isolates. Analysis of the IncI1–pST12 plasmids revealed a low number of SNP differences (range of 0–9 SNPs). The range of SNP differences overlapped in isolates with different epidemiological links. One-hundred and twelve from a total of 113 genes of the pan-genome were present in all plasmid constructs. Next generation sequencing analysis of blaCMY-2-containing IncI1–pST12 plasmids isolated from Enterobacteriaceae with different epidemiological links show a high degree of sequence similarity in terms of SNP differences and the number of shared genes. Therefore, statements on the horizontal transfer of these plasmids based on genetic identity should be made with caution.

Download Full-text

Limited genetic diversity of blaCMY-2-containing IncI1-pST12 plasmids from Enterobacteriaceae of human and broiler chicken origin in the Netherlands

10.1101/2020.07.09.195461 ◽

2020 ◽

Author(s):

Evert den Drijver ◽

Joep J.J.M. Stohr ◽

Jaco J. Verweij ◽

Carlo Verhulst ◽

Francisca C. Velkers ◽

...

Keyword(s):

Escherichia Coli ◽

Sequence Data ◽

Sequence Similarity ◽

Nucleotide Polymorphisms ◽

Single Nucleotide ◽

Pan Genome ◽

Long Read ◽

Short Read Sequence ◽

High Degree ◽

Limited Genetic Diversity

AbstractDistinguishing epidemiologically related and unrelated plasmids is essential to confirm plasmid transmission. We compared IncI1-pST12 plasmids from both human and livestock origin and explored the degree of sequence similarity between plasmids from Enterobacteriaceae with different epidemiological links. Short-read sequence data of Enterobacteriaceae cultured from humans and broilers were screened for the presence of both a blaCMY-2 gene and an IncI1-pST12 replicon. Isolates were long-read sequenced on a MinION sequencer (OxfordNanopore Technologies). After plasmid reconstruction using hybrid assembly, pairwise single nucleotide polymorphisms (SNP) were determined. The plasmids were annotated, and a pan-genome was constructed to compare genes variably present between the different plasmids. Nine Escherichia coli sequences of broiler origin, four Escherichia coli sequences and one Salmonella enterica sequence of human origin were selected for the current analysis. A circular contig with the IncI1-pST12 replicon and blaCMY-2 gene was extracted from the assembly graph of all fourteen isolates. Analysis of the IncI1-pST12 plasmids revealed a low number of SNP differences (range of 0-9 SNPs). The range of SNP differences overlapped in isolates with different epidemiological links. One-hundred and twelve from a total of 113 genes of the pan-genome were present in all plasmid constructs. NGS-analysis of blaCMY--2-containing IncI1-pST12 plasmids isolated from Enterobacteriaceae with different epidemiological links show a high degree of sequence similarity in terms of SNP differences and the number of shared genes. Therefore, statements on the horizontal transfer of these plasmids based on genetic identity should be made with caution.

Download Full-text

Analisis Keragaman Genetik Acremonium yang Berasosiasi dengan Tanaman Gaharu Menggunakan Teknik Random Amplified Polymorphic DNA (RAPD)

Jurnal AgroBiogen ◽

10.21082/jbio.v5n2.2009.p65-70 ◽

2016 ◽

Vol 5 (2) ◽

pp. 65

Author(s):

Dewi Rahmawati ◽

Nurita Toruan-Mathius

Keyword(s):

