{"paper_id":"5feb8af0-68de-4eb0-9c0e-32d36c0cb0b0","body_text":"1  \nSingle-cell sequencin g reveals unexpected gen etic diversity among Bodo spp. flagellates 1 \nand their bacterial endosymbionts  2 \nSally  D. W arring 1 ,  Jamie McGowan 1 , Estelle S. Kilias 2 , James Lipscombe 1 , Elisabet Alacid 2 , 3 \nTom Bar ker 1 , Leah Catchpole 1 , Karim G harbi 1 , Seanna McTaggar t 1 , Thomas A. Richards 2 , 4 \nDavid Swarbr eck 1 , Neil Hall 1,3  5 \n 6 \n1  Ear lham Ins t itute, Norwich Res ear c h  Par k, Nor wich, U nit ed Kingdom, 2  De par tment  of  7 \nBiolo gy, U niversity of Oxford, Ox f ord, United King dom, 3 Sc ho o l of  Biol ogic al  Scienc es , 8 \nUni ver s ity of East Anglia, Norwich, U n ited King d om 9 \n 10 \nAbstract 11 \nBodo is a cosmopolitan genus of free living bacterivorous single-celled flagellates in 12 \nthe class Kinetoplastea. Members of genus B odo  are considered the closest free-living 13 \nrelatives to the parasitic lineages Tr ypan os oma  and Leis h mania , the causative agents of the 14 \nhuman diseases sleeping si ckness, Ch agas disease, and leishmaniasis. Currently, a single 15 \ngenome exists for the one formally d escribed species in the genus, Bodo s altans . Previous 16 \nstudies on B. s alt ans  have show n t hat it is dependent on an endosymbiot ic bacterium from 17 \nthe order Holosporales, “Candidat us  Bodocaedibacter vickermanii”. U sing single cell-18 \nsequencin g, we i solat ed, sequenced, and assembled genomes f or seven uncult ured Bodo 19 \nspp. cells from a single freshwater sample from Royal Leamington Spa, UK. By usin g 20 \ncomparat ive genomics, we show that  t hese seven cell s represent  t hree potentially novel 21 \nBodo species and exhibit unexpected levels of diversity at the genome level. Our results 22 \nindicate that Small Subunit (SSU) rDNA sequencing, of ten used to classif y Bo do  fla gella tes, is 23 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 2  \ninsufficient for determining species delimitation in this genus. In addition, w e recovered a 24 \nHolosporales bacterium genome from all seven Bodo  spp. cells. S urprisingly, these seven 25 \nbacterial endosymbionts also represent three p otentially novel species and one novel genus 26 \nof Ho losporales bacteria. This diversity would be indistinguishable in routin ely-used SSU 27 \nribosomal DN A ( rDNA)  metabarcodin g or bulk sequencing pipelines, thus demonstrating the 28 \nutility of  using single- cell sequencin g to reveal the level o f  genomic diversity within lineages 29 \nof microbial eukaryotes and their cobionts.  30 \n 31 \nKeywords 32 \nSingle-cell sequencing, Kinetoplastea,  Bodo saltans , protist, Holosporales, environmen tal 33 \nsequencin g 34 \n 35 \nBackground 36 \nBodo is a genus of heterotrophic free-living bi-flagellated pro tists common in fresh 37 \nand bracki sh water and soil. They are members of the Kinetoplastea, a class of parasitic and 38 \nfree-living pro tists distingui shed by the presence of a large mass of mitochondrial DNA 39 \nknown as kinetoplast D N A, or kDNA (1). Class Kinetoplastea is divided into  tw o subclasses, 40 \nthe Prokinetoplastina and the Metakinetoplastina, the latter containing fo ur orders, 41 \nEubodonida, Parabodonida , Neobodonida and the Trypanosmat ida (2,3). Order 42 \nTrypanosomatida includes the parasites responsible for human diseases such as sleeping 43 \nsickness, Chagas’ disease, and leishmaniases. Phylogenies based on Small Subunit (SSU) 44 \nribosomal DN A ( rDNA)  and protein sequences place order Eubodonida, and its one genus, 45 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 3  \nBodo, as the sister clade to the Trypanosmatida, and Bodo is considered the closest free-46 \nliving lineage to the parasitic Trypanosomatida lineages (4–6).  47 \nTraditionally, Bodo species were identified  and distinguished by morphology, e.g. cell 48 \nsize and shape, length of the f lagella, position of the nucleus and kinetoplast, and 49 \nultrastructural featu res visible by light and electron m icroscopy (7,8). Later, phylogenies 50 \nbased on molecular data showed the genus Bodo was paraphylet ic and its members were 51 \ndistributed in to three of the aforementioned four orders, with only one clade containing 52 \nthree species, B . saltans , B. edax, and B. unci natus r emaining genus B odo (2,4,9). Further, it 53 \nis suggested that B. edax  and B. uncin atus are not tru e species and are likely isolates of B. 54 \nsaltans ( 2,10). Despite the important  phylogenic position of  genus Bodo, B . saltans  remains 55 \nthe only formally described species in the genus, and a single genome from  bulk culture 56 \ncurrently exists for B. saltans strain Lak e Konstanz (11).  57 \nHolosporales are alphaproteobacterial t hat are widespread obligate endosymbionts 58 \nof eukaryotes, particularly protists (12,13). Holosporales form complex associations with 59 \ntheir eukaryotic hosts, including infectious parasitic species (14) , some that confer 60 \ncompetitive advan tages to their hosts ( 15), and some that may increase host f itness under 61 \ncertain conditions (16,17) . Recent  work has shown that  B. s alta ns  Lake Konst anz harbours 62 \nan endosymbiotic bacterium from the order Holosporales, “ Candida t us  (C a.)  63 \nBodocaedibacter vickermanii” (18). B. saltans appears dependent on its endosymbiont as 64 \nantibiotic treatmen t results in rapid cell death, and it is hypothesized that this dependence 65 \nis due to three pu tative addictive toxin/antitoxin systems encoded in the endosymbionts 66 \ngenome (18) . These systems render the endosymbiont  essential to its host  as t hey encode a 67 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 4  \nlong-lived toxin mo lecule alongside an antitoxin with a shorter half- l if e, and loss of the 68 \nantitoxin results in the activation o f the toxin and host cell dea th (19).   69 \nHere, we use single- cell sequencing t o assemble genomes from seven uncultured 70 \ncells from one environmental sample that were iden tified as belonging to genus Bodo. 71 \nGenomic comparisons show that the seven genomes form t hree clades, pot entially 72 \nrepresent ing three novel Bo do species that diverge significantly from B. s a l t ans  and each 73 \nother. In addition, we recover a Holosporales bacterial genome f rom each raw assembly and 74 \nshow that these putative endosymbionts also represent three poten tially novel species, 75 \nforming three clades that appear congruent with the phylogeny of  the hosts. These results 76 \nhighlight the utility of using single-cel l sequencing and comparative genomics to reveal the 77 \ndiversity within and between uncultured populations of microbial eukaryotes and their 78 \ncobionts.  79 \n 80 \nResults 81 \nSingle-cell sequencin g and genome assembly 82 \nSeven Bodo  spp. cells w ere identified  on tw o 96-well plates containing single cells 83 \nisolated by Fluorescence-Activated Cell Sorting (FACS) from an environmental sample. After 84 \nDNA amplification and short-read sequencing, between 20, 738, 413 – 48,141,972 paired-end 85 \nreads were generated for each cell (Table 1). The cells were named after the well in which 86 \nthey were deposited and are hereafter referred to as A8, A10, B2, B7, F10,  G10 and H10. 87 \nEach raw assembly w as curated by bi nning and taxonomic assignment  to remove scaffolds 88 \nof bacterial and non-target origin, and the curated eukaryotic assemblies r ange in size f rom 89 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 5  \n29,029,105 base pairs ( bp) to 36,294, 242 bp compared to t he bulk- cult ure assembly for B. 90 \nsaltans lake  Konstanz at 39,862,120 bp (Table 1). All assemblies are fragmented, each in 91 \nmore than 2,619 contigs with contig N50s ranging from 22,220 bp to 31, 876 bp (Table 1). All 92 \nassemblies range from 54.84 % t o 56. 02 % G C, a hi gher % GC than B. saltans  lake Konstanz 93 \nat 51.79 % (Table 1). Protein annotat ions of  each genome result ed in 11,121-13,166 protein 94 \nannotations per single cell genome, 30-40 % less than B. saltans lake Konstanz at 18,190 95 \nproteins ( Table 1) . Completen ess anal ysis of the protein sets using BUSCO v5 with the 96 \nEuglenozoa odb10 database gave 72. 3 %-80.8 % complete BUSCOs per assembly, compared 97 \nto  B. saltans lake Konst anz at 88.5 % complete BUSCOs (Table 1).  98 \n 99 \nCe l l  \nIdentifi e r \nraw \npa i r ed -\nen d  \nrea ds  \nA s sem b ly  s ta t i s t ics No. \np r otei ns \na nno ta te d  \nB US C O -   Eu g l eno zoa  o db 1 0 \n \nLen gt h No. \ncon ti gs  N50  GC %  Co m pl et e  \n% S D  F  M \nA8  20,814,8\n14 30,867,202 2,619 26,848 54.94 11,505 72.3 72 0 5 22 \nA10 25,608,2\n25 31,501,931 2,828 25,310 54.9 11,858 76.2 75 1 3 21 \nB2  20,738,4\n13 36,294,242 3,436 26,706 55.02 13,166 80.8 80 1 6 13 \nB7  48,141,9\n72 30,770,982 3,056 22,521 56.02 11,736 77.0 76 1 4 19 \nF1 0  21,853,2\n21 29,029,105 2,665 24,738 56.02 11,121 75.4 75 0 6 18 \nG10 33,945,9\n03 33,462,613 3,452 22,220 55 12,168  74.6 73 2 5 20 \nH10 21,728,8\n38 34,908,636 3,605 22,252 54.97 12,867  80.0 79 2 5 15 \n     \nB. saltans  39,862,120 2,256 31,876 51.79 18,190 88.5 112 3 12 3 \n 100 \nTable 1 - Bodo  spp. sequencing and assembly statistics. Number of  pair ed-end r eads 101 \ngenerated for each single cell assembly, and general assembl y and annotation statistics. The 102 \nassembly statistics for B. s alt ans  lake Konstanz genome are show n as a comparison. BU SCO 103 \ncompleteness f or protein sets, S = complete single copy, D = complete dup licated, F = 104 \nfragmented, M = missing. Euglenozoa odb10 includes 130 BU SCOs.  105 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 6  \nThe seven Bodo cells form three clades  106 \nBLASTn searches against  the GeneBank Nucleotide (nt ) database using the SSU rDNA 107 \nsequences from each assembly returned three B odo  spp. as top hits (Table 2). For A8 a nd 108 \nA10, the top hit is to ATCC isolate Bo do edax  (ATCC30903), originally isolat ed from the Czech 109 \nRepublic (5). For B7 and F10, the best hit is to B. s alt ans  strain HFCC309 i solated from a 110 \neutrophic pond in G ermany (20). For B2, G10 and H10, the top hit is to a B.  saltans  isolate 111 \nfrom soil in Malaysia ( 10).  In all cases, the query sequences are more than 99% identical to 112 \ntheir top hits. Figure 1A shows the pairwise identities of  the SSU  rDN A nucleotide sequences 113 \nfrom th e seven single cell assemblies.  A8 and A10 are 100% identical. B7 and F10 are 99.82%  114 \nidentical, with one m ismatch between the two sequences, a one nucleotide indel o f a G  115 \nresidue in a string of G residues, however, due to the fragmented nature o f the assemblies 116 \nthe SSU rDNA sequence from F10 is truncated, limiting f ull comparison of these sequences 117 \n(Table 2 and Additional File 1). B2 and H10 are 100% identical, with G 10 99.92 % identical to 118 \nB2 and H10 w ith a one nucleotide polymorphism, a C to T transition in G 10. All th ree SSU  119 \nrDN A sequences from B2, H10 and G10 are truncated, missing ~700 nucleotides f rom their 120 \n5’ ends (Table 2 and Additional File 1). A8/A10 differ fro m B2/G10/H10 by one nucleotide, 121 \nan A/G  transition, while B7/F10 differ more substantially from all the others, with pairwise 122 \nidentities ~98% (Figure 1A) .  123 \nA Maximum-Likelihood ( ML) tree of SSU rDN A sequences of Bodo and related 124 \nspecies available in G eneBank places our seven cells onto two clades, all within the B.  125 \nsaltans lineage (Figure 1B).  