- Volume 6, Issue 3, 2020

Bordetella pertussis , the aetiological agent of whooping cough, is re-emerging globally despite widespread vaccination. B. pertussis is highly infectious and, prior to vaccination programmes, was the leading cause of infant mortality. The WHO estimated that over 600 000 deaths are prevented annually by pertussis vaccination, but B. pertussis infection was still responsible for over 63 000 deaths globally in 2013. The re-emergence of B. pertussis has been linked to strains with inactive or absent major virulence factors included in vaccines such as pertactin, pertussis toxin and filamentous haemagglutinin. Thus, the molecular surveillance of currently circulating strains is critical in understanding and controlling B. pertussis . Such information provides data on strains to inform control measures and the identification of future vaccine antigens. Current surveillance and typing methods for B. pertussis rely on the availability of clinical isolates. However, since the 1990s, the majority of pertussis cases have been diagnosed by PCR, where an isolate is not needed. The rapid decline in the availability of B. pertussis isolates impacts our ability to monitor this infection. The growing uptake of next-generation sequencing (NGS) has offered the opportunity for culture-independent genome sequencing and typing of this fastidious pathogen. Therefore, the objective of the study was to optimize respiratory sample preparation, independent of culture, in order to type B. pertussis using NGS. The study compared commercial depletion kits and specimen-processing methods using selective lysis detergents. The goal was to deplete human DNA, the major obstacle for sequencing a pathogen directly from a clinical sample. Samples spiked with a clinically relevant amount of B. pertussis were used to provide comparison between the different methods. Commercial depletion kits including the MolYsis, Qiagen Microbiome and NEBNext Kits were tested. Previously published methods, for Saponin and TritonX-100, were also trialled as a depletion. The ratio of B. pertussis to human DNA was determined by real-time PCR for ERV3 and IS481 (as markers of human and B. pertussis DNA, respectively), then samples were sequenced using the Illumina NextSeq 500 platform. The number of human and B. pertussis sequenced reads were then compared between treatments. The results showed that commercial kits reduced the human DNA present, but also reduced the concentration of target B. pertussis . However, selective lysis with Saponin treatment resulted in almost undetectable levels of human DNA, with minimal loss of target B. pertussis DNA. Sequencing read depth improved five-fold in reads to B. pertussis . Our investigation delivered a potential protocol that will enable the public health laboratory surveillance of B. pertussis in the era of culture-independent testing.

Genome-wide association studies (GWASs) have the potential to reveal the genetics of microbial phenotypes such as antibiotic resistance and virulence. Capitalizing on the growing wealth of bacterial sequence data, microbial GWAS methods aim to identify causal genetic variants while ignoring spurious associations. Bacteria reproduce clonally, leading to strong population structure and genome-wide linkage, making it challenging to separate true ‘hits’ (i.e. mutations that cause a phenotype) from non-causal linked mutations. GWAS methods attempt to correct for population structure in different ways, but their performance has not yet been systematically and comprehensively evaluated under a range of evolutionary scenarios. Here, we developed a bacterial GWAS simulator (BacGWASim) to generate bacterial genomes with varying rates of mutation, recombination and other evolutionary parameters, along with a subset of causal mutations underlying a phenotype of interest. We assessed the performance (recall and precision) of three widely used single-locus GWAS approaches (cluster-based, dimensionality-reduction and linear mixed models, implemented in plink, pyseer and gemma) and one relatively new multi-locus model implemented in pyseer, across a range of simulated sample sizes, recombination rates and causal mutation effect sizes. As expected, all methods performed better with larger sample sizes and effect sizes. The performance of clustering and dimensionality reduction approaches to correct for population structure were considerably variable according to the choice of parameters. Notably, the multi-locus elastic net (lasso) approach was consistently amongst the highest-performing methods, and had the highest power in detecting causal variants with both low and high effect sizes. Most methods reached the level of good performance (recall >0.75) for identifying causal mutations of strong effect size [log odds ratio (OR) ≥2] with a sample size of 2000 genomes. However, only elastic nets reached the level of reasonable performance (recall=0.35) for detecting markers with weaker effects (log OR ~1) in smaller samples. Elastic nets also showed superior precision and recall in controlling for genome-wide linkage, relative to single-locus models. However, all methods performed relatively poorly on highly clonal (low-recombining) genomes, suggesting room for improvement in method development. These findings show the potential for multi-locus models to improve bacterial GWAS performance. BacGWASim code and simulated data are publicly available to enable further comparisons and benchmarking of new methods.