Cluster Analysis ◽

Genetic Variation ◽

Genetic Similarity ◽

Random Amplified Polymorphic Dna ◽

Economic Value ◽

Similarity Index ◽

Similarity Coefficients ◽

West Sumatra ◽

High Degree ◽

Geographical Locations

Agarwood or gaharu is a plant that has a high economic value in Asia, due to its use for production of incense and traditional medicines. The agarwood formation occurs in the trunk and roots of trees that have been infected by a fungus, such as Acremonium spp. Various fungi were associated with the agarwood formation. Acremonium is generally considered as highly polyphyletic, contains distantly related fungi. A study was done to identify genetic diversities in 10 isolates of Acremonium spp. from four different areas in Indonesia that are associated with Aquilaria and Gyrinops verstigii using the Random Amplified Polymorphic DNA (RAPD) technique. Eight RAPD primers, i.e., OPA 02, OPB 04, OPB 07, OPB 17, OPC 11, OPD 03, OPD 05, and OPE 07 were used in the analyses. The results indicated that similarity index values of the genetic variation ranged from 0.21 to 0.97. Based on the Nei and Li’s similarity coefficients, these values indicating the presence of high degree of genetic variability. The lowest degree of genetic similarity were found between isolates F (Acremonium spp., which is associated with G. verstigii from Mataram, Nusa Tenggara Barat), and LM2 from south coastal area of West Sumatra. The highest genetic similarity value (0.97) was found between isolates Sr2 and Sr4 from Sorong, Papua. Results from the cluster analysis indicated that the isolates could be grouped into two major clusters that were associated with their geographical locations.

Download Full-text

Single-molecule sequencing of theDrosophila serratagenome

10.1101/090969 ◽

2016 ◽

Author(s):

Scott L. Allen ◽

Emily K. Delaney ◽

Artyom Kopp ◽

Stephen F. Chenoweth

Keyword(s):

Single Molecule ◽

De Novo ◽

Sequence Data ◽

Error Rates ◽

Model Species ◽

High Coverage ◽

Single Molecule Sequencing ◽

Long Read ◽

Drosophila Serrata ◽

High Degree

ABSTRACTLong read sequencing technology promises to greatly enhancede novoassembly of genomes for non-model species. While error rates have been a large stumbling block, sequencing at high coverage allows reads to be self-corrected. Here we sequence andde novoassemble the genome ofDrosophila serrata, a non-model species from themontiumsubgroup that has been well studied for clines and sexual selection. Using 11 PacBio SMRT cells, we generated 12 Gbp of raw sequence data comprising approximately 65x whole genome coverage. Read lengths averaged 8,940 bp (NRead50 12,200) with the longest read at 53 Kbp. We self-corrected reads using the PBDagCon algorithm and assembled the genome using the MHAP algorithm within the PBcR assembler. Total genome length was 198 Mbp with an N50 just under 1 Mbp. Contigs displayed a high degree of arm-level conservation withD. melanogaster. We also provide an initial annotation for this genome usingin silicogene predictions that were supported by RNA-seq data.

Download Full-text

Isolation, Characterisation, and Lipase Production of a Cold-Adapted Bacterial Strain Pseudomonas sp. LSK25 Isolated from Signy Island, Antarctica

Molecules ◽

10.3390/molecules24040715 ◽

2019 ◽

Vol 24 (4) ◽

pp. 715 ◽

Cited By ~ 8

Author(s):

Leelatulasi Salwoom ◽

Raja Raja Abd Rahman ◽

Abu Salleh ◽

Fairolniza Mohd. Shariff ◽

Peter Convey ◽

...

Keyword(s):

Sequence Data ◽

Sequence Similarity ◽

Lipase Production ◽

Research Station ◽

Pseudomonas Sp ◽

Limited Information ◽

Bacterial Strains ◽

Lipase Gene ◽

16S Rrna Sequence ◽

Signy Island

In recent years, studies on psychrophilic lipases have been an emerging area of research in the field of enzymology. This study focuses on bacterial strains isolated from anthropogenically-influenced soil samples collected around Signy Island Research Station (South Orkney Islands, maritime Antarctic). Limited information on lipase activities from bacteria isolated from Signy station is currently available. The presence of lipase genes was determined using real time quantification PCR (qPCR) in samples obtained from three different locations on Signy Island. Twenty strains from the location with highest lipase gene detection were screened for lipolytic activities at a temperature of 4 °C, and from this one strain was selected for further examination based on the highest enzymatic activities obtained. Analysis of 16S rRNA sequence data of this strain showed the highest level of sequence similarity (98%) to a Pseudomonas sp. strain also isolated from Antarctica. In order to increase lipase production of this psychrophilic strain, optimisation of different parameters of physical and nutritional factors were investigated. Optimal production was obtained at 10 °C and pH 7.0, at 150 rev/min shaking rate over 36 h incubation.