A8/A10 are sister to B2/G 10/H10 and are part of a radiation 126 \nthat includes B. edax  , B. unic in at us  and three o ther B. s alt a ns  isolates. B7/F10 are si ster to 127 \nthis is clade, along w ith two other B. s alt ans  isolates. B . s altan s  lake Konst anz branches with 128 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 7  \nother  B. s alt a ns  isolates on a separate clade that diverges closer to the ro ot of  th e tree 129 \n(Figure 1B). The seven single cell Bod o genomes have SSU rDNA pairwise identities ranging 130 \nbetween 94.7 and 95.7 % with B. saltans  lake Kon stnaz and appear closer to t he SSU rDNA 131 \nsequence f rom B. edax, with pairwise identities ranging between 98.0 and 99.6 % identity 132 \n(Figure 1B) . 133 \n 134 \nAss e m\nbl y  \nSS U  \nleng th  \nTo p hit Bl as t n \nsu bj ec t  ac ce ss i o n \nT op hit B l as t n su bj ect \nor g anis m  \nT op hit B l as t n quer y \ncov er age  \nT op hit B l a s t n \npai rwis e id enti t y \nT op hit B l as t n \nbi t s c or e  \nA8 2,147 AY028451 Bodo edax 99.81% 99.60% 3829.84 \nA10 2,147 AY028451 Bodo edax 99.81% 99.60% 3829.84 \nB7 2,147 DQ207572 Bodo saltans 98.04% 99.30% 3724.34 \nF10 1,124 DQ207572 Bodo saltans 97.86% 99.40% 1951.63 \nB2 1,393 AY490226 Bodo saltans 99.86% 99.70% 2492.64 \nG10 1,263 AY490226 Bodo saltans 100% 99.80% 2266.32 \nH10 1,409 AY490226 Bodo saltans 99.86% 99.80% 2521.49 \n 135 \nTable 2 – BLAST results for SSU rDNA sequences. Top hits resulting from a BLASTn search of  136 \nthe nt database f or the SSU rDNA sequence recovered from each single cell assembly.  137 \n 138 \nExtensive genomic diversity among t he Bodo spp.  139 \nWe next  built  a Maximum Likelihood (ML) tree from  488 single-copy orthogroups 140 \nshared by B. s a l t ans  lake Konstanz and our seven single cell Bodo asse mblies, using 141 \nPer kinsela  sp. as an outgroup ( Figure 2A). Here, the seven single-cell genomes split into 142 \nthree clades with A8/A10 form ing a si ster clade to B2/G 10/H10 and B7/F10 forming another 143 \nclade branching closer to the root o f the tree.  B. sal t a ns lake Konstanz forms its own clade 144 \nclosest to the root of the tree. N ext, we cal culated average pairwise amin o acid identities 145 \n(AAI) for all eight Bodo assemblies, wi th each pairwise comparison averagi ng AAI over a 146 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 8  \nminimum of 7,419 and a maximum of 10,181 reciprocal best hits (Addition al File 2).   The AAI 147 \nwithin each clade is high (>97%) but varies substantially betw een the clades (Figure 2B). B.  148 \nsaltans lake Konstanz i s equally di stant from all seven single cell assemblies, with pairwise 149 \nAAI between 51.3-51.7 % in all comparisons. B7/F10 are equally distant from A8/A10 and 150 \nB7/G 10/H10 with pairwise AAI between 62.1-62.6%. A8/A10 and B7/G10/H10 appear 151 \ncloser, w ith pairwise AAI between 84.8-85.0%.  152 \nWe also compared t he Average N ucleotide Identity (ANI) and the Aligned Fraction 153 \n(AF)  of each Bodo spp. genome using the too l skani (21) (Figure 2C). Here, a result of 0.00 154 \nfor AN I and/or AF suggests that sequence similarity at the nucleotide level is too low for 155 \npairwise comparison using this method (21), e.g. B. s a l t ans  lake Konstanz is too distant to 156 \nany of the single-cell Bod o genomes to compare using this metric, with values of 0.00 in 157 \nevery pairwise comparison. However,  ANI and AF are relat i vely high within each of the three 158 \nsingle-c ell Bodo clades ( >98 %  AN I and > 80 % AF). Between the three clad es B7/F10 is too 159 \ndistant from the ot hers to compare, while A8/A10 and B2/G10/H10 show an ANI of ~84%, 160 \nbut an AF of  ~14-16% indicating that only a small portion of the genome is close enough for 161 \ncomparison using AN I . The AF values shown in Figure 2C are the average of two AF values 162 \ncalculated f or each pairwise compari son. A matrix with bo th AF values calculated for each 163 \npairwise comparison is in Additional File 3. 164 \nNext , w e compared how many orthogroups are shared among the single-cell 165 \nassemblies and B. s alt ans  lake Konstanz ( Figure 2D).  The majority of ortho groups, 55.5% are 166 \nshared by all eight assemblies. The two clades that branch most closel y in the phylogenetic 167 \ntrees, A8/A10 and B2/G10/H10 share the highest  percentage of orthogroups, 83.4%. B7/F10 168 \nshare 66.1% with A8/A10 and 68% wi th B2/G 10/H10. B. saltans  lake Konst anz shares 62.8 %  169 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 9  \nwith B7/F10. 61. 4 % wit h A8/A10 and 63.7% wit h B7/G10/H10. Each clade has a small 170 \npercentage of  orthogroups that are unique, w ith B. s alt ans  lake Konstanz having the most 171 \nunique orthogroups.  172 \nTo compare t he functional complim ent of  each genome, we mapped each protein 173 \nset to the Pfam database. In each case, less than 50%  of  proteins in each annotation contain 174 \none or more Pfam domains (Figure 2E) . W e tabulat ed the Pfams found in each prot ein 175 \nannot at ion and ran a Principal Components Analysi s (PCA)  on that table. The PCA separates 176 \nthe genomes into t hree clust ers congruent with t he phylogeny. Principle Component  ( PC) 1 177 \nseparates B. s alt ans  lake Konstanz f rom all others. While PC 2 separates B2 and F10 f rom 178 \nA8/A10 and B7/G 10/H10, w hi ch clust er together (Figure 2F, Additional File 4).  179 \nTaken together, the phylogeny, AAI, AN I, AF and Pf am clustering indicate that the seven 180 \nsingle cell Bodo assemblies form t hree clades: clade A8/A10, clade B7/F10 and clade 181 \nB2/G 10/H10. All three clades are potentially novel Bodo species, and all appear equally 182 \ndistant from B. salta ns lake Konstanz.  183 \n 184 \nAll Bodo  spp. harbour Holosporales bacterial endosymbionts 185 \nPrevious st udies show that  B . saltans  lake Konst anz is dependent on an 186 \nendosymbiotic alphaprot eobacterium f rom t he order Holosporales, “ Ca. B. vickermanii” 187 \n(18). We went back t o our binned raw genome assemblies and were able t o identify a single 188 \nbin in five of the genomes, 2 bins in B2 and 4 bins in B7 that were > 80 % complete 189 \naccording to CheckM (Additional File 5). CheckM analysi s of  these genomes showed that all 190 \nwere at least 89 % complete, and all but one had contamination lev els < 5% (Table 3).  191 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 10  \nAss e m b l y  C h e c k M  An not ati o n \nCell  Le n g t h ( b p)  No . \nc ont ig s N50 GC %  \nCom p-\nle t en e ss   \n% \nC ont -\nam ina t ion  \nstr ai n \nhe te r o-\nge neit y  \nno. \nge nes  \nno. \ntR NA s rDNA s  \nA8 1,591,871 98 28,883 37.82 90.32 0 0 1,591 39 (20) 23s, \n16s \nA1 0 1,758,263 139 26,374 37.27 90.32 0 0 1,591 41 (19) 0 \nB2  2,174,152 145 31,250 34.85 93.55 26.34 10.34 2,011 55 (19) 0 \nG1 0 1,604,499 119 27,931 38.14 89.01 2.2 100 1,363 39 (20) 23s, \n16s \nH1 0 1,693,415 108 27,858 37.65 94.62 0 0 1,463 39 (17) 0 \nF10 1,279,012 83 33,152 41.61 94.62 0 0 1,234 35 (18) 0 \nB7  1,292,737 53 55,440 41.89 95.6 0 0 1,220 38 (19) 0 \n           \n“Ca. B.  \nvi cker ma nii”  1,391,311 1 NA 40.61 94.62 0 0 1,214   \n 192 \nTable 3 – Assembly statistics, CheckM  results for Holosporales genomes. G eneral assembly 193 \nand annot at ion st atist ics for each Holosporales genomes. No. tRNAs shows the tot al number 194 \nof tRN As, and the number of dif feren t amino acids in brackets. Values from “ Ca. B. 195 \nvickermanii” are shown where appropriat e for comparison. 196 \n 197 \nGTDB-TK classified one bin from B7 and the bin in F10 in th e order Ho losporales, 198 \nfamily “ Ca. Paracaedibacteraceae”, genus U BA6184, as the same genus as “ Ca. B. 199 \nvickermanii” (Additional File 5). Six other bin, including a single bin from B2  were classif ied 200 \nin order Paracaedi bacterales , family U BA11393, genus J AGOTX01, whi ch currently contains 201 \na single bacterial met agenom e assembled genome (MAG) isolat ed from wastewater ( 22) 202 \n(Additional Fil e 5) . The additional bin s f rom B2 and B7 were not classified as 203 \nalphaproteobacterial and were no t analysed further (Additional File 5). Order 204 \nPara c aedibactera l es is considered a heterotypic synonym of  order Holosporales ( 12). 205 \nRecently it has been proposed that due to phylogenetic nesting of order H olosporales, it be 206 \ndown-ranked to suborder Ho losporin eae, which w ould include the families H ol o s por ac eae , 207 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 11  \n“ Caedimonidace a ” and “Ca. Paracaedibacteracea” ( 13,23). U nder this ranki ng, all our 208 \ngenomes would fall w ithin suborder H olos por ineae , two falling in fam ily “Ca. 209 \nParacaedibacteracea” and the rest as unclassified H olos por ineae . 210 \nTo better place the bod o Ho losporlaes, we built a ML tree fro m 24 single-copy 211 \nort hogroups shared among Holosporales genomes and MAG s available in GeneBank and 212 \nusing two alphaprot eobact erial as an out group (Figure 3A). This t ree places the seven 213 \ngenomes into two lineages. F10/B7 branch on a clade wit h “ Ca. B. vickermanii” and two 214 \nMAG s classified as family “ Ca. Paracaedibacteracea” assembled from metagenomes of 215 \nwast ewat er samples (24, 25). A8/A10 and B2/H10/G 10 form another clade with t hree MAG s 216 \nassembled from m etagenomes of  water samples, including wastewater (22,26, 27). “ Ca. 217 \nFinniella inopinata”, an endosymbiont of the amoebaflagellat e V iri dirap t or  invadens  218 \nbranches at the base of this clade  (23,28).  W e calculated pairwise AAI for all the genomes in 219 \nthe tree (Figure 3A, Add itional File 6). The pairs B7 and F10, A8 and A10, and G10 and H10 220 \nall have a pairwise A AI greater than or equal to 95 %, considered a cutoff for species 221 \ndelimitation in the literature (29,30) , indicating that they are most likely the same species. 222 \nB2 is more distant from  G10 and H10 with AAI values of 58.8 and 59.62, respectively 223 \n(Additional file 6). However,  CheckM result s show that  t his genome cont ains contaminat ion, 224 \nwhich could be interfering with these analyses. The AAI boundary for genus delimitatio n 225 \nvaries depending on the taxa under consideration (31–34). Here, we used a cutof f of 55%, as 226 \nwas used in Midha et al. (2021)  w hen describing “ Ca. B. vi ckermanii”. This cutoff place clade 227 \nB7/F10 as a novel species in the same genus as “ Ca. B. vi ckermanii”, while clade A8/A10 and 228 \nG10/H10 form t wo novel species in a novel genus.  229 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 12  \nWe also compared AN I and AF values f or all pairwise comparisons between the Bodo spp. 230 \nendosymbionts and “ Ca. B. vickermanii” (Figure 3B). AN I and AF  are hig h w ithin clades (ANI 231 \n>97%, AF >85%), but, bet ween the clades, only A8/A10 and H10/G10 are close enough to 232 \ncompare, with AN I values of 85.42-86.31 % and AF at 27.73-29. 52 %, indicating that like 233 \ntheir hosts, these genomes share high sequence similarity only over a small proportion o f 234 \ntheir genomes. The B2 genome shows high ANI ( > 99% )  w it h the genomes from H10 and 235 \nG10, but low AF (<25%) indicating that this genome is only similar to those genomes over a 236 \nsmall proportion of the genome. How ever, contamination in this genome may mean that 237 \nthis result is an artifact. The AF values shown in Figure 3B are the average of two AF values 238 \ncalculated f or each pairwise compari son. A matrix with bo th AF values calculated for each 239 \npairwise comparison is in Additional File 7. 240 \nFinally, we compared t he number of shared ort hogroups between the Holosporales 241 \ngenomes and “ Ca . B. vickermanii” (Figure 3C). 23. 2% of  orthogroups are shared by all 242 \ngenomes. The genomes t hat are closest in t he phylogeny, A8/10 and B2/G10/H10, share t he 243 \nhighest proportion of  orthogroups at 52.6 %. G enomes B2/G10/H10 have the highest 244 \nproportion of un ique orthogroups at 12.35, ho w ever, the contamination in  the B2 genome 245 \nis likely to be inflating this number.  