Phylogenomic evaluation of Wolbachia genomes from the Darwin Tree of Life biodiversity genomics challenge

We screened uncooked genomic sequence knowledge and first assemblies for 368 insect species (204 Lepidoptera, 61 Diptera, 52 Hymenoptera, 24 Coleoptera, 9 Hemiptera, 5 Trichoptera, 4 Orthoptera, 3 Ephemeroptera, 3 Plecoptera, 2 Odonata, and 1 Neuroptera) generated by DToL for the presence of Wolbachia (S1 Desk) utilizing the small subunit ribosomal RNA (SSU rRNA) as a marker gene. Wolbachia SSU sequences have been detected in 111 (30%) of the species. This degree of an infection shouldn’t be reflective of complete incidence, the proportion of host species inclined to an infection, as just one particular person was analysed for every taxon screened. Wolbachia prevalence, the proportion of contaminated people in a inhabitants, and an infection depth differ between species and between populations inside a species [20,21]. Due to this fact, the true incidence of an infection throughout the insect biota surveyed by DToL is probably going a lot greater. Nonetheless, the measured incidence of an infection is just like earlier survey-based estimates (roughly 22%) [22,23] however, as anticipated, is decrease than estimates deploying mathematical fashions to account for sampling bias (40% to 50%) [11,24]. An infection incidence was decrease in Coleoptera (4/24, 17%) in comparison with Lepidoptera (55/204, 27%), Diptera (21/61, 34%), and Hymenoptera (23/52, 44%) (Fig 1A).

Fig 1. Prevalence and relative abundance of Wolbachia in DToL insect genomes.

(A, B) Prevalence of Wolbachia in insect hosts, cut up by taxonomic order (A) and by intercourse (B). The cladogram of insect ordinal relationships is predicated on Misof and colleagues [28]. Orders with greater than 10 analysed species are proven in daring. Silhouettes are from PhyloPic (http://phylopic.org/). Intercourse of bugs was categorized as F (feminine), M (male), or U (unknown, the place not recorded on assortment). The information underlying this Determine will be present in S1 Knowledge. (C, D) The estimated variety of Wolbachia genomes per copy of the host nuclear genome cut up by taxonomic order (C) and by intercourse (D). The information underlying this Determine will be present in S1 Knowledge.

https://doi.org/10.1371/journal.pbio.3001972.g001

Though maternal inheritance requires that Wolbachia are predominantly localised within the germline, tropism to somatic cell sorts has been proven to be extremely regulated throughout host improvement [25,26]. We didn’t observe a bias in an infection degree by analysed tissue sort (S1 Fig), or by gender, with an equal prevalence of an infection in samples recognized as feminine (39/138, 28%) and male (45/153, 29%) (Fig 1B). Whereas the DToL challenge goals to sequence eukaryotes from throughout Britain and Eire, 82% of the samples screened have been sampled from the Wytham Woods Ecological Observatory, Oxfordshire (https://www.wythamwoods.ox.ac.uk/) [27]. No correlation between sampling location and an infection degree was detected, with 29% of all samples collected in Wytham Woods being Wolbachia optimistic, reflective of the general incidence degree (S2 Fig).

The DToL species have been sequenced utilizing PacBio Sequel II HiFi extremely correct lengthy learn platform, producing consensus uncooked reads of 10 to twenty kb with base degree accuracy of >99% (roughly Q30 to Q40). These lengthy, correct reads are perfect for meeting, notably for bacterial genomes the place the data content material per base is greater than in repeat-rich eukaryotes. The common sequence size of HiFi reads recognized as being derived from Wolbachia was 12 kb, indistinguishable from host HiFi reads. We separated and assembled all Wolbachia reads in every optimistic pattern and screened these assemblies to establish full genomes. We generated 110 full genomes, from 93 species, of which 77 have been round (S2 Desk). The common completeness of those genomes, assessed utilizing BUSCO, was 99.3%, with a imply duplication degree of 0.37%. The imply genome measurement of the brand new genomes was 1.47 Mb, which is considerably bigger than the common genome measurement of public Wolbachia genomes (1.32 Mb; Wilcoxon rank sum take a look at, p-value = 4.576 × 10⁻⁹) (S3 Fig). That is probably as a result of it’s potential to assemble throughout repeated loci (corresponding to built-in Wolbachia phage) with the lengthy, correct HiFi reads. The imply variety of contigs generated for the 33 genomes that might not be circularised was 2.12 (starting from 1 to six).