Download Full-text

DNA BARCODING IKAN HIAS INTRODUKSI

Jurnal Riset Akuakultur ◽

10.15578/jra.12.1.2017.29-40 ◽

2017 ◽

Vol 12 (1) ◽

pp. 29 ◽

Cited By ~ 3

Author(s):

Melta Rini Fahmi ◽

Ruby Vidia Kusumah ◽

Idil Ardi ◽

Shofihar Sinansari ◽

Eni Kusrini

Keyword(s):

Sequence Data ◽

Sequence Similarity ◽

Similarity Index ◽

Ornamental Fish ◽

Coi Gene ◽

Correct Identification ◽

Global Marketing ◽

Fish Samples ◽

Hybridization Product ◽

Eigenmannia Virescens

Identifikasi spesies menjadi tantangan dalam pengelolaan ikan hias introduksi baik untuk tujuan budidaya maupun konservasi. Penelitian ini bertujuan untuk melakukan identifikasi molekuler ikan hias introduksi yang beredar di pembudidaya dan pasar ikan hias Indonesia dengan menggunakan barcode DNA gen COI. Sampel ikan diperoleh dari pembudidaya dan importir ikan hias di kawasan Bandung dan Jakarta. Total DNA diekstraksi dari jaringan sirip ekor dengan menggunakan metode kolom. Amplifikasi gen target dilakukan dengan menggunakan primer FishF1, FishF2, FishR1, dan FishR2. Hasil pembacaan untai DNA disejajarkan dengan sekuen yang terdapat pada genbank melalui program BLAST. Identifikasi dilakukan melalui kekerabatan pohon filogenetik dan presentasi indeks kesamaan dengan sekuen genbank. Hasil identifikasi menunjukkan sampel yang diuji terbagi menjadi lima grup, yaitu: Synodontis terdiri atas lima spesies, Corydoras: empat spesies, Phseudoplatystoma: tiga spesies, Botia: tiga spesies, dan Leporinus: tiga spesies dengan nilai boostrap 99-100. Indeks kesamaan sekuen menunjukkan sebanyak 11 spesies memiliki indeks kesamaan 99%-100% dengan data genbank yaitu Synodontis decorus, Synodontis eupterus, Synodontis greshoffi, Botia kubotai, Botia lohachata, Rasbora erythromicron, Corydoras aeneus, Gyrinocheilus aymonieri, Eigenmannia virescens, Leporinus affinis, Phractocephalus hemioliopterus. Dua spesies teridentifikasi sebagai hasil hibridisasi (kawin silang) yaitu Leopard catfish (100% identik dengan Pseudoplatystoma faciatum) dan Synodontis leopard (100% identik dengan Synodontis notatus). Hasil analisis nukleotida penciri diperoleh tujuh nukleotida untuk Synodontis decora, 10 nukleotida untuk Synodontis tanganyicae, 13 nukleotida untuk Synodontis euterus, empat nukleotida untuk Synodontis notatus, dan 14 untuk Synodontis grashoffi. Kejelasan identifikasi spesies ikan menjadi kunci utama dalam budidaya, perdagangan, manajemen, konservasi, dan pengembangan ilmu pengetahuan.Species identification becomes a new challenge in the management of ornamental fish either for cultivation, or for conservation proposes. The objective of this study was to identify currently existing introduced ornamental fish in Indonesian farmers and markets using DNA barcodes COI gene. Fish samples were collected from farmers and importers of ornamental fish in Bandung and Jakarta. Total genome was extracted from caudal fin tissue using the column method. Amplification of the target gene was done by using FishF1, FishF2, FishR1, and FishR2 primers. DNA sequence was aligned with the sequences from genbank by BLAST program. Species identification was decided through the phylogenetic tree and similarity index with genbank sequences. The results showed that all of tested samples fall into five groups; Synodontis consisted of five species, Corydoras four species, Phseudoplatystoma four species, Botia three species, and Leporinus three species with 99-100 boostrap value. Sequence similarity index showed around 11 species have 99%-100% similarity index with sequence data on genbank which are Synodontis decorus, Synodontis eupterus, Synodontis greshoffi, Botia kubotai, Botia lohachata, Rasbora erythromicron, Corydoras aeneus, Gyrinocheilus aymonieri, Eigenmannia virescens, Leporinus affinis, Phractocephalus hemioliopterus. Two species were identified as hybridization product (interbreeding) including leopard catfish (100% identical with Pseudoplatystoma faciatum) and the leopard Synodontis (100% identical with Synodontis notatus). Analysis of nucleotide diagnostic showed Synodontis decora has seven nucleotides diagnostic, Synodontis tanganyicae 10 nucleotides, Synodontis euterus 13 nucleotides, Synodontis notatus four nucleotides, and Synodontis grashoffi 14 nucleotides. The correct identification of fish species is a useful tool for aquaculture, global marketing, management or conservation, and academic/scientific purpose.