246 \nIn summary, B. saltans  and each novel Bodo clade presented in this paper harbour a 247 \nunique species of Holosporales endosymbiont . These seven endosymbionts split into t wo 248 \ngenera (genus “ Ca. Bodocaedibacter” and a novel genus) t hat appear congruent with the 249 \nphylogeny of the host and further support the conclusion that these are three novel Bodo 250 \nspecies.   251 \n 252 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 13  \nThe Bodo  Holosporales differ in their metabolic capacity. 253 \nWe compared the secretion systems and met abolic modules present  in t he bact erial 254 \ngenomes and compared them to “ Ca. B. vickermanii” (Figure 4) .  Genus “ Ca . 255 \nBodocaedibacter” (B7/F10) and the novel genus described in this study (A8/A10 and 256 \nB2/G 10/H10)  encode a SEC-SRP and a Type VI secretion system (18). In addition, the n ovel 257 \ngenus encodes two out the fou r proteins of the TAT secretion system, while a partial Type IV 258 \nconjugal transfer pilus assembly protein systems i s spottily distributed across the taxa.  259 \nAll seven genomes have a partially complete (>= 66% complete) pat hway for lysine 260 \nbiosynthesis, which is absent in all seven hosts ( <=22% complete) (Figure 4 ). In addition, the 261 \nnovel genus encodes several met abolic pat hways absent in genus “ Ca. Bodocaedibacter”, 262 \nnotably including heme biosynthesi s ( Figure 4). Heme is an essential nutrient, and to date, 263 \nno complete heme pat hway has been described for a Kinet oplast ea member, which either 264 \nrequire an exogenous source of heme, or contain endosymbiotic bact eria that  produce it 265 \n(35).  Figure 4 only shows KEGG  modules that  were at least  50%  complet e in one or m ore 266 \ngenomes. A full list of all the KEGG modules, including those at less than 50% completeness 267 \nis presented in Additional File  8. 268 \nMidha et al. (2021) reported that “ Ca. B. vickermanii” encodes three putative 269 \npolymorphic t oxin/anitoxin system s which may be responsible for its host ’s dependency 270 \n(18). Polymorphic toxin/antitoxin systems are typically composed of a large multi-domain 271 \nprotein contain ing an N - terminal secretion signal f ollowed by a toxin domain, a protective 272 \nimmunit y protein, and multiple casset tes encoding alt ernative toxic domains and associat ed 273 \nimmunit y proteins (19). The t oxin genes can cont ain a homologous repeat region t hat 274 \nenables recombination betw een the full-length toxin gene and the alternative toxins. We 275 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 14  \nidentified th ree genes belonging to one putative toxin/antitoxin system in both B7 and F10, 276 \nbut none in the other five genomes from the novel genus. These genes show  some of t he 277 \ncharacteristics of  a polymorphic toxin /antitoxin system including a homolo gous region. 278 \nHow ever, the pu tative large mul tido main protein in B7 d oes not contain an N-terminal 279 \nsignal pept ide, and the one put ative orphan module in bot h B7 and F10 contains a t oxin but 280 \nno cognate anti-toxin (Additional File 9). Therefore, the l ocus’ identity and f unction as a 281 \npolymorphic toxin/antitoxin system is tentative. In addition, the incompleteness of the 282 \nsingle cell genomes presented here raises the possibility that key proteins and/or pathways 283 \nare missing from these assemblies due to under sampling. 284 \n 285 \nGenus  Bodo  encodes many unique protein families, but core metabolism is generally 286 \nconserved with parasitic Kinetoplastea 287 \nWe next  compared ort hogroups distribut ed across the Kinet oplast ea lineage (Figure 288 \n5). For each kinetoplastid genome we tallied the total numb er of orthogrou ps in that 289 \ngenome, and the number of  orthogroups unique to that  genome (Figure 5). Species on long 290 \nbranches or located at the base of lin eages have the highest number of unique orthogroups, 291 \nwith  B. saltans lake Konstanz  having t he most unique orthogroups. We next calculated t he 292 \ntot al number of  orthogroups and the number of unique orthogroups in t hree kinetoplastid 293 \ngenera, Bodo, Leis h mania , and Trypanosoma (Figure 5). Consistent with previous studies 294 \nshowing gene number reduct ion in parasit ic lineages compared to  B. salta ns (11), genus 295 \nBodo has the greatest number o f orth ogroups, and the highest proportion of unique 296 \nort hogroups, 53%, compared t o 35% and 5.5 % in Tr y panosoma  and L e is hma n i a,  297 \nrespectively.  298 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 15  \nFinally, we compared t he completeness of  metabolic pathways present  in B odo  spp. 299 \nand kinetoplastid species available in the Kyot o Encyclopedia of Genes and Genomes (KEGG) 300 \ndatabase (Figure 6). The four Bodo spp. are nearly identical in their metabo lic capacity, with 301 \nalmost no KEGG  modules mi ssing from one lineage and present in another (Figure 6). 302 \nExceptions are Arginine biosynthesis and Ribof lavin biosynthesi s, whi ch are missing in B. 303 \nsaltans lake Konstanz and partially present (25-80% complete) in  the o ther species. The four 304 \nBodo spp. also share most of  their metabolic capacity w ith the parasitic lineages, as 305 \npreviously reported (11). However, we do observe some pat hways that  are complete or 306 \nnearly complete (>=80% complet e) in the Bod o lineage and absent or nearly absent (<=20% 307 \ncomplete)  in Tr y panosoma  and Leis hm ania. N amely, Trypt ophan metabolism, Tyrosine 308 \ndegradat ion, Molybdenum cofactor biosynt hesis, and Pyrimidine degradat ion (uracil => 309 \nbeta-alanin e, thymine => 3-am inoisobutanoate) .  Conversely, the G lyoxylate cycle is partial  310 \n(60 % complete) in the Trypan osoma and L eishmania, and near absent in all Bodo spp., 311 \nconsistent w ith previous findings in B. saltans (36), while Nucleotide sugar biosynthesis 312 \n(galactose => UDP- galactose) is partially complete (50 %) in all Leishmania and T. c r uz i, but 313 \nabsent in all Bo d o  spp. Figure 6 only shows modules that are at least 60% complete in at 314 \nleast one or more genomes. The f ull list of all KEGG  modules identified in each genome, 315 \nincluding those less than 50% i s in Additional File 10.  316 \n 317 \nDiscu ssion  318 \nIn this study, we show that single-cell sequencing reveals unexpected levels of 319 \ndiversity among uncultured candidat e Bodo spp. cells f rom a single enviro nmental sample. 320 \nMet abarcoding, part icularly of SSU rDNA, and met agenomics have become the norm for 321 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 16  \nassessing the diversity of microbial eukaryotes in environmen tal samples. For protists, 322 \ntraditionally these methods involve clustering reads and/or amplicons into Operational 323 \nTaxonomic units (OTUs) based on similarity thresholds, usually at around 97-99% identity, or 324 \ninto Amplicon Sequence Variants (ASVs) produced b y t he DADA2 pipeline (37). The seven 325 \nsingle cells presented here all have a pairw ise SSU  rD N A identity greater th en 99% and 326 \nwould therefore have b een clustered into one O TU in some metabarcoding pipelines. 327 \nInstead, by generating single-cell genomes, we show that these seven cells form th ree 328 \ndistinct clades, with each clade potentially representing a novel species. ANI has been used 329 \nto compare distances within and between closel y-related species of eukaryotic microbes 330 \nincluding microsporidia, yeast and some protists, with an interspecies ANI cutoff of  95%  331 \nconsidered appropriate (38–40) . Here, we found that genomes within each of the three 332 \nputative B odo  species have an ANI > 98%, supporting the conclusion that each clade 333 \nrepresent s a Bodo species. However, ANI is not useful when comparing more dissimilar 334 \ngenomes, where AAI becomes a more useful met ric (41,42).  While more commonly used to 335 \ncompare prokaryotic genomes, AAI has also been applied to eukaryotes. A comparison of 336 \n1,196 Human and Mouse prot ein sequences show ed an AAI of 85% (43), while recent 337 \nstudies in f ungal linages found AAI values for members of the same species were often 338 \n>97% and demonstrated an est imated a genus boundary threshold of ~70-75% AAI f or the 339 \nfamily Hypoxylaceae (Ascomycota) ( 44,45). W hile further w ork is needed to explore 340 \nappropriate species and genus cutoffs for AAI and ANI within and between  protist lineages, 341 \nour results showing that AAI with in each Bodo spe cies w as > 97% suggest that the species 342 \nthresholds of 95- 98% used in bacteria and fungi may be appropriate here too.  343 \nAll three novel Bo do species presented here are more similar by SSU rDN A identity 344 \nto B. edax ATCC strain 30903 than they are t o B. s alt ans  lake Konstanz. However, as no 345 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 17  \nother molecular sequence dat a exists from t his organism, we are unable t o determine if  any 346 \nor all are indeed B. ed ax . Callahan et al. (2002) mentions that B. edax contains 347 \nendosymbiotic bacteria and differs morphologically f rom B. saltans by lacking 348 \nmastigonemes on the anterior flagellum (5) , yet  several authors have proposed that B. edax 349 \nbe reclassified as an isolate of B. salta ns  (2,10). Given the extensive genetic diversity 350 \nobserved among the B odo spp. present ed in this paper, more work is needed to determine 351 \nif B. edax is a true species.   Importantly , this study sugg ests that the SS U rDN A locus is 352 \ninsufficient for species delimitation within genus Bodo as it does not reflect the genomic 353 \ndiversity within the group.  354 \nFurt her support ing our conclusion t hat the three Bo do clades represent three 355 \nspecies i s the f inding that each putative species harbors a unique species of Holosporlaes 356 \nbacterium. The three novel bacterial species split into two genera, one species belonging to 357 \ngenus “ Ca. Bodovickermani,  and the other two species forming a novel genus. The 358 \nphylogeny of the host Bod o spp. appears congruent  wit h the phylogeny of the 359 \nendosymbionts, however more sampling covering greater taxonomic dist ribution is needed 360 \nto test if  this congruence indicates co -evolution, or is the result of  oth er processes ( 46,47). 361 \nInfectivity has been demonstrated in the Ho losporaceae family of Holosporales, 362 \nendosymbionts of various Param ecium  ciliates ( 16,48), and some Holosporales have 363 \ndemonstrated the ab ility to invade n ovel hosts experimentally (49,50). However, the 364 \ndistribut ion and genetic differences observed between the Holosporales genomes 365 \npresented here suggest that these endosymbionts are and have been associated with their 366 \nrespective host s for a long period.  367 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 18  \nThe nature of the symbioses between H olos por ales  bacteria and their hosts is 368 \ncomplex and varied (13). Midha et  al. ( 2021) hypothesi zed that B. saltans  is dependent  on 369 \nits endosymbiont due to the presence of three putative addiction p olymorphic toxin/anti-370 \ntoxin systems encoded in the “ Ca. B. vickermanii” genome ( 18). These systems typically 371 \nencode a long-lived toxin molecule alongside an antitoxin with a shorter h alf-life, so that if 372 \nthe system is lost, the toxin is activated and becomes lethal. We found evidence for a 373 \nputative toxin/anti- toxin system in the two genomes belonging to genus “ Ca . 374 \nBodovickermani”, but not in five gen omes in the novel genus. However, th e identity and/or 375 \nfunctionality of  th e toxin/anti-tox in systems described in this study is speculative, as the 376 \nproteins and loci differ from those in “ Ca. B. vickermanii” by encoding fewer orphan 377 \nmodules. The lack of a similar system in the genomes of the n ovel Holosporales genus 378 \npresented in t his study suggests that  alt ernative processes may be maintaing this symbiosis. 