The dataset contains the primary full round Wolbachia genomes assembled from two insect orders, Odonata (dragonflies and damselflies) and Orthoptera (grasshoppers and crickets). Each species of dragonfly surveyed (Odonata) harboured Wolbachia (Fig 1A). The biggest round Wolbachia genome generated, 2.19 Mb, was remoted from the blue-tailed damselfly. That is the longest full Wolbachia genome but reported (S3 Fig). Though in most samples an infection by solely a single Wolbachia pressure was detected, 15 of 93 specimens (16%) have been contaminated with not less than two Wolbachia genomes. Inside Phalera bucephala (Lepidoptera) and Lasioglossum morio (Hymenoptera), three genomes have been assembled, whereas all different coinfections concerned two strains.

Having chromosomally full insect host genomes, in addition to full Wolbachia, permits for the estimation of the relative numbers of Wolbachia genomes per host genome. Wolbachia proliferation appears to be tightly managed and a relative abundance under ten Wolbachia genomes per host nuclear genome was noticed in most contaminated hosts. Significantly excessive abundances have been noticed in Thymelicus sylvestris and Athalia cordata (48 and 47 Wolbachia per host genome, respectively) (S2 Desk) (Fig 1C). The imply relative abundance in numerous taxonomic orders lay between 3 and 12, aside from the 2 crickets (Orthoptera), Chorthippus brunneus and Chorthippus parallelus, which have a 33 and 20 Wolbachia genome copies per host genome, respectively (Fig 1C). No important distinction was noticed between relative Wolbachia abundance and intercourse of the host (Fig 1D), with each female and male having a imply between 9 and ten copies.

We chosen 93 high-contiguity and high-completeness Wolbachia genomes from the general public INSDC databases, together with genomes from Wolbachia infecting Nematoda (13 genomes), Arachnida [4], Isopoda [1], and a number of other orders of Hexapoda [75] (S3 Desk). Including the 110 newly assembled genomes yielded a dataset of over 200 high-quality assemblies. We annotated all protein-coding genes in these genomes utilizing Prodigal [29] and clustered the expected protein units into orthologous teams utilizing OrthoFinder2 [30]. The ensuing 634 near-single copy genes have been used to deduce a phylogeny of Wolbachia (Figs 2A and S4). From this phylogeny, we assigned every genome to the beforehand outlined Wolbachia supergroups [12,13]. All newly assembled Wolbachia genomes belonged to both supergroup A or B. Whereas Lepidoptera have been predominantly contaminated with supergroup B Wolbachia (42/53, 80%), Wolbachia supergroup A was most frequent in all different insect courses (46/57, 81%). It has been beforehand noticed that supergroup B is the commonest Wolbachia sort in Lepidoptera [22,31–33]. Of the 15 species the place coinfections occurred, Endotricha flammealis, Phalera bucephala, Philonthus cognatus, Protocalliphora azurea, and Sphaerophoria taeniata have been coinfected with strains from each A and B supergroups, and the opposite ten coinfections have been of distinct strains throughout the identical supergroup (S2 Desk).

Fig 2. Wolbachia DToL genomes broaden recognized phylogeny.

(A) Round phylogeny of supergroup A and B Wolbachia, visualised with the foundation positioned between the A and B supergroups and the remaining supergroups (C, D, E, F, J, S; nodes collapsed as gray wedge), highlighting newly sequenced genomes (black tip labels) and genomes from public databases (white). (B) Incongruence between host topology (left) and supergroup A and B Wolbachia topology (proper) is proven as a tanglegram. Overview of the supergroups infecting numerous insect orders is given in a desk (inset, backside proper). A purple field is drawn to level to a bunch switching occasion; see panel C. (C) Instance of a bunch switching occasion, the place the Wolbachia of the hoverfly Eupeodes latifasciatus has excessive nuclear sequence identification and genome colinearity to 4 Wolbachia genomes assembled from Lepidoptera.