Coniothyrium minitans is a mycoparasite of the notorious plant pathogen Sclerotinia sclerotiorum. To further understand the parasitism of C. minitans, we assembled and analysed its genome and performed transcriptome analyses. The genome of C. minitans strain ZS-1 was assembled into 350 scaffolds and had a size of 39.8 Mb. A total of 11 437 predicted genes and proteins were annotated, and 30.8 % of the blast hits matched proteins encoded by another member of the Pleosporales, Paraphaeosphaeria sporulosa, a worldwide soilborne fungus with biocontrol ability. The transcriptome of strain ZS-1 during the early interaction with S. sclerotiorum at 0, 4 and 12 h was analysed. The detected expressed genes were involved in responses to host defenses, including cell-wall-degrading enzymes, transporters, secretory proteins and secondary metabolite productions. Seventeen differentially expressed genes (DEGs) of fungal cell-wall-degrading enzymes (FCWDs) were up-regulated during parasitism, with only one down-regulated. Most of the monocarboxylate transporter genes of the major facilitator superfamily and all the detected ABC transporters, especially the heavy metal transporters, were significantly up-regulated. Approximately 8 % of the 11 437 proteins in C. minitans were predicted to be secretory proteins with catalytic activity. In the molecular function category, hydrolase activity, peptidase activity and serine hydrolase activity were enriched. Most genes involved in serine hydrolase activity were significantly up-regulated. This genomic analysis and genome-wide expression study demonstrates that the mycoparasitism process of C. minitans is complex and a broad range of proteins are deployed by C. minitans to successfully invade its host. Our study provides insights into the mechanisms of the mycoparasitism between C. minitans and S. sclerotiorum and identifies potential secondary metabolites from C. minitans for application as a biocontrol agent.

Fowl cholera, caused by Pasteurella multocida, continues to be a challenge in meat-chicken-breeder operations and has emerged as a problem for free-range meat chickens. Here, using whole-genome sequencing (WGS) and phylogenomic analysis, we investigate isolate relatedness during outbreaks of fowl cholera on a free-range meat chicken farm over a 5-year period. Our genomic analysis revealed that while all outbreak isolates were sequence type (ST) 20, they could be separated into two distinct clades (clade 1 and clade 2) consistent with difference in their lipopolysaccharide (LPS) type. The isolates from the earlier outbreaks (clade 1) were carrying LPS type L3 while those from the more recent outbreaks (clade 2) were LPS type L1. Additionally, WGS data indicated high inter- and intra-chicken genetic diversity during a single outbreak. Furthermore, we demonstrate that while a killed autogenous vaccine carrying LPS type L3 had been successful in protecting against challenge from L3 isolates it might have driven the emergence of the closely related clade 2, against which the vaccine was ineffective. The genomic results also revealed a 14 bp deletion in the galactosyltransferase gene gatG in LPS type L3 isolates, which would result in producing a semi-truncated LPS in those isolates. In conclusion, our study clearly demonstrates the advantages of genomic analysis over the conventional PCR-based approaches in providing clear insights in terms of linkage of isolate within and between outbreaks. More importantly, it provides more detailed information than the multiplex PCR on the possible structure of outer LPS, which is very important in the case of strain selection for killed autogenous vaccines.