379 \nAll kineto plasts including B. saltans lack the biosynthet ic pat hway required for heme 380 \nbiosynthesis (36,51) , as do the Bodo  spp. presented in this study. The Kinetoplastids 381 \nAngom o nas  and Str igomonas  harbour endosymbiotic beta-proteobacteria which provide 382 \ntheir hosts with nutrien ts including heme, essential ammino acids, and vitamins (35, 52,53).  383 \nWe did f ind partially complete pathw ays for heme biosynthesi s ( >=70% complet e) in the 384 \nnovel genus, and a partially complet e pathway for Lysine biosynthesi s (>=66% complete)  in 385 \nall the B odo  Holosporlaes genomes. Theref ore, the two H olosporales genera presented here 386 \ncould demonstrate two alternative symbiotic strategies; with genus “ Ca. Bodovickermani” 387 \nbecoming addictive to its host through the toxin/antitoxin system, while the novel genus is 388 \npossibly a source of heme and/or other essential nut rients. All seven Holosporales genomes 389 \npresented in this paper encode a Type VI secretion system, con served throughout 390 \nHolosporales and hypot hesized to play a role in host-endosymbiont  int eractions (13).  391 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 19  \n 392 \nPrevious st udies have shown that B. s altans  sha res the majority of its meta bolic 393 \npathways w ith the parasitic lineages Trypanosoma and leis hma ni a  ( 11,36).  Our results show 394 \nthat  the metabolic capacit y of  B. s alt ans  is shared with the novel Bodo spp. presented here.  395 \nWe also show t hat the Bod o  genus contains an abundance of unexplored and uncategorized 396 \nprotein d iversity. The lineage contains far more unique o rthogroups and proteins than the 397 \nparasitic lineages. However, as most of these protein sequences have no known homology 398 \nto functional do mains in databases such as UniProt and KEGG; the functional implications of 399 \nthis unique gene repertoire remains unknown.  400 \n 401 \n Conclusions  402 \nMost of  our knowledge of protist  genomic diversity and their symbionts comes from 403 \nstudies of species that are culturable. However, we know that environmen tal samples 404 \ncontain a wealth of underexplored d iversity. Thi s study uses single-cell sequencing and 405 \ncomparat ive genomics to show that seven B odo  spp. cells f rom a single environmen tal 406 \nsample represent three po tentially novel species, each harbouring a novel and unique 407 \nspecies of bacterial endosymbiont.  C omparing these data, we demonstrate the varied 408 \nnature of these symbioses and show that single-cell sequencing i s a power ful meth od for 409 \nexploring the diversity of uncultured protists and their cobionts. 410 \n  411 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 20  \nMethods 412 \nSample collection 413 \nSurf ace water (~1 m) was collected fr om t he River  Leam (52.287295, - 1.547563), 414 \nRoyal Leamington Spa (UK) in August  2022. Init ially, t he sample was prefiltered t o remove 415 \nlarger debris and then con cent rated using diff erent polycarbonat e filt er pore sizes 416 \n(Millipore) t o obtain concent rated subsamples of different protist size ranges (10 t o 40 µm  417 \nand 0.8 t o 5 µm). The subsamples were supplement ed with 2- 3 autoclaved barley grains to 418 \nsupport  heterotrophic/mixotrophic growth via bact eria increase and incubat ed for one 419 \nweek prior to cell sorting. Single-cell nuclei were stained with 1xSybr G reen for 10 minut es  420 \nand sorted int o 96- w ell microplates (pre-filled wit h 5ul autoclaved/sterile filtered media), 421 \nusing fluorescence- activated cell sorting (FACS; flow rate=1) and selecting against 422 \nchlorophyll a- negative cells while selectiong for SybrGreen-positive cells. After cell sorting, 423 \n10 µl of RLT-plus lysi s buffer (Qiagen) was added to t he wells and the plat e was frozen at -424 \n80°C until further processing. 425 \n 426 \nWhole genome amplification, library preparation, and sequencing 427 \nA modif ied G&T- seq protocol (54) was carried out as follows. U sing a magnet ic  428 \nseparator, Dynabeads MyOne Streptavidin C1 (Invitrogen) beads were washed according to 429 \nthe manufactur er’s guidance and then incubated with 2 × Binding &Wash buffer  (10 mM  430 \nTris-HCl pH 7.5, 1 mM  EDTA, 2 M NaCl) and Biotinylated Oligo- dT primer (IDT, 5’-431 \n/BiotinTEG / AAG  CAG  TGG  TAT CAA CG C AG A GTA CTT TTT TTT TTT TTT TTT TTT TTT TTT  T TT  432 \nTVN-3’) at 100 μM for 30 minutes at  room temperature on a rotator. The oligo-t reated 433 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 21  \nbeads were washed four t imes in 1 × Binding &  Wash buffer  (5 mM Tr is- HCl pH 7.5, 0.5 mM 434 \nEDTA, 1 M NaCl) and then suspended in 1 × SuperScript  II First Strand Buff er (Invit rogen) 435 \nsupplemented with SUPERaseIn R N ase Inhibit or (Invitrogen)  to a final concentration of 1 436 \nU/μl. The lysate w as thawed on ice. 10 μl of prepared oligo-dT beads was added to each well 437 \ncontaining 12 μl cell lysate using a Dragonf ly Discovery liquid dispenser ( SPT Labtech). The 438 \nlysat e plat e was sealed and incubated on a ThermoMixer C (Eppendorf) w it h a heated lid at 439 \n21°C for 20 minutes shaking at 1000 rpm. Using a Fluent 480 liquid handling robot ( Tecan) 440 \nand a Magnum FLX magnet ic separator  (Alpaqua) , the lysat e super natant was transferr ed to 441 \na new plate, and t he beads wer e was hed twice in a cust om wash buffer ( 50 mM Tr is-HCl pH 442 \n8.3, 75 mM KCl,  3 mM MgCl2, 10 mM DTT, 0. 5% Tween-20). The supernatant  f rom the 443 \nwashes was added t o the left-over cell lysate containing t he genomic DNA which w as stored 444 \nat -20°C overnight. The mRN A was not used for this study.  The r emaining cell lysat e was 445 \nthawed and subjected to a 0.6 × vols Ampure XP clean- up with 80% et hanol. The bead-446 \nbound gD N A was i sothermally amplified for 3 hours at 30°C t hen 10 minut es at 65° C using a 447 \nminiaturised (1/5 vol s) Repli-g Single-Cell assay ( Q iagen).  The amplified gDN A was cleaned 448 \nup with 0.8 × vols Ampure XP and 80% ethanol, then elu ted in 10 mM Tris-HCl.  449 \nSequencing libraries for t his project were const ructed by t he Techni cal Genomics 450 \nGroup at  the Earlham Inst itute, N orwich, U K. Initial libraries w ere const ruct ed for shallow  451 \ndepth sequencing as follows: gD NA was quantified by f luorescence ( Quant- iT HS-DNA, 452 \nInvitrogen) on an Infinite Pro 200 plate reader (Tecan) t hen normalised to a f ina l 453 \nconcentrat ion of  0.2 ng/μl in 10 mM Tris-HCl.  The Mosquito HV and D ragonfly Di scover y 454 \nliquid handling inst rument s (SPT Labtech) were used to prepare miniaturised (1/12.5 vols) 455 \nNext era XT ( Illumina) dual-indexed sequencing libraries as follows:  A tagmentation  456 \nmastermix comprising two part s TD buffer (Illumina)  and one part ATM (Illumina) was 457 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 22  \nprepared. Using the Dragonfly Discovery, 1.2 µl tagmentat ion mast ermix was dispensed per 458 \nwell to a 384-well skirted PCR pl ate (Eppendorf) . U sing the Mosqui to HV, 0.4 µl of 459 \nnormalised DN A was transferred to each well containing tagment ation mastermix. The plate 460 \nwas sealed and spun down then incubated at  55°C for  10 minutes on a ther mal cycler. Using 461 \nthe Dragonfly Discovery, 0.4 µl 0.2% SDS was dispensed to each reaction. The plat e was 462 \nsealed, spun down and then incubated at room  temperature for 5 m inut es. Using the 463 \nDragonfly Discovery 1. 2 µl NPM (Illumina) was added to each react ion. Using t he Mosquito 464 \nHV, 0.8 µl index primers pairs (i5 + i7)  at a concentration of 0.5 µM was added to each 465 \nreaction ensuring a unique combination in each well. The libraries were amplified under the 466 \nfollowing thermal cycler condit ions: 72° C for 3 minutes, 95°C for 30 seconds, 12 cycles (95°C 467 \nfor 10 seconds, 55°C for 30 seconds, 72° C for 60 seconds), 72°C for 5 minutes, 4°C hold. The 468 \nlibrar ies w er e pooled and cleaned up using 0.8 × vols Ampure XP and 80% et hanol. The 469 \nlibrar y pools wer e eluted in 20 µl  10mM Tr is-HCl and assessed using a Bioanalyzer HS D N A 470 \nassay (Agilent) , HS DN A Q ubit assay (Invitrogen) and finally an  Illumina Library  471 \nQuantification Kit  assay (KAPA).  These r eads f rom these libraries were used f or t axonomic  472 \nassessment of the cells. 473 \nAft er t axonomic assessment deep sequencing libraries f or t he Bodo cells were 474 \nconst ructed using t he KAPA High Throughout  Library Prep Kit (Roche Part N o: 475 \nKK8234/07961901001). W her e possi ble, 1µg of genomic DNA was sheared t o 450bp usin g 476 \nthe Covaris ML230 Sonicator ( Covaris)  and the ends of t he DNA were repaired; 3' t o 5' 477 \nexonuclease activity removed t he 3' overhangs and the polymerase activity filled in the 5'  478 \noverhangs creating blunt  ends. A single ‘A’ nucleot ide was added t o the 3’ ends of t he blunt 479 \nfragments to allow for the ligation of barcoded adapters ( 6bp - Perkin Elm er NEX TFLEX DNA  480 \nBarcodes 1-48 ( NOVA- 514101/2/3/4) ) or  ( 12bp - Perkin Elmer N EXTFLEX-HT (NOVA-481 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 23  \n51474/ 5/6/7)) at  a concent ration of 6µM prior  to a 0. 8x clean up using Beckman Coulter 482 \nAMPure XP beads (A63882). The size of  the libraries was estimated using an Agilent  High 483 \nSensit ivit y DNA chip (5067- 4626) and the concent rations w ere quant ified by fluorescence 484 \nwith a High Sensitivity Q ubit assay ( ThermoFisher Q32854).  485 \nBot h shallow and deep sequencing l ibraries wer e sequenced on either an Illumina 486 \nNovaSeq 6000 (cells A8, A10, B2, F10, G10, H10) with SP Reagent Kit v1.5 kit (300 cycles), , 487 \nusing  Real- tim e A nalysi s (RTA) ( version 3.4.4) and Control Software (version 1.7.5) or an  488 \nIllumina NovaSeq X Plus (cell B7) wit h  10B Reagent Kit  (300 Cycle) using RTA (version 4.6.7) 489 \nand Cont rol Software  (version 1.2.2)  to produce 150 bp paired-end reads. The result ing BCL  490 \nfiles were converted to  f astq with bcl2f astq ( version 2.0) .  491 \n 492 \nGenome assembly, curation and classification 493 \nSequencing reads were trimmed using Trim Galore (version 0.6.6)  494 \n(ht tps://github. com/FelixKrueger/Tri mGalore ) with Cutadapt (version 3.4) (55) . G enome 495 \nassemblies were generated using SPAdes (version 3.15. 5) (56) with single-cell mode enabled 496 \n(-- sc) and k-mer sizes 21, 33, 55, and 77.  497 \nScaf folds less than 1,000 nt long were discarded from the assembly. Each assembly 498 \nwas manually curated and cont aminant scaf folds/bins w ere removed using a combinat ion of 499 \nmetagenomic binning with MetaBAT2 ( 57) based on t etra- nucleotide frequencies and 500 \ntaxonomic classification with CAT (v5.2) (58), Blobtools (v 1.1.1) (59)   and Tiara ( v1.0.1) (60) 501 \nand EukRep (v0.6.6)( 61). Bins that were classified as majority eukaryotic by all four 502 \nclassifiers and majority “Euglenozoa” by Blobtools were retained, as well as any unbinned 503 \nscaff old classified as “Euglenozoa” by Blobtoo ls or CAT.  504 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 24  \nThe Holosporales bins w ere first identif ied from  the taxon omic classif ications 505 \ndescribed above. However, all raw assembly bins were assessed for complet eness using 506 \nCheckM (v1.1.2) ( 62) , and all bins > 70% complete were furth er classified GTDB- TK ( v2.3.2) 507 \non KBase. In all the cases, only one bin in each assembly was classified as 508 \nAlphapro teobacteria by GTDB-TK, and w as the bin retained as the Ho losporlaes  509 \nendosymbiont MAG. The retained scaff old set s were manually checked whi ch resulted in 510 \nthe removal of one scaff old from t he B2 Holosporales bin that  was a cont aminate from 511 \nanother well on the 96 well plate.  