https://doi.org/10.1371/journal.pbio.3001972.g002

Wolbachia usually don’t present strict cophylogeny with their hosts [7,21]. This sample was additionally noticed when evaluating host and Wolbachia phylogenies for the supergroup A and B genomes (Fig 2B). Carefully associated insect species could also be contaminated by dissimilar Wolbachia strains, and, conversely, intently associated Wolbachia can infect a various set of bugs. For instance, the Wolbachia strains infecting the hoverfly Eupeodes latifasciatus and 4 Lepidoptera (Pararge aegeria, Celastrina argiolus, Hylaea fasciara, and Watsonella binaria) (Fig 2C) share over 99% nucleotide identification. Though horizontal transmission appears to have been a dominant sample within the evolutionary historical past of Wolbachia, the propensity of Lepidoptera to be contaminated by Wolbachia sort B underlines the significance of distribution by cospeciation. As a result of most of our new samples got here from a single website (Wytham Woods Genomic Observatory), we have been additionally in a position to discover the horizontal switch of Wolbachia between hosts in a neighborhood context. Wytham Woods–derived Wolbachia have been no extra prone to be associated than every other Wolbachia subset (S5 Fig).

The completeness of the brand new genomes and, particularly, the round assemblies achieved for 77 of them permits analyses of genome properties that aren’t potential with fragmented and partial genomes. All circularised genomes, together with these from public databases, have been rotated to start out on the presumed origin of replication. The common pairwise whole-genome nucleotide identification between all Wolbachia genomes ranged between 77.3% and 100.0%, with not less than 92.8% and 93.5% identification inside supergroups A and B, respectively (Fig 3A). The variety of breakpoints interrupting pairwise whole-genome alignments was counted, normalised for the whole alignable size, and in comparison with common nucleotide identification (ANI) of the in contrast genomes (Fig 3A). A big correlation was noticed between nucleotide divergence and the variety of breakpoints in supergroups A (0.90, p < 2.2 × 10⁻¹⁶, Spearman correlation) and B (0.69, p < 2.2 × 10⁻¹⁶, Spearman correlation) (Fig 3A). This broad vary of nucleotide variety, even inside a supergroup, is indicative of the low degree of conserved synteny inside supergroups and the extent of rearrangements occurring.

Fig 3. Comparative genomics of Wolbachia.

(A) Complete-genome common nucleotide identification (ANI) plotted towards the variety of breakpoints in comparisons inside A supergroup genomes, inside B, between A and B and between different supergroup Wolbachia. The information underlying this Determine will be present in S1 Knowledge. (B) Index of skewness in comparison with GC content material for all circularised Wolbachia genomes. The information underlying this Determine will be present in S1 Knowledge.

https://doi.org/10.1371/journal.pbio.3001972.g003

Secure bacterial genomes accumulate extra guanines than cytosines on the strand within the path of replication. This phenomenon, GC skew, arises on account of differential mutation pressures on main versus lagging strands. Genomes which have undergone frequent rearrangement are anticipated to have decrease general GC skew, which will be summarised throughout the genome as a single metric, SkewI [34]. Genomes from supergroups A and B had distinct GC contents (Fig 3B), with supergroup A having the next imply GC (35.2%, normal deviation 0.15%), in comparison with B (34.0%, normal deviation 0.16%) (two-sample t take a look at p-value < 2.2 × 10⁻¹⁶). Genomes from different supergroups had distinct GC content material, typically very totally different from A and B genomes, however as so few examples have been sequenced, basic patterns should not discernible. In each A and B supergroups, SkewI values have been comparatively low, however genomes from Wolbachia from nematode hosts (C, D, J) had greater SkewI values (Fig 3B). A excessive diploma of GC skew was beforehand reported in supergroup C Wolbachia strains infecting filarial nematodes [35], and these genomes even have low rearrangement ranges and excessive gene-level synteny. In supergroups A and B, the low degree of skew is related to excessive ranges of chromosomal rearrangement (Fig 3A).

Wolbachia, as a result of they’re sheltered throughout the cells of their hosts, could also be comparatively remoted from different micro organism and thus have considerably closed pan-genomes. One path to acquisition and sharing of latest genes is thru the Wolbachia phage (WO phage), which alongside the important phage particle structural genes carry a cargo of genes which were implicated in host manipulation. We reannotated all 203 Wolbachia with the identical, normal gene discovering toolkit, Prodigal, to normalise annotations. Whereas this may increasingly have misplaced cautious handbook revision in beforehand decided gene units, it avoids points of knowledge incompatibility. Gene quantity correlated with genome measurement and the common gene quantity within the newly assembled set of supergroup A and B Wolbachia was bigger than in A and B genomes from the general public databases (S6 Fig). Evaluating all genomes, the imply variety of predicted genes was bigger in supergroup B (1,467) in comparison with A (1,385).