Download Full-text

Ophiocordyceps salganeicola, a parasite of social cockroaches in Japan and insights into the evolution of other closely-related Blattodea-associated lineages

IMA Fungus ◽

10.1186/s43008-020-00053-9 ◽

2021 ◽

Vol 12 (1) ◽

Author(s):

João P. M. Araújo ◽

Mitsuru G. Moriguchi ◽

Shigeru Uchiyama ◽

Noriko Kinjo ◽

Yu Matsuura

Keyword(s):

Genetic Similarity ◽

Evolutionary History ◽

Fungal Species ◽

Evolutionary Relationships ◽

Ryukyu Archipelago ◽

Phylogenetic Placement ◽

The Social ◽

Host Jump ◽

High Degree ◽

A New Species

AbstractThe entomopathogenic genus Ophiocordyceps includes a highly diverse group of fungal species, predominantly parasitizing insects in the orders Coleoptera, Hemiptera, Hymenoptera and Lepidoptera. However, other insect orders are also parasitized by these fungi, for example the Blattodea (termites and cockroaches). Despite their ubiquity in nearly all environments insects occur, blattodeans are rarely found infected by filamentous fungi and thus, their ecology and evolutionary history remain obscure. In this study, we propose a new species of Ophiocordyceps infecting the social cockroaches Salganea esakii and S. taiwanensis, based on 16 years of collections and field observations in Japan, especially in the Ryukyu Archipelago. We found a high degree of genetic similarity between specimens from different islands, infecting these two Salganea species and that this relationship is ancient, likely not originating from a recent host jump. Furthermore, we found that Ophiocordyceps lineages infecting cockroaches evolved around the same time, at least twice, one from beetles and the other from termites. We have also investigated the evolutionary relationships between Ophiocordyceps and termites and present the phylogenetic placement of O. cf. blattae. Our analyses also show that O. sinensis could have originated from an ancestor infecting termite, instead of beetle larvae as previously proposed.

Download Full-text

Genome sequences of human cytomegalovirus strain TB40/E variants propagated in fibroblasts and epithelial cells

Virology Journal ◽

10.1186/s12985-021-01583-3 ◽

2021 ◽

Vol 18 (1) ◽

Author(s):

Ahmed Al Qaffas ◽

Salvatore Camiolo ◽

Mai Vo ◽

Alexis Aguiar ◽

Amine Ourahmane ◽

...