512 \nThe general statistics, classification summaries, CheckM results for all g enome bins, 513 \nand the GTDB- TK classifications are tabulated in Additional File 5. 514 \nThe SSU  rDNA BLASTs were done t hrough Geneious Prime. In each case, the subject 515 \nsequence with the highest bitscore was considered the top hit. 516 \n 517 \nGene prediction annotation 518 \nEukaryotic gene prediction was done using Companion (v2.2.11) (63)  at  the 519 \nWebServer, with default  sett ings and using Bodo sal t ans  as the ref erence. rDN As were 520 \nannot at ed with Barrnap (v0.9). For one sample (B7) , t he SSU  rDN A sequence was annotated 521 \nas t wo overlapping fragment s, which were manually merged into one sequence by 522 \nalignment in Geneious Prime. Four of t he seven SSU  rDNA sequences are truncated (See 523 \nTable 1 and Additional File 1). Three of  these, B2, G 10, and H10 w ere manually extended by 524 \nalignment of the scaffolds and truncated SSU  rDNA sequences annotated b y Barrnap in 525 \nGeneious Prime and ext racting the longest region with shared homology from each scaff old. 526 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 25  \nAll three sequences were extended by ~70 bp. For G10, the SSU  rDNA annotation w as 527 \ntruncated due to  end of  th e scaff old, w hile B2 and H10 are truncated due to loss of 528 \nhomology, possibly due to miss-assembly of those scaf folds. Holosporales gene predictions 529 \nwere done using PROKKA (v1.14. 6) ( 64).  530 \nCompleteness of the genomes and protein sets was assesses with BU SCO ( v5.3.2) 531 \n(65). 532 \n 533 \n Phylogenetic analyses of SSU  rDNA sequences 534 \nSSU  rDNA sequences were collect ed from GeneBank, filtering f or sequences that  535 \nwere at least 1,000 nucleotides long ( Additional File 9). The sequences were aligned with 536 \nMAFFT (v7.520)  using the FFT-NS-i model (66). The alignment was trimmed manually in 537 \nGeneious Prime to remove all positio ns with less than 50% coverage. The untrimm ed 538 \nalignment used to build the tree is in Additional file 1, while th e sequence accessions are in 539 \nAdditional file 11. The Maximum-Likelihood (ML) tree was built with IQ-TREE (v2.3. 2)  (67) 540 \nusing the TIM3e+I+R3 model, which was t he best  fit model determined by ModelFinder (68), 541 \nwith 1,000 non- parametric bootstrap replicates, and rooting at two outgro up species 542 \nDima s tigella tr ypa nif orm i s  and Rhynchomonas nas ut a . The tree was visualized and plotted 543 \nwith its associated distance matrix using Interactive Tree Of Life ( iTOL) (v7.2)(69) . 544 \n 545 \nPhylogenomic analyses  546 \nFor t he Kinetoplastea, prot ein sets for related species were collect ed from GeneBank 547 \n(Additional File 12) . O rthologous gene sets were identif ied using Orthofind er ( v2.5.4)  (70). 548 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 26  \nProtein sequences for shared single copy orthogroups ( n = 488) were aligned w it h MAFFT 549 \n(v7.520) using the FFT- N S- 2 model (66), and trimmed with TrimAI (v2.0) using the gappyout 550 \noption ( 71). The trimmed alignments were concat enated with AMAS (72) int o a matrix with 551 \n234,727 amino acid sites. The ML tree w as produced f rom a partition ed analysis undertaken 552 \nusing IQ-TREE ( v2. 3.2)  w ith a partitioning scheme that merged the 488 pro teins into 20 553 \npart itions, wit h root ing at  the outgroup Perkin sela  sp.. Model selection was performed by 554 \nModelFinder, and 1,000 non-parametric boot st raps w ere run.  555 \nFor t he Holosporales, prot ein sets were collect ed from GeneBank for related species 556 \nand several closel y related MAG s (Addit ional File 10). For one M AG, G CA_002422845.1, a 557 \nprotein set w as not available on G eneBank, so an annotation w as done using PRO K KA 558 \n(v1.14.6). Orthologous gene sets were identified using O rthof inder (v2.5.4). Protein 559 \nsequences from shared single copy orthogroups (n = 24) were aligned with MAFFT ( v7.520) 560 \nusing the FFT- NS-2 model and t rimm ed wit h TrimAI (v2.0)  using the gappyout  option. The 561 \ntrimm ed alignments were concatenated with AM AS into a matrix w ith 8,004 amino acid 562 \nsit es. The ML t ree was produced from a partitioned analysis undertaken using I Q-TREE 563 \n(v2.3.2) with a partitioning scheme that merged the 24 pro teins into 6 partitions, with 564 \nrooting at the out group containing Temper atibacter mar in us  and Kor diimonas pumila . 565 \nModel selection was performed by ModelFinder, and 1000 non-parametric bootstraps were 566 \nrun.   567 \nThe Kinetoplastea tree (Figure 5) is th e Species Tree from All G enes (STAG) inferred  568 \nspecies tree produced by O rthofinder.  It is inf erred from 331 gene trees from the 569 \nort hogroups where all species are present. 570 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 27  \nAll trees were visualized and plotted either using the Interactive Tree of Life (iTOL)  or 571 \nin R using the package ggtree ( v3.14. 0). 572 \n 573 \nFunctional analyses 574 \nPfams were assigned to each protein set  using Int erProScan (v5. 52-86.0)  using 575 \noption - aap Pfams and with the Pre- calculated match lookup service disabl ed (----disable-576 \nprecalc). Pfam counts were tabulat ed in R. A PCA was run on the count  matrix for Bod o spp. 577 \nafter removal o f Pfam with no variation, which reduced the dataset f rom 2938 variables to 578 \n2545. The PCA was run in R using the function prcomp with scaling and centering.  579 \nKEGG  identifiers were assigned t o each protein set using KofamKOALA or kofamscan (v1.3.0) 580 \n(73). The complet eness of each KEGG module was cal culat ed using kegg-pathway-581 \ncompleteness-tool (v1.3.0) ( https://gi thub.com/EBI- Metagenomics/kegg-pathways-582 \ncompleteness-tool ).  583 \nVenn diagrams showing the proportion of shared orthogroups were generat ed in R 584 \nusing the package ggven (v0.1.10). For an orthogroup to be missing from a species, it must 585 \nbe missing from all genomes in that species. For an orthogroup to be present, it can be 586 \npresent in one or more genomes in that species.   587 \nThe presence of bacterial secretion systems was assessed manually by visualizing the 588 \nresults of KofamKO ALA on the KEGG website using the KEGG  mapper tool Reconstruct. The 589 \nnumber of unique components present in each protein set that  mapped to each secretions 590 \nsystem was counted and compared t o the tot al number of  unique component s list ed for 591 \neach system. The one exception was for the Type VI system, which was annot ated using a 592 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 28  \ncombination of  KEG G Reconstruct, BLASTp and psiblast, where the Type VI proteins from Ca . 593 \nB. vickermanii w ere used as queries to search a database of  the protein sets from all seven 594 \nHolosporales MAG s. The query sequence identifiers, and the significant hit subject 595 \nidentifiers are listed in Additional File 13.  596 \nThe toxin anti/toxin  systems were investigated manually using BLASTp and psiblast, 597 \nwhere t he proteins of Ca. B. vickermanii systems were used as queries to search a database 598 \nof the protein sets from all seven Holosporales MAGs. N terminal signal peptides were 599 \nidentified using SignalP v 6. 0 at the webserver.  600 \n 601 \nAAI, ANI, and AF analyses 602 \nPairwise amino acid identity (AAI)  was calculated f or the Bo d o spp. using the aai.rb 603 \nruby script and BLAST+ v(2.16.0) that is part of the Enveomics Collection at 604 \nhttps: //github.com/lmrodriguezr/enveomics ( 74). For the Holosporales, pairwise AAI w as 605 \ncalculated using FastAAI (34) . AN I and AF were calculated using the tool Skani (v0.2.2) (21).  606 \n 607 \nFigures 608 \nAll p lots were generated in R and edited for publication using Illustrator.  609 \n 610 \nDeclarations 611 \nAvailability of data and materials 612 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 29  \nThe raw reads and annotat ed assemblies f or Bodo and Bodo Holosporales have been 613 \ndeposited in the European N ucleotid e Archive under accession PRJEB97217. The assembly 614 \nsequence f iles, protein sequence files and gff files used in this study have b een deposited in 615 \nZenodo at 10.5281/zenodo.16948154. 616 \n 617 \nCompeting interests 618 \nThe authors declare that there are no  conflicts of interest. 619 \n 620 \nFunding 621 \nThis work was funded by the Wellcome Trust Darw in Tree of Life Awards (218328 622 \nand 226458), and by the Biotechnology and Biological Sciences Research Council (BBSRC), 623 \npart  of UK Research and Innovat ion, through the Earlham Institute’s Core Capabilit y G rant s 624 \n(BB/CCG1720/1 and BB/CCG 2220/1),  it s Strat egic Programme Grant Decoding Biodiversit y  625 \n(BBX011089/1)  and its constituent Work Package 2 BBS/E/ER/230002B, its N ational 626 \nCapability BBS/E/T/000PR9816 (NC1 - Supporting EI’s I SP s and the UK Communit y with 627 \nGenomics and Single Cell Analysis),  and Transf ormat ive Genomics, National Bioscience 628 \nResearch I nfrastruct ure (BBS/E/ER/23N B0006). TAR is supported by a Royal Societ y 629 \nUniversit y Resear ch Fellowship (URF/R/191005). Part of this work was delivered with 630 \nsupport  for t he physical HPC inf rastructure and data center delivered via the N BI Comput ing 631 \ninfrastructure f or Science (CiS) group. 632 \n 633 \nAuthors' contributions 634 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 30  \nSally D. W arring, Conceptualization, Data curation, Formal analysi s, Investigation, 635 \nMethodology, Resources, Visualization, W riting | Jamie McG owan, Con ceptualization, Data 636 \ncuration, Investigation, Methodology,  Resources, Software, Writing |Estelle S . Kilias, 637 \nInvest igation, Met hodology, Resources, Writing | James Lipscombe, Invest igation, 638 \nMethodology, Resources, Writing | El isabet Alacid, Investigation, Methodology, Resources, 639 \nWriting | Tom Barker, Investigation, Methodology | Leah Catchpole, Investigation, 640 \nMethodology, Project administration, Writing | Seanna McTaggart, Funding acquisition, 641 \nProject administration, Writing | Karim Gharbi, Funding acquisition, Methodology, 642 \nResources, Supervision, Writing | Thomas A. Richards, Conceptualization, Funding 643 \nacquisition, Resources, Supervision, W riting | David Swarbreck, Conceptualization, Funding 644 \nacquisition, Supervision | Neil Hall, Conceptualization, Funding acquisition, Resources, 645 \nSupervision, W riting.  646 \n 647 \nAcknowledgements 648 \nWe would like to acknowledge the members of  t he Technical Genomics and Core 649 \nBioinformatics groups at the Earlham Institute, and note the specific contributions of  Chris 650 \nWatkins, Sacha Lucchini, Kendall Baker, and N eil Shearer. W e also acknowledge the work 651 \ndelivered via the Laboratory Managers and Resear ch Computing Groups at EI who manage 652 \nand deliver High Performance Computing at EI.  653 \n 654 \nReferences 655 \n1. Adl SM, S imp son A G B,  Lan e CE, L uke š J, Bas s D, Bow se r SS,  et al . The Rev i s e d Cla s s i f i catio n of 656 \nEukary ot e s . J E ukaryot Mic ro biol. 20 12 S ep t ;5 9 (5) :429–514 .  657 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 31  \n2. Mor e ira  D, Lóp e z - Ga rcía  P, Vic kerman K . An updat ed view  of kine t op la stid  phylog eny using  658 \nenvironme n t a l s e qu ence s  a nd a c lo se r o utgroup: p ropo sal for a n e w c las s i fic ati o n of th e clas s 659 \nKinetopla s tea . I n t  J  Sys t Evol Microbiol . 2004 S ept 1;54(5 ):18 61 –75.  660 \n3. Simpson A GB, S tev en s  JR,  Luk e š  J. The ev olution a nd dive rsity of k in etopl a stid fla g ella tes . Trend s  661 \nPara si t o l. 2006  Ap r ;2 2 (4):168 –74.   662 \n4. Simpson A GB, Luke š J , Roger A J. Th e Evo l u tionary Hi sto r y  of Kine topl a stid s and Th eir 663 \nKinetopla s t s . M o l Bi ol E vol. 2 002 Dec  1;1 9(12):20 71 –83.  