A hybrid and poly-polish workflow for the complete and accurate assembly of phage genomes: a case study of ten przondoviruses

Abstract

Bacteriophages (phages) within the genus Przondovirus are T7-like podoviruses belonging to the subfamily Studiervirinae, within the family Autographiviridae, and have a highly conserved genome organisation. The genomes of these phages range from 37 to 42 kb in size, encode 50-60 genes and are characterised by the presence of direct terminal repeats (DTRs) flanking the linear chromosome. These DTRs are often deleted during short-read-only and hybrid assemblies. Moreover, long-read-only assemblies are often littered with sequencing and/or assembly errors and require additional curation. Here, we present the isolation and characterisation of ten novel przondoviruses targeting Klebsiella spp. We describe HYPPA, a HYbrid and Poly-polish Phage Assembly workflow, which utilises long-read assemblies in combination with short-read sequencing to resolve phage DTRs and correcting errors, negating the need for laborious primer walking and Sanger sequencing validation. Our assembly workflow utilised Oxford Nanopore Technologies for long-read sequencing for its accessibility, making it the more relevant long-read sequencing technology at this time, and Illumina DNA Prep for short-read sequencing, representing the most commonly used technologies globally. Our data demonstrate the importance of careful curation of phage assemblies before publication, and prior to using them for comparative genomics.

Keywords: Klebsiella; Przondovirus; assembly; bacteriophage; phage; sequencing.

Fig. 1.

Heatmap for host range of the przondoviruses in the collection by plaque assay against a diverse range of Klebsiella spp. Top panel: isolate type, capsular type and sequence type. The source of each isolate is given as isolate type, with grey indicating an unknown source. Capsular loci determined by Kaptive and/or Kleborate, green; unknown or no match confidence, grey. Sequence type (ST) was determined by MLST, blue; unknown or incomplete matches, grey. No sequencing data available, undetermined. Bottom panel: host range heatmap. Productive infection (positive) is the observation of individual plaques, purple; lysis without productive infection is the observation of clearance without individual plaques, green; no productive infection or clearance (negative), yellow.

Fig. 2.

Genome map and gene clustering for przondoviruses in the collection and a selection of related phages. Arrows represent coding sequences and pairwise comparisons of gene similarities are indicated by percentage identity given as links in greyscale, with darker shading representing areas of higher similarity. Genes without any sequence similarity are indicated without links. Some phages had a hypothetical protein following the tail fibre protein and protein blast revealed high homology to tail spike proteins. DTRs are present but not annotated.

Fig. 3.

Nucleotide-based intergenomic similarities of przondoviruses in the collection and a selection of related phages within the subfamily Studiervirinae, using VIRIDIC. A heatmap of hierarchical clustering of the intergenomic similarity values was generated and given as percentages (right half, blue–green heatmap). Each genome pair is represented by three values (left half), where the top and bottom (blue scale) represent the aligned genome fraction for the genome in the row and column, respectively, where darker colour indicates that a lower fraction of the genome was aligned. The middle value (grey scale) represents the genome length ratio for each genome pair, where darker colour indicates increasing distance between phages. The przondoviruses within our collection are highlighted in blue–grey. Yersinia phage vB_YenP_AP10 is in the genus Apdecimavirus.