Decoding huge phage diversity: a taxonomic classification of Lak megaphages

Abstract

High-throughput sequencing for uncultivated viruses has accelerated the understanding of global viral diversity and uncovered viral genomes substantially larger than any that have so far been cultured. Notably, the Lak phages are an enigmatic group of viruses that present some of the largest known phage genomes identified in human and animal microbiomes, and are dissimilar to any cultivated viruses. Despite the wealth of viral diversity that exists within sequencing datasets, uncultivated viruses have rarely been used for taxonomic classification. We investigated the evolutionary relationships of 23 Lak phages and propose a taxonomy for their classification. Predicted protein analysis revealed the Lak phages formed a deeply branching monophyletic clade within the class Caudoviricetes which contained no other phage genomes. One of the interesting features of this clade is that all current members are characterised by an alternative genetic code. We propose the Lak phages belong to a new order, the 'Grandevirales'. Protein and nucleotide-based analyses support the creation of two families, three sub-families, and four genera within the order 'Grandevirales'. We anticipate that the proposed taxonomy of Lak megaphages will simplify the future classification of related viral genomes as they are uncovered. Continued efforts to classify divergent viruses are crucial to aid common analyses of viral genomes and metagenomes.

Keywords: lak; megaphage; phage genomics; phage taxonomy.

Fig. 1.. Proteomic tree of megaphages amongst currently classified Duplodnaviria. (a) ViPTree proteomic tree of ‘megaphages’ and Duplodnaviria with viral family shown in the coloured ring. Blue bar chart (inner) represents genome length and red bar chart (outer) shows difference in coding capacity when using translation table 15 rather than 11 (i.e. coding capacity of 90 % using 15 and 70 % using 11 would lead to a difference of 20). The ‘Grandevirales’ members are shown with red branches and other megaphages are shown with green and highlighted by a star with corresponding colour. (b) A pruned tree showing the ‘Grandevirales’ with nearest sister clades only. The distances shown in ViPTree were calculated from genomic distances based upon normalised tBLASTx scores and the tree was rooted at the mid-point.

Fig. 2.. Shared protein clusters for phages of ‘Grandevirales’. Heatmap showing presence/absence of proteins clustered at 70 % identity with the dendrogram showing hierarchical clustering. Colour strips on the y-axis show proposed taxonomy, and colour strips on the x-axis show predicted function of the protein cluster derived from PHROGs. X-axis labels show strain name with proposed species name in brackets.

Fig. 3.. Core genome phylogeny of ‘Lakviridae’ phages. A concatenated protein phylogeny of translated sequences of 72 ‘core’ genes present on the 22 members of the proposed family ‘Lakviridae’. Proposed species names for each strain are indicated in brackets. Alignments were performed using MAFFT, and the tree was produced using IQ-Tree with 1000 rapid bootstraps and -m TEST to optimize model fits for each alignment. Tree is rooted at the midpoint and bootstraps ≥95 % are shown. The coloured strips indicate proposed genera and subfamilies. Node labels are based on strain names with proposed species names shown in brackets.

Fig. 4.. Intergenomic similarity of ‘Grandevirales’ phages. Heatmap showing intergenomic distance (ANI multiplied by aligned fraction) of ‘Grandevirales’ phages with proposed taxonomy shown in coloured strips. Those shown as having a similarity of 100.0* are highly similar but not identical with a low number of single nucleotide variations between them. Axis labels are based on strain names with proposed species names shown in brackets.