664 \n5. Calla h an HA, Li t a k er RW, Noga E J. Molec ul ar Tax onomy of t he  Su bord er Bod oni n a (Or d er 665 \nKinetopla s tida ), I ncludi ng t he Impo rtan t Fi s h P ara s i t e , Ichthyobodo necator . J E uk aryot 666 \nM i crobiol. 2002 Ma r;49 (2):1 19 –28.  667 \n6. D e s c hamp s P , L ara E, Ma ra nde W , Lo pez-G arci a P , E kelund F, Mor eira D. Phy loge n omic Ana lysi s  668 \nof Kinet opla s ti ds Suppor t s That Tryp ano so matid s  Ar o s e f rom wi t hi n Bodoni ds . Mo l Bi ol Ev ol. 669 \n2011 Jan  1;28(1 ):53– 8 .  670 \n7. Br oo ker BE. F ine st r uc ture  o f Bodo sal tan s and  Bo do ca uda tu s (Zooma stigo phora : Pr o tozoa ) an d 671 \nt he ir a f finiti e s with the  Trypan o somati da e. Bull Br  Mus Nat Hi st Z ool. 197 1;22 (3): 89–1 02.  672 \n8. Vick er m an K, P re st on T. C omp ara tive cell  biology  of the k ine topla s tid fl ag ella te s. B iol  673 \nKinetopla s tida . 1 979;2:35 –130 .  674 \n9. D o leze l D, Jirků M, M asl ov  DA ,  Luk e s  J . P hyloge ny of the b odonid fla gella t e s  (Ki n e t o pla st i da) 675 \nbase d on small- subu nit r RNA gen e s eque nces . Int J S yst E vol Mic robiol. 2 000 Se p t  1;50(5) :1943–676 \n51.  677 \n10. Von D er H e yd en S , Ch ao EE, Vic kerman  K , Cav alie r-Smi th T. Ribo somal R N A P hy logen y of 678 \nBodonid a nd Diplon emi d Fla gella t e s an d the Ev olution o f Eug lenoz oa. J E ukaryot Mic r o biol. 20 04 679 \nJ ul y ; 5 1 ( 4) : 4 02 – 1 6.   680 \n11. Jack s o n AP , O t to TD, As let t M, A r m stron g SD, B r ing aud F , Sc hlac h t  A ,  e t al. Kin eto plas tid 681 \nPhylog enomic s  Re veal s t he E volution a r y  Innova tion s  A s s ociat ed with th e Origin s of P a r a siti sm . 682 \nCur r Bi ol. 2016  Jan ;26(2 ):161 –72 .  683 \n12. Szokoli F , Ca s telli M ,  S aba neyev a E, Schr a ll hammer M, Kren e k S, D o a k TG, e t  a l. Di sen tangli ng 684 \nt he  Tax onomy o f  Rick ett sial e s  a nd D es c ri ption o f Tw o Novel S ymbion ts (“C a ndida tus Beal eia 685 \nparamac r o n uclea r i s ”  and “ Ca ndida tu s  F o kinia  cryptic a”) S haring t h e  Cyt opla sm of the Cili at e 686 \nPr o tis t P arame cium b iau relia .  Drak e HL, e dito r .  Appl Env iron Mic robiol . 2016 D e c 687 \n15;82(24) :72 36–47.  688 \n13. Ca stel li M, P e troni G. An Evolution a r y -F o cu s ed R eview  of th e Holo spo rale s 689 \n(A l phap rot eobac t e r i a ) : Diver s i t y , Ho s t In terac tion s ,  an d Ta xonomi c Re-r ank ing a s  Holos porin eae  690 \nSubord. Nov. Mi crob Ec ol . 202 5 Mar 14;8 8(1):1 5 .  691 \n14. Schral lhammer M , P o t e khin A . Epidemio l ogy o f  N uc leu s -D w ellin g Hol os por a : In fec tion , 692 \nTrans m i ssion , Ada p t a t i on, and I nt e r a cti o n w ith Param ec ium. In : Kloc  M, edito r. Sy mbiosi s : 693 \nCellula r , Mol ecul ar, M edic al and  Evolu tio n ary Asp ect s [In t e r n e t]. Cham: S p r in ger I ntern ation al 694 \nPubli shing; 2020 [c i ted 2025  Sep t 24] . p. 105– 35. (R e sult s  a nd Probl ems in Cell Dif fer enti ation ; 695 \nvol. 6 9). Avai labl e f rom: ht tp s: / /li nk . s prin ge r . com/10 .1007 / 978 -3-030 -51849 -3 _4  696 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 32  \n15. Schral lhammer M , C a stell i M, Pe troni G. Phy logene t i c rela tio n s hip s am ong en do s ymbi ot i c R-697 \nbody produc er : Bac te r ia  pr ovidin g their hos t t h e kil l er tra it.  Sys t Appl Mic r o bi ol. 201 8 698 \nM a y;41(3) :21 3–20 .  699 \n16. Fujishima M. Inf ection  a nd Mai n t e nanc e  of Holo s po ra Spec i e s in Pa ram ecium cau da t u m. In : 700 \nFujishima M ,  e dito r. Endo symbi onts in P aramec ium [ Int ern et ]. B erlin, H ei delb erg : S pr i nger 701 \nBerlin Heid el be rg; 2009  [cit ed 2 025 June 1 0]. p. 20 1 –25. ( Mic r o biol ogy  Monograp hs; vol . 12 ). 702 \nA v aila ble from: h ttp s : / /link. spring er.c om / 1 0.1007/978 -3 -54 0-92677 -1_ 8  703 \n17. Fujishima M, Kawa i M, Ya mamo to R. Paramecium caudatum  ac qui re s  he at - s hoc k re si st a nc e i n 704 \ncili ary move ment by infec tion with th e e nd onucle ar s y mbiotic bac te rium Holospora obtusa . 705 \nFEMS Mi crobiol L et t. 2005 Feb ;243 ( 1 ) : 10 1–5.  706 \n18. Midha S , Rigde n DJ , Si ozio s S, Hur st G DD ,  Jack s on AP .  Bodo sal tan s (Kine topl a stida ) i s  d e pend ent 707 \non a nov el Pa r a cae di bac t e r- lik e endo sym bi ont th at po sse s se s  mul t i ple pu tativ e to xin-an t i tox in 708 \ns y stem s. ISME J. 2021 June ;15 (6) :1680– 9 4.  709 \n19. Jamet A, Na s s i f X. N e w P la yer s  in th e Tox in Fi eld: P ol ymorphic  Toxi n System s in Ba c teria . mBio . 710 \n2015 July ;6(3) :e0028 5 -15.  711 \n20. Sche cke nbac h F, Wyl ez ic h C, My lnikov  A P, Wei te re M, A rndt H . Molec ul ar C ompa r iso ns o f 712 \nFre s hw at er and Ma r i ne Is olat e s of the Sa me Morpho s p ecie s of H et ero tr o phic F lag e llat es . Appl 713 \nEnviron Mic robiol. 2 006 Oc t;72(10 ):6 638 –43.  714 \n21. Shaw  J, Yu  Y W . Fa st an d robust me tagen o mic sequen ce c ompari so n thro u gh spa r s e  cha ini ng 715 \nwith skani . N a t  M e th od s . 2 023 Nov;20 ( 1 1):1661–5 .   716 \n22. Schn eider D, Züh lke D, Po e hlein A, Riedel  K, Daniel  R. M et agen ome -A ss embled  Ge n ome 717 \nSequen ce s from Dif f e r e n t  W a st ewa te r  Tre atmen t St age s  i n Ge r ma ny. T hr a s h J C, e ditor . 718 \nM i crobiol Re sour Ann oun c. 20 21 July 8;1 0(27): e00 504-21 .  719 \n23. Muñoz-Góme z SA, He s s  S, Bu rger G, L ang BF , Susko E, Sl amovit s CH, e t al . A n upda ted phylog eny 720 \nof th e Alphap rot eoba c teri a rev ea ls tha t t he para si tic Rick ettsi ale s  a nd Hol o s p or a l es have  721 \nindepend e nt o r i gins . eL if e. 201 9 Feb 25; 8:e425 35.  722 \n24. Park s DH, Rinke C , Chuvoc hina M, Chau mei l PA, W oodc r o ft B J, Ev an s P N, et al. R ecov ery of 723 \nnearly 8,000 me t a g enome -a ssembl ed ge nomes sub stan t ia lly expan d s  the t re e o f  life. Nat 724 \nM i crobiol. 2017 S ep t 11;2 (11):1 533 –42.   725 \n25. Liu L , Wang  Y, Ya ng Y, W a ng D, Che ng SH , Z heng C, et al . Ch art i ng  the c omplex ity of th e 726 \nactiv ated sludg e microbiome  th r oug h a h y brid seq uenc ing s tr a t e gy. Mic r obiome . 2 021 727 \nD e c;9 (1) :205 .  728 \n26. Park s DH, Chuvoc hina M , Wai t e  DW, Ri n ke C,  Ska rshew ski  A, Chaume il PA, et a l . A s t a nda rdized 729 \nbacteri al  taxonom y ba sed on  ge nome ph ylog eny sub st a n tially  revi se s t he  t r e e o f l ife. N at 730 \nBiotechn ol . 2018 N ov ;36 (10):996 – 1004.  731 \n27. Prob st AJ, L add B , Ja ret t JK, Gelle r - Mc G r ath DE, Si eber CMK, Emer s on  J B,  e t al. Di ff ere ntia l d epth 732 \ndis t ribu t ion o f microbia l function an d putativ e s y mbiont s t hr ough sedim ent -ho ste d aqu if er s in 733 \nt he  de ep te rre s t ria l s ub s urf ace . Nat Micr obiol.  2018 Ja n 29;3 (3) :328 –36 .  734 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 33  \n28. H e ss S, S uthaus A , Melko ni an M. “ Can did a t us F inniell a” (Ric ket t siale s, Alpha prot e oba cte ria) , 735 \nN ov el End o s ym biont s  o f Vir i dirap tor i d A moebofl ag ella te s ( Ce rcozoa, R hiz a r i a ) . V oordouw G, 736 \nedito r. Appl En viron Mic robi ol . 2016  Jan 15;82 (2) :659–70.   737 \n29. Jain C , Ro dr i gue z-R LM, Phi llippy  AM,  Konsta nt i ni dis KT, Alu ru S . High th roughp ut  ANI a n alysi s o f 738 \n90K proka r y otic g enome s r evea l s  c lear s pec ie s bounda rie s. Nat Commun.  2018 N ov 739 \n30;9(1) :5114 .  740 \n30. G ori s J , Kons tan tinidi s KT,  Klappe nbach  JA, Co enye T, Van da mme P, Tie dje J M . DNA -DNA 741 \nhybridizatio n v alue s and t h e ir  re lati on shi p to who le -genome  s e q uence  s im ilari tie s . Int J Sy s t Ev ol 742 \nM i crobiol. 2007 J an; 57 (P t 1):81 –91 .  743 \n31. Konsta nt i ni dis KT, T iedje J M. T oward s a Genome -B a s ed Taxon omy for Proka r y ote s. J Ba c teri ol. 744 \n2005 S ept 15;187 (18 ):6258–6 4 .  745 \n32. Rodriguez -R LM, K on s t antini di s  KT. Bypa ssing Cul tivation To Id enti fy  Bact eria l Sp e c ies : Cult ure -746 \nindepend e nt gen omi c app r oa ch e s iden tif y  credibl y distinc t c lus t e rs , av oid c ultivati on bias , and 747 \npr ov ide true in s ig hts int o mi crobial  s pe ci es . Mic r ob e  Mag. 2 014 Mar 1 ;9(3 ):1 11–8 .  748 \n33. Park MJ , Kim  YJ, Park M ,  Yu J , Namirimu T, Roh Y R, et al . Establ i s hm en t of Gen ome Ba sed 749 \nCr i te r ia  for Cla s s if ic ation o f th e  Famil y D esul fovibri ona cea e  and P r opo s al of Two Nov el Ge ner a, 750 \nA l kali de s ul f o vibrio g en . nov.  and S alid e s ulfov ibr i o gen . nov. F r o n t  Mi crobi ol. 2 02 2 May  751 \n25;13:73 8205.  752 \n34. G e r h a r dt K,  Rui z - Pe r e z  C A, Ro d r ig uez - R LM, Jai n C, Tie dje JM , Col e  JR, et al . Fa stA AI: e ffi cient 753 \nestima t i on of gen o me av era ge a mino ac i d ide nt i t y  and phylum -le vel r ela t i on s hi p s using 754 \nt e trame rs o f univ er sal  prote in s .  N uc leic  Acids R e s. 20 25 Apr 22;53 (8) :g kaf348 .  755 \n35. H a r me r  J, Yurchenk o V, N enarokov a A , L ukeš J , Ginge r  ML . Fa r mi ng, slavi ng and  e nslave m ent: 756 \nhis t orie s o f end o s ym biose s  during  kine to pl astid evolu tion . Para si tology . 2018 757 \nSept;145 (10 ):1 311–23 .  758 \n36. O ppe r d oe s  F R, Butenko A ,  Fleg on tov P, Y u rchenk o V , Lukeš J. C ompara tiv e Me tab olism o f 759 \nFree -liv ing Bodo saltans  a nd P ara si tic T r y pano somatid s . J Euk aryo t Microbi ol . 2016  760 \nSept;63 (5) :657 –78.  761 \n37. Bur k i F,  Sandi n MM , Jamy M. D iv er s ity  a nd ec ology  of pro ti s t s r evea led by  met ab arc oding. C u r r 762 \nBiol. 20 21 Oct ;31(19) :R126 7–80 .  763 \n38. Majda S , Boeni gk J , Be i sser D. Intra s pe ci f ic  Varia tion i n P rot i st s: Clue s fo r Micro evo lu tion from 764 \nPot er i o s pum e lla l acu st r i s  (Ch rys ophyc ea e).  Ar c hibald  J ,  edi tor . Gen ome Bi ol Evo l. 2019 Sept 765 \n1;11(9) :2492 –504.  766 \n39. Lac hanc e MA, Le e DK, H s i ang T . De linea ti ng  yea s t sp eci e s  with ge nome av e rage n ucleotid e  767 \nidenti ty: a  cal ibra tion o f AN I with h aplo n tic , het ero thallic  Me t s c hniko wia spec ie s . Antoni e Va n 768 \nLeeuwenho ek . 2020  Dec;113 ( 1 2) :2097–1 06.  769 \n40. D e  Albuq uerq ue N R M, Haa g KL.  Using averag e nuc le otid e id en tity ( A NI ) to eva lu ate 770 \nmicrospo ridia spec i e s  bounda rie s ba sed on  t h eir ge netic  r ela tedn e ss . J Euk aryot Microbio l. 2023 771 \nM a r ; 70 (2): e12944 .  772 \n41. Konsta nt i ni dis KT, Ti edje J M. P rok aryo tic ta xonomy  and phy logen y in the genomic  era: 773 \nadva ncemen ts an d chal lenge s ah e ad . Curr Opin Micr obio l. 2007  O c t ; 10(5 ):504 –9.   774 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 34  \n42. Q in Q L , Xie B B,  Zhang  XY, Ch en XL,  Zhou BC , Zhou J, e t  al . A P ropo sed Genu s B oun da r y  for t h e 775 \nPr o karyo te s Ba s ed on G enom ic In sigh ts . J Bac t eriol. 201 4 J une 15 ;196 (12):221 0 –5 .  776 \n43. Makało wski W , Zha ng J, Bog us k i MS. Co m parative an aly s i s o f 119 6 ort h ologous mou se and 777 \nhuman full -lengt h mRNA and p r ot ei n s e q uence s. Genome  R e s. 1996 Sep t;6 (9):846 – 57.  778 \n44. W i bberg D , Rupp O , Blo m J, J elon ek  L, Kr öb er  M ,  Ve rwaai jen B, e t  a l. Deve lo pmen t of a  779 \nRhiz oc toni a s olan i A G 1 -I B Spec i fic G e n e Model E nabl es Compa r a t iv e Genom e A n a lyse s be twee n  780 \nPhytopa thog enic R . sol ani A G1- IA, AG1 -I B, AG3 and AG8 I sol at e s. Yu J H, e dito r .  P LOS ONE. 2015  781 \nD e c 21;1 0(12 ): e0144 769.  