We used OrthoFinder with default settings to outline clusters of orthologous proteins throughout all Wolbachia genomes. Every genome contained between 0 and 184 novel, strain-specific genes (common 19). These novel genes have been shorter than all genes (common gene size general was 875 nucleotides or roughly 290 amino acids, whereas novel genes averaged 434 nucleotides or roughly 145 amino acids). As anticipated, supergroups that weren’t nicely represented typically contained extra strain-specific genes. For instance, wCfeT from supergroup E (which infects cat fleas, Ctenocephalides felis) uniquely encoded genes for pantothenate (panC-panG-panD-panB) [36] and thiamine (thiG-thiC) biosynthesis. Nonetheless, out of the ten genomes with most strain-specific genes, seven belonged to both supergroup A or B. These novel genes weren’t preferentially related to WO phage areas (S7 Fig), however the majority (78%) had annotations that related them with transposon and cell aspect operate. This implies that a lot of the novelty is related to cell components apart from WO phage, however we word that the growth in gene quantity could also be on account of cell element-driven pseudogenisation. Apart from clusters with one or two members, probably the most often noticed cluster sizes have been 203 ± 2. These clusters contained the single-copy (and near-single-copy) orthologs deployed in phylogenetic analyses (Fig 4A). Total, the vast majority of the proteins encoded within the Wolbachia genomes have been members of orthology clusters that have been current in not less than 95% of all strains.

Fig 4. Exploration of Wolbachia protein-coding gene variety.

(A) Histogram of protein household measurement per supergroup. The information underlying this Determine will be present in S1 Knowledge. (B) Rarefaction evaluation of pan- and core proteomes of supergroups A and B, based mostly on 500,000 random addition-order permutations of co-occurring orthogroups excluding novel genes. The information underlying this Determine will be present in S1 Knowledge. (C) Synteny of the biotin cluster reveals conserved gene order and punctuated sample of species presence (inset, species with biotin cluster current are highlighted with purple circles).

https://doi.org/10.1371/journal.pbio.3001972.g004

The considerable sampling of supergroup A and B genomes allowed us to deal with and examine the sizes of the core- and pan-proteomes of those teams. The bigger genome and proteome measurement present in supergroup B was mirrored in a bigger core proteome (Fig 4B), however supergroup A had a bigger pan-proteome (Fig 4B). Whereas the core proteomes differed, only a few of the protein households that have been a part of every supergroup’s core proteome have been distinctive to that supergroup. One supergroup-restricted set of protein households was discovered to comprise the operon for arginine transport (ArtM, ArtQ, and ArtP and the repressor of arginine degradation ArgR) [37], which was uniquely detected and conserved in supergroup A (current in 83/103 or 80% of all Wolbachia A genomes). Though the periplasmic arginine-specific binding protein (ArtI or ArtJ) was not detected, the presence of this ATP-binding cassette-type (ABC) transporter means that these Wolbachia are buying arginine from their hosts.

The operon-producing biotin (vitamin B7) [38] was detected in seven of the 110 new genomes, all belonging to supergroup A (Fig 4C). One derived from Icerya purchasi (Hemiptera), and 6 have been from Hymenoptera (two from Lasioglossus malacharum, which carried two strains, and single strains from three Andrena and a Nomada species). The biotin synthesis cluster has been described beforehand from a restricted however numerous set of supergroups, together with two A genomes from extra Nomada bee hosts. This distribution suggests potential ecological linkage [39], as Andrena bees are kleptoparasitised by Nomada cuckoo bees and phylogenetic analyses of each the biotin gene clusters and the Wolbachia core proteomes present shut relationships between these clusters of genomes (S8 Fig). The gene cluster is strongly conserved in bodily organisation of all six obligatory genes (bioA-D,F,H). Within the genomic area instantly surrounding the operon, we recognized recombinase and transposase genes, in addition to ankyrin repeat containing genes and toxin–antitoxin CI Cin gene pairs. In three genomes (from Andrena dorsata, Nomada fabricium, and one of many L. malacharum strains), the operon was independently disrupted by transposases. The area containing the biotin operon thus has the hallmarks of a “virulence island” that could be cell between genomes and will have accrued extra genes (ankyrin, Cin) that hitchhike with the biotin operon.