Over the last 35 years in the UK, the burden of Shiga toxin-producing Escherichia coli (STEC) O157:H7 infection has, during different periods of time, been associated with five different sub-lineages (1983–1995, Ia, I/IIa and I/IIb; 1996–2014, Ic; and 2015–2018, IIb). The acquisition of a stx2a-encoding bacteriophage by these five sub-lineages appears to have coincided with their respective emergences. The Oxford Nanopore Technologies (ONT) system was used to sequence, characterize and compare the stx-encoding prophages harboured by each sub-lineage to investigate the integration of this key virulence factor. The stx2a-encoding prophages from each of the lineages causing clinical disease in the UK were all different, including the two UK sub-lineages (Ia and I/IIa) circulating concurrently and causing severe disease in the early 1980s. Comparisons between the stx2a-encoding prophage in sub-lineages I/IIb and IIb revealed similarity to the prophage commonly found to encode stx2c, and the same site of bacteriophage integration (sbcB) as stx2c-encoding prophage. These data suggest independent acquisition of previously unobserved stx2a-encoding phage is more likely to have contributed to the emergence of STEC O157:H7 sub-lineages in the UK than intra-UK lineage to lineage phage transmission. In contrast, the stx2c-encoding prophage showed a high level of similarity across lineages and time, consistent with the model of stx2c being present in the common ancestor to extant STEC O157:H7 and maintained by vertical inheritance in the majority of the population. Studying the nature of the stx-encoding bacteriophage contributes to our understanding of the emergence of highly pathogenic strains of STEC O157:H7.

Interspecific hybridization has played an important role in the evolution of eukaryotic organisms by favouring genetic interchange between divergent lineages to generate new phenotypic diversity involved in the adaptation to new environments. This way, hybridization between Saccharomyces species, involving the fusion between their metabolic capabilities, is a recurrent adaptive strategy in industrial environments. In the present study, whole-genome sequences of natural hybrids between Saccharomyces cerevisiae and Saccharomyces kudriavzevii were obtained to unveil the mechanisms involved in the origin and evolution of hybrids, as well as the ecological and geographic contexts in which spontaneous hybridization and hybrid persistence take place. Although Saccharomyces species can mate using different mechanisms, we concluded that rare-mating is the most commonly used, but other mechanisms were also observed in specific hybrids. The preponderance of rare-mating was confirmed by performing artificial hybridization experiments. The mechanism used to mate determines the genomic structure of the hybrid and its final evolutionary outcome. The evolution and adaptability of the hybrids are triggered by genomic instability, resulting in a wide diversity of genomic rearrangements. Some of these rearrangements could be adaptive under the stressful conditions of the industrial environment.

The accessory genes of prokaryote and eukaryote pangenomes accumulate by horizontal gene transfer, differential gene loss, and the effects of selection and drift. We have developed Coinfinder, a software program that assesses whether sets of homologous genes (gene families) in pangenomes associate or dissociate with each other (i.e. are ‘coincident’) more often than would be expected by chance. Coinfinder employs a user-supplied phylogenetic tree in order to assess the lineage-dependence (i.e. the phylogenetic distribution) of each accessory gene, allowing Coinfinder to focus on coincident gene pairs whose joint presence is not simply because they happened to appear in the same clade, but rather that they tend to appear together more often than expected across the phylogeny. Coinfinder is implemented in C++, Python3 and R and is freely available under the GNU license from https://github.com/fwhelan/coinfinder.