782 \n45. W i bberg D , Stad l er M, La mb ert  C , Bunk B,  Spröer C, Rüc ke rt C , e t  a l . High q uality g enome 783 \ns eq uence s o f thi rte en Hypo xyla cea e ( A s c omyc ot a ) s tr ength en t h e  phylog ene tic  fa mily  back bone 784 \nand ena bl e the d i s c ove ry o f  new ta xa . F ungal D i ver s . 2021 Ja n; 106(1 ):7 –28.   785 \n46. G rou s s i n M, M azel F,  Alm EJ. C o- ev olutio n a nd Co - s pe cia tio n o f Ho s t -G ut  Bac teri a  S ystem s . Cell 786 \nH os t Microb e.  2020 Jul y;28 (1):12 – 22.  787 \n47. Russo L , Mill er A D, Too ker J, Bjorn s t a d ON, S hea K. Q u an t i tativ e ev olution ary p at tern s in 788 \nbiparti te ne twork s: Vic arianc e, phy logen etic  tr ack ing or dif fuse c o -evolu ti on?  M e thod s Ecol  Evol.  789 \n2018 Mar;9(3 ):76 1 –72.  790 \n48. Schral lhammer M , P o t e khin A . Epidemio l ogy o f  N uc leu s -D w ellin g Hol os por a : In fec tion , 791 \nTrans m i ssion , Ada p t a t i on, and I nt e r a cti o n w ith Param ec ium. In : Kloc  M, edito r. Sy mbiosi s : 792 \nCellula r , Mol ecul ar, M edic al and  Evolu tio n ary Asp ect s [In t e r n e t]. Cham: S p r in ger I ntern ation al 793 \nPubli shing; 2020 [c i ted 2025  Jun e 10] . p.  105–35 . (R e sult s  a nd Probl ems in Ce ll Dif fer enti ation ; 794 \nvol. 6 9). Avai labl e f rom: ht tp s: / /li nk . s prin ge r . com/10 .1007 / 978 -3-030 -51849 -3 _4  795 \n49. Frelie r PF , Lo y JK, Kruppe nb ach B . Tr an sm ission o f Nec ro t izing  H e pat opanc rea ti t i s  in  P enae u s  796 \nvanna mei. J I nv erte br  Path ol. 19 93 Jan ;6 1(1):4 4 –8.  797 \n50. Schu lz  F , La gkouv ardos I, Wa sch er F,  Ai st lei t ne r  K , Kos t a nj š e k R, Horn M. L if e in a n  unusual 798 \nintr a cell ular nic he : a b act erial s y mbiont i nfec t i ng the nuc le us o f am o ebae . I SME J .  201 4 Aug 799 \n1;8(8) :1634– 44.  800 \n51. Kořený L , Lukeš J, O b or ní k M . Ev olu tion o f the h aem synthe tic  pathw ay in kinet opl a stid 801 \nflagel la te s: An e ssen tia l pa thwa y tha t is not e s s e n t i al a ft er al l? Int J Pa ra s itol . 2010 802 \nFeb;40 ( 2 ) :1 49 –56.  803 \n52. Alves JMP , Vo egtly  L, M a tveye v A V, L ara AM, D a Si lva F M, Ser rano M G, e t a l. I den tific a tion a nd 804 \nPhylog ene tic Analy si s of H eme S ynthe si s  Gene s  in  Trypan os omatid s and  Th eir Bac terial 805 \nEndosymbion ts . Lo pez-G arcia  P, ed i t or . P LoS O NE. 20 11 Aug 10 ;6(8 ): e23518 .  806 \n53. Klein C C , Alve s JM P, S e r ran o M G , Buck GA, Va s c onc el o s  ATR,  Sag o t  M F , e t al.  Bi o sy nthe s i s  of 807 \nVitamin s an d Co fa ct or s  in  Bact erium -Har bouring Trypano soma tid s Dep end s on th e Symbi ot i c 808 \nA s socia t i on a s R evea led by  G enomic  Ana lyse s. Parki n s on J , edi tor . PLoS  O N E. 201 3 Nov 809 \n19;8(11) :e79 786.  810 \n54. Macau lay I C , Hae rty W, Kumar P,  Li Y I, Hu TX, Teng MJ, e t  al. G & T - s eq : par allel  s e qu encin g of 811 \ns ing le -ce ll ge nom e s  and tr an script ome s. Nat M etho ds. 2015 June ;1 2(6 ):519–22 .  812 \n55. Mart in  M . Cutada pt r emov es  adap ter se qu enc e s  from h igh- throu ghput sequenc i ng read s. 813 \nEMBnet .journal . 2011 May  2;17(1 ):10 –2 .   814 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 35  \n56. Banke vich  A, Nur k  S, A ntipov  D , Gure vic h AA, Dvorkin M, Kulikov  AS, e t  al. S P Ad e s :  A N ew  815 \nG e nome A s s embl y Algorit hm a nd It s  A p p lica t io n s  to S ingl e -Ce ll S e quenc ing. J Com pu t  Biol. 201 2 816 \nM a y;19(5) :45 5–77 .  817 \n57. Kang DD, L i F, Kirt o n E, Tho ma s  A,  Egan  R, An H, et al . Me taB AT 2: a n ada p t i ve bi n ning a lgorithm 818 \nfor  r obust and e f f i cien t g enome r ec on s t r u ction  f r om metage nom e a s s em bli e s. P e erJ . 819 \n2019;7:e 7359.  820 \n58. von Mei jenf eldt FA B, Ar k hip ova K, Camb uy  DD, C outinh o F H , D u tilh  BE. Ro bu st t a xonom ic 821 \nclas si fica t io n  of unch a rted  microbia l seq uenc e s  a nd bin s wi t h  C AT and BA T . Gen o me Biol . 2019 822 \nO c t  22;20(1 ):217 .  823 \n59. Lae tsch D R, Blaxter ML . BlobToo l s: In te r r oga t io n o f geno me a s sem blie s. F10 00R es earch. 2017 824 \nJ uly 31;6 :1287.  825 \n60. Karlick i M, Anton ow icz S, K arnkowska  A . Tiara: de ep le a rning-b a sed c la s s i fi cation sys tem for 826 \neukary ot i c sequ enc e s .  B ioin forma Oxf E n gl. 20 22 Jan 3 ;38(2 ):344–5 0 .  827 \n61. W e s t PT,  Pro bst A J, Grigo riev I V, Thoma s  BC, Ban field JF . Genom e- recon structio n for eu k aryot es 828 \nfr om comple x natural microbia l comm u n itie s. G e nom e Re s. 201 8 Ap r;28(4 ):569 –8 0.  829 \n62. Park s DH, Imel for t  M, Ske nn erton  CT, Hugen holtz  P,  Tyson GW. Chec kM : a sse s sin g the q u ality o f  830 \nmicrobia l geno me s  r ec overed f rom i s ola te s, sing l e ce ll s, and me t a genom es . G eno me Re s .  2015 831 \nJ uly;25(7 ):10 43–55 .  832 \n63. H a e s e - Hil l W, Crouch K , Otto T D. Ann ota tion  and v isua liza tion o f para si te , fung i a nd art hr opod 833 \ngenome s w ith Compa nio n. Nuclei c Aci d s  Re s. 2 024 July  5;52( W1): W39–44 .  834 \n64. Seema nn T. P r ok ka: r api d proka ryotic  ge no me anno tatio n . Bioin fo rmatic s . 2 014 J uly  835 \n15;30(14) :20 68–9.  836 \n65. Manni M, Be r k eley  MR , Seppe y M, S imã o F A, Zdobnov  EM. BUS CO Upd at e: Nove l  and 837 \nStr e amline d Wo r k flow s along  with B road e r  and D eep er P hylog e ne tic Cov er age f o r S coring of 838 \nEukary ot i c, Pr oka r y otic, a nd Vi ral G enom e s. K elley  J, edito r. Mol Bi o l Evol . 2021 Se pt 839 \n27;38(10) :46 47–54.  840 \n66. Katoh K, S tandl ey  DM. MA F FT multipl e s equ enc e al ignm ent so ftwar e v er s ion 7: i mprove ment s  841 \nin perf ormanc e a nd us abi lity. M ol Bi ol Ev ol. 2013  Apr;30 (4) :772–80 .  842 \n67. Minh B Q, Sc hmidt HA,  Che rnomor O, Sc h r e mpf D ,  W o o dham s MD, von Hae sel e r  A , et al. I Q-843 \nTREE 2 : N e w Model s an d Ef fici en t Me tho ds  f or Ph ylog e net ic Inf e r e nc e in  t h e Gen omic  Era. Mol 844 \nBiol Ev ol. 202 0 May 1;37(5) :1530 –4.  845 \n68. Kalya anamoo rthy S, Minh BQ, W ong TKF, v on Haes eler A , Je rmii n LS . Model Fi nder : fa st  mod el 846 \ns el ection for a ccurat e phylog ene tic e s tim ate s. Nat Meth od s. 2017 Ju n e;14 (6):587 – 9.  847 \n69. Letunic  I, Bo rk P. I nte rac t iv e Tr ee o f Life (i TO L ) v6: r ec en t upda te s to t he p hyloge n etic tr ee 848 \ndis p lay a nd a nnot ation  t o ol . Nucle ic Ac id s  Re s. 2024 July  5;52(W1 ): W78–82 .  849 \n70. Emms DM ,  Kelly  S. O rt h o Finde r: phy loge ne t i c or thology  inf erenc e fo r compa r a t i ve g enomic s. 850 \nG e nome Biol.  2019 D ec ;20(1 ):238 .   851 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 36  \n71. Capell a- G ut ié rrez S, S illa -M artí nez J M, G abald ón T. t rimAl: a  t o ol for au tomat ed a lignme nt  852 \nt rimmi ng in l arge -sc ale  phylog en etic  anal y s e s. Bi o inf or m atic s . 2009  Aug 1;25 ( 1 5) : 1972–3.   853 \n72. Bor o wiec  ML. AM AS: a fa st t oo l for alig n ment manip ula tion a nd compu t in g of su mmary 854 \ns t ati sti cs . Pe er J. 201 6 Jan 28 ;4:e1660 .   855 \n73. Ar a maki T, Blanc -Ma thieu R, End o H, Oh kubo K , Kaneh isa  M, Go t o S, et a l . Kofam KOALA: KE G G 856 \nO rth olog a s signm e nt ba s ed on pro file  HMM a nd adap tiv e score thre sh old. Val en cia  A, edit or. 857 \nBioinfo rmatic s . 20 20 Apr 1;36(7 ):22 51 –2.  858 \n74. Rodriguez -R LM, K on s t antini di s  KT. The  e nveo mics col lec tion:  a toolbox  f or s pe cial iz ed  analy s e s 859 \nof microbia l ge n ome s  and met ag enome s  [In tern et ]. Pe erJ Prep rint s; 2016 [c it ed 2025  Apr  1]. 860 \nA v aila ble from: h ttp s : / /pe erj.com / p rep ri nts/1900 v1  861 \n 862 \n  863 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 37  \nFigure Legends 864 \nFigure 1 – SSU rDNA phylo gen y of Bodo spp. A. Pa irwise ANI of SSU  rDNA sequences from 865 \nthe seven single cell genomes B. ML phylogeny of SSU rDNA sequences from Bodo spp. and 866 \nrelated genera. The tree scale is substitutions per site. The seven single cell genomes are 867 \nshaded yellow, blue and pink to distin guish the three clades they form. B. s altans lake 868 \nKonstanz is shaded purple. The tw o columns on the right side of the tree show the pairwise 869 \nANI between each taxon and the SSU  rDNA from B. s alt ans  lake Konstanz and B. edax . 870 \n 871 \nFigure 2 – Genomic diversity of Bodo  spp. A. ML phylogeny of 488 single copy ort hologous 872 \nproteins. Tree scale is substitutions per site B . Heatmap showing pairwise average AAI 873 \nvalues for all Bodo genomes C.  Heatmap showing pairwi se AN I (lower triangle) and AF 874 \n(upper triangle) D.  N umber and proport ion of shared ort hogroups for each Bodo  spp. E.  Bar 875 \nplot showing the proportion of prot ein annotat ions in each genome that  have a Pfam 876 \nannot at ion F. PCA plot generated fro m the tally of Pfam domain p resent in each genome. In 877 \nall plots the th ree B odo  spp. from sin gle cell genomes are shaded yellow, blue and pink to 878 \ndistinguish the three clades they form while B. saltans lake Konst anz i s shaded purple. 879 \n 880 \nFigure 3 – Genomic diversity of Holosporales endosymbionts. A. ML phyl ogeny of 24 single 881 \ncopy orthologous proteins. Tree scale is substitutions per site B.  Heatmap showing pairwise 882 \nANI (lower triangle) and AF (upper triangle) for the Holosporlaes associated with B odo  single 883 \ncells C. Number and proportion of shared orthogroups for each Holosporales species. In all 884 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 38  \nplot the Holosporales from the seven single cell genomes are shaded yellow, blue and pink 885 \nto distinguish the three clades they form while “ Ca. B. vickermanii” is shaded purple. 886 \n 887 \nFigure 4 – Functional diversity among Holosporales endosymbionts. Heatmap in blue 888 \nshowing t he completeness values f or KEGG  modules found in each Bodo Holosporales 889 \ngenome and Ca. B. vickermanii genome. Only modules that are at least 50% complete in one 890 \nor more genome are shown. Heat map in red shows t he number of  prot eins belonging to 891 \neach of the bacterial secretion systems found in the genomes. The block t o the right o f  the 892 \nheatmap shows the tot al number of proteins belonging to each secret ion system, as li sted 893 \non the KEGG Brite database.  894 \n 895 \nFigure 5 – Genomic uniqueness in Genus Bodo. S pecies tree infe rred by STAG. Tree scale is 896 \nsubstitutions per site. The support values are the number o f individual gen e trees that 897 \ncontain that  bipartit ion. The bar plots show t he tot al number of orthogroups and t he 898 \nnumber of unique orthogroups in each genome protein set . Pie charts are shaded t o show 899 \nthe proport ion of orthogroups unique t o the three genera Bodo, Tr y pano s oma  and 900 \nLei shmania .  901 \n 902 \nFigure 6 – Functional comparison of Kinetoplastida. Heat map showing t he completeness 903 \nvalues for KEGG modules in each genome. Only modules that are at  least 60% complete in 904 \none or more genomes are shown.  905 \n  906 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\n 39  \nAdditional files 907 \nAdditional File 1 - Bodo  and related spp. SSU rDNA alignment.  FASTA f ile cont aining t he 908 \nuntrimm ed SSU  rDNA alignment used to build t he tree in Figure 1. 909 \nAdditional File 2 - AAI summary statistics for Bodo  spp. Excel sheet listing the summary 910 \nmetrics out put from the aai.rb script. Column “no. proteins used” shows t hat number of 911 \npairwise comparisons used to calculate the average AAI. 912 \nAdditional File 3 – Bodo  skani AF full matrix. TSV file showing the AF calcul ated by skani in 913 \nboth directions for each pairwise comparison.  