Wolbachia can itself be contaminated by double-stranded DNA temperate bacteriophages, WO phage, which might combine within the genome of its host as a prophage. 4 modules are obligatory for building and performance of phage particles throughout the lytic stage: head, baseplate, tail, and fibre, and inserted and pseudogenised WO phage will be recognized and discriminated based mostly on the presence and completeness of those parts. Areas of a Wolbachia genome flanked by WO phage modules are prone to type parts which are transduced by the phage throughout an infection of latest cells, “cargo” loci that type the eukaryotic affiliation module (EAM) [40,41]. All of the Wolbachia genomes have been screened for prophage areas utilizing important module genes from beforehand annotated WO insertions (S4 Desk). Prophage areas have been deemed putatively full when all 4 modules have been noticed with not less than 80% of genes of every module current. An abundance of putative intact and pseudogenised WO phage have been recognized. For instance, the supergroup B Wolbachia from Ischneura elegans (the bluetail damselfly; the biggest Wolbachia genome assembled) contained three putative intact prophage and 9 WO phage fragments (Fig 5A) summing to 0.8 Mb of the genome.

Fig 5. WO prophage in Wolbachia.

(A) Annotation of the WO prophage built-in within the genome of the Wolbachia pressure infecting Ischnura elegans. (B) Wolbachia genome measurement is strongly correlated with built-in prophage span in supergroups with WO phage affiliation. Phylogenetic generalised least squares (PGLS) analyses have been carried out to evaluate the correlation between prophage size and genome measurement in a phylogenetically conscious method. The information underlying this Determine will be present in S1 Knowledge.

https://doi.org/10.1371/journal.pbio.3001972.g005

The fraction of complete prophage area in every genome ranged from 0% to 38%. Nematode-associated Wolbachia sometimes should not contaminated by WO phage [42], and no prophage areas have been detected in genomes of supergroups C, D, J, and nematode-infecting F (Fig 5B). A big correlation was discovered between genome measurement and WO prophage span in supergroups A and B (Fig 5B). This affiliation was sturdy to correction for phylogenetic relatedness of the genomes (mannequin match elevated to 0.84 and 0.87, respectively, with p-values <10⁻¹⁶).

We recognized a number of potential cargo genes inside intact and fragmented prophage. These included transposases and integrases related to cell components, and different loci beforehand related to eukaryotic manipulation, corresponding to CI loci and ankyrin repeat containing genes, as anticipated from the EAM mannequin [40,41].

Wolbachia produces a set of poisons [43] that may have dramatic results on their hosts, corresponding to CI. The CI phenotype is brought on by two adjoining genes, CifA and CifB, which operate as a toxin–antitoxin pair [44,45]. Phylogenetic evaluation categorized most Wolbachia Cif gene pairs into 4 sorts (I to IV) [46]. A fifth sort (V) is rather more variable in construction. The toxin element can have nuclease exercise (wherein case the gene pair is often known as CinA-CinB), deubiquitinase (CidA-CidB), or each (CndA,CndB) [47]. All sort II, III, and IV pairs have nuclease domains, whereas all sort I’ve deubiquitinase and most have nuclease [46]. 300 and 5 full-length and certain purposeful Cif pairs have been detected in 140 of the 181 (77%) supergroup A and B genomes. One Cif pair was detected in most genomes, however many had a number of, with seven copies within the Wolbachia pressure infecting the holly tortrix moth (Rhopobota naevana). A lot of the gene pairs contained a deubiquitinase area (sort I, Cid) [87] or belonged to sort V (90), whereas the opposite three sorts occurred in roughly equal proportions (II: 39, III: 44, IV: 34) (S9 and S10 Figs). Many pairs (213/305; 70%) have been positioned within the predicted EAM of the prophage.

Loci encoding extra toxins corresponding to RelE/RelB and latrotoxin have been recognized in a number of Wolbachia genomes, often in prophage areas (175/586 [30%], 130/256 [51%] genes, respectively) (summarised in S5 Desk). The Tc pore-forming toxin complicated, which consists of two genes TcA (S11 Fig) and TcB-C (S12 Fig), was detected in a restricted variety of A and B supergroup genomes and likewise confirmed a predisposition to happen inside prophage (42/69 [61%] and 19/35 [54%], respectively). Extra toxin-encoding loci had restricted presence in numerous subgroups and weren’t related to prophage areas. ParD/ParE (S13 and S14 Figs) solely occurred in supergroups A, B, and E, and FIC (S15 Fig) solely occurred in supergroups A, E, F, and S. The sort IV toxin–antitoxin gene pair AbiEii/AbiGii-AbiEi, which protects towards the unfold of phage an infection [48], was solely detected in two genomes in supergroup E. It’s noteworthy that these two genomes had very low ranges of prophage-derived DNA (4.3% of their genome span).