Multiply antibiotic-resistant Acinetobacter baumannii infections are a global public health concern and accurate tracking of the spread of specific lineages is needed. Variation in the composition and structure of capsular polysaccharide (CPS), a critical determinant of virulence and phage susceptibility, makes it an attractive epidemiological marker. The outer core (OC) of lipooligosaccharide also exhibits variation. To take better advantage of the untapped information available in whole genome sequences, we have created a curated reference database of 92 publicly available gene clusters at the locus encoding proteins responsible for biosynthesis and export of CPS (K locus), and a second database for 12 gene clusters at the locus for outer core biosynthesis (OC locus). Each entry has been assigned a unique KL or OCL number, and is fully annotated using a simple, transparent and standardized nomenclature. These databases are compatible with Kaptive, a tool for in silico typing of bacterial surface polysaccharide loci, and their utility was validated using (a) >630 assembled A. baumannii draft genomes for which the KL and OCL regions had been previously typed manually, and (b) 3386 A. baumannii genome assemblies downloaded from NCBI. Among the previously typed genomes, Kaptive was able to confidently assign KL and OCL types with 100 % accuracy. Among the genomes retrieved from NCBI, Kaptive detected known KL and OCL in 87 and 90 % of genomes, respectively, indicating that the majority of common KL and OCL types are captured within the databases; 13 of the 92 KL in the database were not detected in any publicly available whole genome assembly. The failure to assign a KL or OCL type may indicate incomplete or poor-quality genomes. However, further novel variants may remain to be documented. Combining outputs with multilocus sequence typing (Institut Pasteur scheme) revealed multiple KL and OCL types in collections of a single sequence type (ST) representing each of the two predominant globally distributed clones, ST1 of GC1 and ST2 of GC2, and in collections of other clones comprising >20 isolates each (ST10, ST25, and ST140), indicating extensive within-clone replacement of these loci. The databases are available at https://github.com/katholt/Kaptive and will be updated as further locus types become available.

Although gene-finding in bacterial genomes is relatively straightforward, the automated assignment of gene function is still challenging, resulting in a vast quantity of hypothetical sequences of unknown function. But how prevalent are hypothetical sequences across bacteria, what proportion of genes in different bacterial genomes remain unannotated, and what factors affect annotation completeness? To address these questions, we surveyed over 27 000 bacterial genomes from the Genome Taxonomy Database, and measured genome annotation completeness as a function of annotation method, taxonomy, genome size, 'research bias' and publication date. Our analysis revealed that 52 and 79 % of the average bacterial proteome could be functionally annotated based on protein and domain-based homology searches, respectively. Annotation coverage using protein homology search varied significantly from as low as 14 % in some species to as high as 98 % in others. We found that taxonomy is a major factor influencing annotation completeness, with distinct trends observed across the microbial tree (e.g. the lowest level of completeness was found in the Patescibacteria lineage). Most lineages showed a significant association between genome size and annotation incompleteness, likely reflecting a greater degree of uncharacterized sequences in 'accessory' proteomes than in 'core' proteomes. Finally, research bias, as measured by publication volume, was also an important factor influencing genome annotation completeness, with early model organisms showing high completeness levels relative to other genomes in their own taxonomic lineages. Our work highlights the disparity in annotation coverage across the bacterial tree of life and emphasizes a need for more experimental characterization of accessory proteomes as well as understudied lineages.

Multilocus sequence typing (MLST) is one of the most commonly used methods for studying microbial lineage worldwide. However, the traditional MLST process using Sanger sequencing is time-consuming and expensive. We have designed a workflow that simultaneously sequenced seven full-length housekeeping genes of 96 meticillin-resistant Staphylococcus aureus isolates with dual-barcode multiplexing using just a single flow cell of an Oxford Nanopore Technologies MinION system, and then we performed bioinformatic analysis for strain typing. Fifty-one of the isolates comprising 34 sequence types had been characterized using Sanger sequencing. We demonstrate that the allele assignments obtained by our nanopore workflow (nanoMLST, available at https://github.com/jade-nhri/nanoMLST) were identical to those obtained by Sanger sequencing (359/359, with 100 % agreement rate). In addition, we estimate that our multiplex system is able to perform MLST for up to 1000 samples simultaneously; thus, providing a rapid and cost-effective solution for molecular typing.