Keyword(s):

Epithelial Cells ◽

Human Cytomegalovirus ◽

Viral Entry ◽

Sequence Data ◽

Laboratory Strain ◽

Serial Passage ◽

Wild Type Virus ◽

Protein Coding ◽

Genetic Changes ◽

Long Read

AbstractThe advent of whole genome sequencing has revealed that common laboratory strains of human cytomegalovirus (HCMV) have major genetic deficiencies resulting from serial passage in fibroblasts. In particular, tropism for epithelial and endothelial cells is lost due to mutations disrupting genes UL128, UL130, or UL131A, which encode subunits of a virion-associated pentameric complex (PC) important for viral entry into these cells but not for entry into fibroblasts. The endothelial cell-adapted strain TB40/E has a relatively intact genome and has emerged as a laboratory strain that closely resembles wild-type virus. However, several heterogeneous TB40/E stocks and cloned variants exist that display a range of sequence and tropism properties. Here, we report the use of PacBio sequencing to elucidate the genetic changes that occurred, both at the consensus level and within subpopulations, upon passaging a TB40/E stock on ARPE-19 epithelial cells. The long-read data also facilitated examination of the linkage between mutations. Consistent with inefficient ARPE-19 cell entry, at least 83% of viral genomes present before adaptation contained changes impacting PC subunits. In contrast, and consistent with the importance of the PC for entry into endothelial and epithelial cells, genomes after adaptation lacked these or additional mutations impacting PC subunits. The sequence data also revealed six single noncoding substitutions in the inverted repeat regions, single nonsynonymous substitutions in genes UL26, UL69, US28, and UL122, and a frameshift truncating gene UL141. Among the changes affecting protein-coding regions, only the one in UL122 was strongly selected. This change, resulting in a D390H substitution in the encoded protein IE2, has been previously implicated in rendering another viral protein, UL84, essential for viral replication in fibroblasts. This finding suggests that IE2, and perhaps its interactions with UL84, have important functions unique to HCMV replication in epithelial cells.

Download Full-text

Predicting bacteriophage hosts based on sequences of annotated receptor-binding proteins

Scientific Reports ◽

10.1038/s41598-021-81063-4 ◽

2021 ◽

Vol 11 (1) ◽

Author(s):

Dimitri Boeckaerts ◽

Michiel Stock ◽

Bjorn Criel ◽

Hans Gerstmans ◽

Bernard De Baets ◽

...

Keyword(s):

Machine Learning ◽

Predictive Model ◽

Receptor Binding ◽

Bacterial Infections ◽

Sequence Data ◽

Sequence Similarity ◽

Area Under The Curve ◽

Local Alignment ◽

Search Tool ◽

Different Levels

AbstractNowadays, bacteriophages are increasingly considered as an alternative treatment for a variety of bacterial infections in cases where classical antibiotics have become ineffective. However, characterizing the host specificity of phages remains a labor- and time-intensive process. In order to alleviate this burden, we have developed a new machine-learning-based pipeline to predict bacteriophage hosts based on annotated receptor-binding protein (RBP) sequence data. We focus on predicting bacterial hosts from the ESKAPE group, Escherichia coli, Salmonella enterica and Clostridium difficile. We compare the performance of our predictive model with that of the widely used Basic Local Alignment Search Tool (BLAST). Our best-performing predictive model reaches Precision-Recall Area Under the Curve (PR-AUC) scores between 73.6 and 93.8% for different levels of sequence similarity in the collected data. Our model reaches a performance comparable to that of BLASTp when sequence similarity in the data is high and starts outperforming BLASTp when sequence similarity drops below 75%. Therefore, our machine learning methods can be especially useful in settings in which sequence similarity to other known sequences is low. Predicting the hosts of novel metagenomic RBP sequences could extend our toolbox to tune the host spectrum of phages or phage tail-like bacteriocins by swapping RBPs.

Download Full-text