DToL uncooked knowledge are generated from entire or partial single specimens and thus comprise sequence from any cobionts in or on the specimen on the time of sampling. We screened knowledge for 368 insect genomes generated by the DToL challenge [16] for the presence of the intracellular endosymbiont Wolbachia (S1 Desk) utilizing a marker gene scan method by trying to find the SSU rRNA locus. The prokaryotic 16S rRNA alignment from RFAM (RF00177) [58] was remodeled right into a HMMER profile, and the profile was used to display contigs with nhmmscan [59]. We outlined a optimistic match as having an e-value <10⁻¹⁵⁰ or an aligned size of >1,000 nucleotides. Putative optimistic areas have been extracted from the sequences and in comparison with the SILVA SSU database (model 138.1) [60] utilizing sina [61]. Matches have been filtered to retain solely these with >90% identification. Taxonomic classification of every optimistic was decided by way of a consensus rule of 80% of the highest 20 greatest hits, utilizing each the NCBI [62] and SILVA [63] taxonomies.

For Wolbachia-positive samples, all PacBio HiFi reads have been analysed utilizing kraken2 [64] with a customized database consisting of a genome from a species intently associated to the host, all RefSeq genomes of Anaplasmataceae, and reference genomes of moreover detected cobionts downloaded utilizing NCBI datasets and masked utilizing dustmasker [65]. Horizontal switch of fragments of endosymbiont and organellar DNA to the nuclear genome is a typical phenomenon. To keep away from inadvertently classifying nuclear Wolbachia insertions (NUWTs) as deriving from an unbiased bacterial replicon, Wolbachia reads recognized by kraken2 have been mapped to the insect genome meeting, and solely contigs totally lined by these reads have been retained. The Wolbachia reads have been additionally independently reassembled utilizing a number of meeting instruments: flye (model 2.9) (flye—pacbio-hifi {reads} -o {dir} -t {threads}—asm-coverage 50—genome-size 1.6m —scaffold) [66], hifiasm (model 0.14) (hifiasm -o {prefix} -t {threads} {reads} -D 10 -l 1 -s 0.999) [67], and hifiasm-meta (model 0.1-r022) (hifiasm_meta -o {prefix} -t {threads} {reads} -l 1) [68]. The a number of assemblies generated for every pattern have been ranked based mostly on their completeness utilizing BUSCO model 5.2.2 [69] and the Rickettsiales_odb10 dataset, alignment to reference genomes utilizing nucmer (model 4.0.0) [70], evenness of protection, and circularity. One of the best (most full, single-contig round most well-liked) meeting per pattern was chosen. For samples the place 10X Genomics Chromium knowledge have been out there, sprucing was carried out utilizing FreeBayes-called variants [71] from 10X quick reads aligned with LongRanger. The host origin, span, and completeness of all Wolbachia detected are introduced in S2 Desk.

All out there Wolbachia genomes have been downloaded from NCBI GenBank on 01/02/2022 and supplemented with assemblies generated from short-read insect datasets by Scholz and colleagues [14]. This dataset contained replicate genomes for very intently associated Wolbachia from the identical host, and lots of fragmented and partial assemblies. Solely probably the most contiguous meeting per host species was retained. These genomes have been renamed utilizing the schema “R_Xyz_GenSpec_§”, the place Xyz is the primary three letters of the insect order of the host, GenSpec is an abbreviation derived from the generic and particular epithets of the host, and § signifies the supergroup. Retained assemblies have been assessed for the presence of contamination by performing a contig evaluation by kraken2 utilizing a database of solely round Wolbachia genomes. A listing of all eliminated contigs will be present in S3 Desk. Moreover, we solely included database-sourced Wolbachia genomes with not less than 90% BUSCO completeness [69] and at most 3% duplication with the Rickettsiales_odb10 dataset (S3 Desk). The exception to this filtering was the inclusion of genomes belonging to probably the most divergent supergroup S.

The entire publicly out there and newly assembled genomes have been annotated utilizing Prodigal (model 2.6.3) [29]. Protein households have been inferred utilizing OrthoFinder (model 2.4.0) [30]. We recognized 624 protein households, which have been single-copy in additional than 95% of all Wolbachia genomes. These have been individually aligned utilizing mafft in automated mode (model 7.490) [72]. Particular person most chance gene timber have been calculated utilizing iqtree (model 2.1.4) (iqtree -s {alignment} -nt {threads}) [73], and coalescence of those gene timber was decided utilizing ASTRAL (model 5.7.4) [74]. The person alignments have been trimmed utilizing trimAl (model 1.4) [75] and concatenated to type a supermatrix. This was used to deduce a most chance phylogeny with iqtree utilizing 1,000 ultrafast bootstrap approximation iterations (model 2.1.4) (iqtree -s {supermatrix} -m LG+G4 -bb 1000 -nt {threads}) [73]. The insect topology was subsampled from Chesters [76]. Incongruence in topology between the insect host and Wolbachia, host phylogeny was decided with ggtree in R [77].

All round genomes have been rotated to start out with HemE (OG0000716) on the optimistic strand, as this gene is positioned subsequent to the origin of replication [78]. All pairwise alignments have been calculated utilizing nucmer (model 4.0.0) [70], and breakpoints have been inferred and adjusted for the aligned protection. Complete-genome common nucleotide variety was calculated utilizing FastANI (model 1.33) [79]. GC and GC skew index values have been calculated for all genomes utilizing SkewIT [34].

To functionally annotate predicted genes, each Prokka (model 1.14.6) [80] and InterProScan (model 5.54–87.0) [81] have been run. The synteny plot of the biotin locus was created utilizing gggenes [82]. All six genes that make up the biotin locus (BioA-D, BioF, BioH) have been individually aligned with mafft in automated mode (model 7.490) [72] and remodeled right into a concatenated nucleotide alignment. A phylogenetic tree was constructed utilizing the mannequin GTR+F+G4 in iqtree (model 2.1.4) [73]. Genes answerable for CI have been recognized by a BLAST search [83] utilizing the next genes as queries: CidA: WP_010962721.1, WP_182158704.1, WP_012673228.1, WP_006014162.1, CAQ54402.1, NZ_MUIX01000001.1_1324, OAM06111.1; CifB: WP_010962722.1, WP_182158703.1, WP_012673227.1, WP_006014164.1, CAQ54403.1, NZ_MUIX01000001.1_1323, OAM06112.1. Furthermore, extra CifB sort V genes have been added as reference genes (Diachasma_alloeum_pair1, Diploeciton_nevermanni_pair5, wBor_pair2, wStri_pair1, wStri_pair2 and wTri-2_pair1). Solely pairs of recognized neighbouring genes (e-value 1 × 10⁻³⁰, protection 80% to 120%) have been retained. Each CifA and CifB have been aligned utilizing mafft in automated mode (model 7.490) [72], adopted by most chance estimation utilizing iqtree (model 2.1.4) (iqtree -s {alignment} -nt {threads} -bb 10000).

A listing of recognized prophage sequences was generated based mostly on annotated areas described within the literature [41,44,84] (S4 Desk) for a set of genomes (R_Dip_DroSim_A, R_Hym_NasVit_A, R_Dip_DroAna_A, R_Dip_HaeIrr_A, R_Hym_CerSol_A, and R_Hym_WiePum_A) and linked to their respective gene households. Every Wolbachia genome was screened for steady stretches of linked prophage genes with at most 5 different genes in-between, and these have been annotated as prophage areas in the event that they contained not less than one gene from one of many 4 core phage modules (head, baseplate, tail, and fibre). This permitted detection of novel prophage-associated genes. Areas that contained not less than 5 of 6 head, 7 of 8 baseplate, 5 of 6 fibre, and 5 of 6 tail module genes have been deemed putatively full. Genomic maps of prophage integration have been created with circos [85]. Phylogenetic generalised least squares analyses have been carried out to evaluate the correlation between prophage size and genome measurement utilizing the ape R package deal [86], utilizing a Brownian mannequin of evolution and the phylogenetic tree in Fig 2A. R squared values have been calculated utilizing the package deal rr2 [87].