914 \nAdditional File 4 - Pfam counts and PCA loadings. Excell workbook wit h two sheet s. The 915 \nfirst sheet shows the Pfam frequencies in each B odo  genome. The second sheet shows t he 916 \nloadings given to each Pfam for PC1-8.  917 \nAdditional File 5 - Bodo  SC raw genome assemblies statistics per bin. Excell sheet 918 \ntabulating statistics for all the raw single-cell genome assemblies, split by MetaBat2 bins.   919 \nAdditional File 6 - FastAAI distance matrix. Excell sheet containing the A AI values f or the 920 \nHolsoporales heatmap in Figure 3A.   921 \nAdditional File 7 – Holosporlaes skani AF full matrix. TSV file showing the AF calculated by 922 \nskani in both directions for each pairwise comparison.  923 \nAdditional File 8 - Holosporales all KEGG completeness.  Excell sheet showing the 924 \ncompleteness values for KEGG  modul es ident if ied in each Holosporales protein set.  925 \nAdditional File 9 - Toxin antitoxins in Bodo  Holosporlaes B7 F10 figure.  P DF do c um e n t 926 \nshowing a: A. schematic of the putative toxin/antitoxin loci in B7 and F10 B. alignment 927 \nbetween the large toxin pro tein and alternative toxin, showing region of homology C. 928 \nSingnalP plots of N  termini of the p utative multi-domain toxins from B7 and F10.  929 \nAdditional File 10 - Kinetoplastids all KEGG completeness. Excell sheet showing the 930 \ncompleteness values for KEGG  modul es ident if ied in each Bodo  and Kinet oplast ea protein 931 \nset.  932 \nAdditional File 11 - SSU  rDNA sequences u sed in ph ylogeny . Excell sheet with N CBI 933 \naccession identifiers for all SSU  rDNA sequences used to construct the phylogeny in Figure 1.  934 \nAdditional File 12- all NCBI protein sets . Excell workbook with two sheets listing the N CBI  935 \naccession identifiers for all pro tein sets used in this study. Kinetoplastea sets are on sheet 1 936 \nand Holosporales sets are on sheet  two. 937 \nAdditional File 13 - TypeVI secretion system protein IDs . Excell sheet listing the Type VI 938 \nsecretion system protein id entifiers used as queries from Ca. B. vickermanii, and the protein 939 \nidentifiers f rom the Holosporales genomes that were signif icant hits.    940 \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\nA\nB\nBootstrap\n   100\n   80-99\n   50-80\nB. saltans AY490226.1\nBodo H10\nBodo G10\nBodo B2\nBodo A8\nBodo A10\nB. edax AY028451.1\nB. saltans JC02 AY490227.1\nB. saltans JF693632.1\nB. uncinatus AF208884.1\nB. saltans strain Petersburg AF208887.1\nB. saltans strain HFCC309 DQ207572.1\nBodo B7\nBodo F10\nB. saltans AY490224.1\nB. saltans isolate NG MF962814.1\nB. saltans JC18 AY490229.1\nB. saltans PP5 AY490223.1\nB. saltans JC03 AY490228.1\nB. saltans AY490222.1\nB. saltans AY490232.1\nB. saltans AY490230.\nB. saltans AY490231.1\nB. sp. TGS2 AB585965.1\nB. saltans strain Konstanz AF208889.1\nB. sp. TGKH8 LC000678.1\nB. saltans AY028452.1\nB. saltans strain HFCC12 DQ207569.1\nB. saltans AY490234.1\nB. saltans strain SCCAP BS364 AY998648.1\nB. sp. ATCC 50149 AY028449.1\nB. saltans strain HFCC311 DQ207574.1\nB. saltans strain HFCC14 DQ207571.1\nB. saltans MH614643.1\nB. saltans strain IOW92 KX431511.1\nB. saltans AY490233.1\nB. saltans strain HFCC310 DQ207573.1\nB. saltans strain HFCC323 DQ207575.1\nB. saltans strain HFCC13 DQ207570.1\nB. caudatus AY028450.1\nB. curvifilus AY425015.1\nB. sp. isolate COLPROT774 MW355419.1\nB. caudatus AY490218.1\nB. caudatus strain SCCAP BC330 AY998649.1\nB. caudatus AY490215.1\nB. sorokini AF208888.1\nB. sorokini strain ATCC 50641 AY425018.1\nN. designis strain SCCAP BD54 AY998650.1\nN. designis strain SCCAP BD55 AY998651.1\nN. designis strain SCCAP BD56 AY998652.1\nN. designis strain SCCAP BD57 AY998653.1\nB. saliens AF174379.1\nB. designis AY425016.1\nB. designis AF209856.1\nN. designis strain SCCAP BD52 AY998646.1\nB. rostratus AY425017.1\nN. designis strain SCCAP BD50 AY998643.1\nN. designis strain SCCAP BD23 AY998644.1\nN. designis strain SCCAP BD51 AY998645.1\nB. designis AY490235.1\nN. designis strain SCCAP BD53 AY998647.1\nB. designis strain DH AF464896.1\nB. celer AY490221.1\nD. trypaniformis strain SCCAP DIM74 AY998641.1\nR. nasuta strain SCCAP RH3 AY998642.1\n89.6\n91.0\n82.5\n83.4\n76.4\n76.0\n81.6\n81.7\n94.5\n95.0\n95.3\n0.0\n82.5\n77.6\n83.8\n83.6\n97.6\n95.9\n82.3\n82.7\n96.7\n89.5\n94.6\n95.9\n99.9\n95.4\n82.8\n83.6\n97.8\n94.7\n81.7\n81.7\n80.7\n81.4\n99.7\n88.5\n81.4\n82.3\n97.6\n96.1\n96.2\n96.9\n82.3\n82.4\n80.5\n79.6\n89.5\n89.0\n98.4\n95.0\n94.2\n94.8\n79.0\n79.1\n95.1\n100\n95.3\n99.6\n95.6\n98.2\n95.5\n96.8\n84.6\n84.4\n97.6\n96.0\n83.3\n84.1\n93.5\n94.8\n94.8\n96.0\n94.7\n95.8\n82.4\n83.1\n82.0\n82.7\n81.3\n81.5\n87.0\n87.6\n0.0\n95.3\n96.7\n94.6\n77.5\n77.4\n85.9\n85.9\n95.7\n98.0\n82.0\n82.9\n95.2\n99.6\n94.9\n99.0\n99.9\n95.4\n98.7\n94.6\n95.3\n99.6\n95.2\n99.6\n78.4\n78.5\n93.9\n94.4\n94.7\n99.4\n98.4\n95.0\n95.3\n99.9\n82.8\n83.8\n94.7\n98.0\n96.2\n96.8\n93.4\n96.0\n87.4\n88.0\n82.0\n82.3\n94.2\n94.9\nB. saltans strain Konstanz\nB. edax\nTree scale: 0.1\nPercent identity\n75\n77\n80\n82\n85\n87\n90\n92\n95\n97\n100\n 99.82\n 98.05\n 98.05\n 97.68\n 98.12\n 97.94\n 98.05\n 98.05\n 97.63\n 97.99\n 97.94\n 99.84\n 99.93\n 99.93\n 99.84\n 99.93\n 99.93\n100\n 99.92\n 99.92\n95 96 97 98 99 100\nBodo F10\nBodo B7\nBodo A10\nBodo A8\nBodo G10\nBodo B2\nBodo H10\n100\n100\n100\n100\n100\n100\n100\n100\nF10\nB7\nA10\nA8\nG10\nB2\nH10\n18s rDNA pairwise nucleotide identity %\nFigure 1\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\nA\nC D\nE F\nB\nBootstrap\n   100\nBodo H10\nBodo G10\nBodo B2\nBodo A8\nBodo A10\nBodo B7\nBodo F10\nBodo saltans \nPerkinsela\n51.3\n62.2\n62.2\n98.8\n100\n51.5\n62.6\n62.6\n85.0\n85.0\n97.9\n98.7\n100\n51.4\n62.3\n62.1\n100\n51.6\n62.5\n62.5\n84.9\n85.0\n100\n51.7\n100\n51.5\n98.0\n100\n51.5\n62.4\n62.5\n84.8\n84.8\n97.7\n100\n100\nTree scale: 0.1\nAverage pairwise amino acid identity (AAI) %\nH10\nG10\nB2\nA8\nA10\nB7\nF10\nB. saltans\n50\n55\n60\n65\n70\n75\n80 85 90 95 100\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n98.39\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n99.24\n84.17\n84.05\n84.16\n84.24\n84.14\n84.09\n98.42\n98.23 99.02\n0 10 20 30 40 50 60 70 80 90 100\nB.sal.\nB.sal.\nF10\nB7\nA10\nA8\nB2\nH10\nG10\n 0.00  0.00\n81.56\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n83.12\n 0.00\n 0.00\n 0.00\n15.63\n15.46\n 0.00\n 0.00\n 0.00\n15.30\n15.15\n80.14\n 0.00\n 0.00\n 0.00\n14.70\n14.53\n76.72\n82.20\n0\n10\n20\n30\n40\n50\n60\n70\n80\n90\n100\nF10\nB7\nA10\nA8\nB2\nH10\nG10\nAligned Fractiom (AF) %\nAverage nucleotide Identity (ANI) %\nB. saltans\nB7/F10 A8/A10\nB2/\nG10/H10\n806\n(5.8)\n571(4.1) 199(1.4)\n570(4.1)257(1.9)\n11(0.8) 898\n(6.5)\n77(0.6) 181(1.3)\n138(1.0)\n280(2.0) 1076(7.8)\n530(3.8) 467(3.4)\n7663(55.5)\nShared orthogroups\nH10\nG10\nB2\nA10\nA8\nF10\nB7\nB. sal\n0.00 0.50 1.00\nProportion proteins with Pfam annotations\nOther proteinsProteins with Pfam annotation\nB. saltans A8A10\nB7\nF10\nB2\nG10H10−0.25\n0.00\n0.25\n0.50\n−0.9 −0.6 −0.3 0.0\nPC1 (26.38%)\nPC2 (20.4%)\nPrinciple components analysis base on Pfam frequencies\nFigure 2\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\nB C\nA\nA10\nA8\nB2\nMAG GCA_018061565\nB7\nF10\nCa. B. vickermanii\nMAG GCA_002422845\nCa. C. indipagum\nCa. C. primus\nCa. C. primus 43202\nCa. N. abundans\nCa. N. abundans 44043\nCa. G. agglomerans\nCa. H. endosymbioticus\nH. curviuscula\nH. elegans E1\nH. undulata HU1\nH. obtusa F1\nCa. H. penaei\nHolosp. bac. Namur\nCa. B. paramacronuclearis 11III1\nCa. B. paramacronuclearis 15I1\nCa. N. amoebiphila\nCa. P. acanthamoebae\nC. varicaedens\nCa. O.thessalonicensis L13\nCa. O. acanthamoebae\nCa. P.symbiosus\nC. Finniella inopinata\nMAG GCA_002791835\nMAG GCA_003542655\nG10\nH10\nMAG GCA_024236035\nK. pumila\nT. marinus\neoAcanthamoeba sp. UWC8\nA10\nA8\nB2\nMAG GCA_018061565\nB7\nF10\nCa. B. vickermanii\nMAG GCA_002422845\nCa. C. indipagum\nCa. C. primus\nCa. C. primus 43202\nCa. N. abundans\nCa. N. abundans 44043\nCa. G. agglomerans\nCa. H. endosymbioticus\nH. curviuscula\nH. elegans E1\nH. undulata HU1\nH. obtusa F1\nCa. H. penaei\nHolosp. bac. Namur\nCa. B. paramacronuclearis 11III1\nCa. B. paramacronuclearis 15I1\nCa. N. amoebiphila\nCa. P. acanthamoebae\nC. varicaedens\nCa. O.thessalonicensis L13\nCa. O. acanthamoebae\nCa. P. symbiosus\nC. Finniella inopinata\nMAG GCA_002791835\nMAG GCA_003542655\nG10\nH10\nMAG GCA_024236035\nK. pumila\nT. marinus\neoAcanthamoeba sp. UWC8\nTree scale: 1\nBootstrap\n   100\n   80-99\n   50-80\nAAI >= 55 % , AAI <=85 %\nAAI >= 95%\nPairwise Average Amino Acid Identity (AAI) (%)\n34\n40\n46\n52\n58\n64\n70\n76\n82\n88\n95\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n 97.55\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n  0.00\n 99.95\n  0.00\n 85.84\n 86.31\n  0.00\n 85.42\n 86.02\n 99.98\n100 100\n0 10 20 30 40 50 60 70 80 90 100\nCa. B.v\nCa. B.v\nF10\nB7\nA10\nA8\nB2\nH10\nG10\n 0.00  0.00\n85.33\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n93.45\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n 0.00\n28.09\n27.73\n27.89\n 0.00\n 0.00\n 0.00\n28.35\n29.52\n20.79\n90.89\n0\n10\n20\n30\n40\n50\n60\n70\n80\n90\n100\nF10\nB7\nA10\nA8\nB2\nH10\nG10\nCa. B.v\nB7/F10 A8/A10\nB2/G10/H10\n117\n(4.9)\n184\n(7.7)\n156\n(6.5)\n299\n(12.5)\n302\n(12.7)\n1\n(0.0)\n633\n(26.2)\n6\n(0.3)\n19(0.8)\n6\n(0.3)\n3\n(0.1)\n27\n(1.1)\n10\n(0.4)\n49\n(2.1)\n555\n(23.2)\nAligned Fractiom (AF) %\nAverage nucleotide Identity (ANI) % Shared orthogroups\nFigure 3 .CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\nMethionine degradation (M00035)\nLysine biosynth., acetyl-DAP pathway, aspartate => lysine (M00525)\nLysine biosynth., DAP aminotransferase pathway, aspartate => lysine (M00527)\nLysine biosynth., DAP dehydrogenase pathway, aspartate => lysine (M00526)\nLysine biosynth., succinyl-DAP pathway, aspartate => lysine (M00016)\nCitrate cycle (TCA cycle, Krebs cycle) (M00009)\nCitrate cycle, first carbon oxidation, oxaloacetate => 2-oxoglutarate (M00010)\nCitrate cycle, second carbon oxidation, 2-oxoglutarate => oxaloacetate (M00011)\nGluconeogenesis, oxaloacetate => fructose-6P (M00003)\nGlycolysis (Embden-Meyerhof pathway), glucose => pyruvate (M00001)\nGlycolysis, core module involving three-carbon compounds (M00002)\nPentose phosphate pathway (Pentose phosphate cycle) (M00004)\nPentose phosphate pathway, non-oxidative phase, fructose 6P => ribose 5P (M00007)\nPRPP biosynth., ribose 5P => PRPP (M00005)\nPyruvate oxidation, pyruvate => acetyl-CoA (M00307)\nGlycogen biosynth., glucose-1P => glycogen/starch (M00854)\nGlyoxylate cycle (M00012)\nMethylcitrate cycle (M00982)\nUDP-N-acetyl-D-glucosamine biosynth., prokaryotes, glucose => UDP-GlcNAc (M00909)\nUndecaprenyl phosphate (UP) \nα\n-L-Ara4N biosynth., UDP-GlcA => UP \nα\n-L-Ara4N (M00761)\nCytochrome bc1 complex respiratory unit (M00151)\nCytochrome bd ubiquinol oxidase (M00153)\nCytochrome c oxidase, prokaryotes (M00155)\nF-type ATPase, prokaryotes and chloroplasts (M00157)\nNADH (M00144)\nSuccinate dehydrogenase, prokaryotes (M00149)\nCAM (Crassulacean acid metabolism), dark (M00168)\nCAM (Crassulacean acid metabolism), light (M00169)\nReductive citrate cycle (Arnon-Buchanan cycle) (M00173)\nReductive pentose phosphate cycle (Calvin cycle) (M00165)\nCMP-KDO biosynth. (M00063)\nKDO2-lipid A biosynth., Raetz pathway, LpxL-LpxM type (M00060)\nKDO2-lipid A biosynth., Raetz pathway, non-LpxL-LpxM type (M00866)\nFatty acid biosynth., elongation (M00083)\nFatty acid biosynth., initiation (M00082)\nPhosphatidylcholine (PC) biosynth., PE => PC (M00091)\nPhosphatidylethanolamine (PE) biosynth., PA => PS => PE (M00093)\nC1-unit interconversion, eukaryotes (M00141)\nHeme biosynth., animals and fungi, glycine => heme (M00868)\nHeme biosynth., bacteria, glutamyl-tRNA => coproporphyrin III => heme (M00926)\nHeme biosynth., plants and bacteria, glutamate => heme (M00121)\nLipoic acid biosynth., eukaryotes, octanoyl-ACP => dihydrolipoyl-H (M00882)\nLipoic acid biosynth., octanoyl-CoA => dihydrolipoyl-E2 (M00884)\nLipoic acid biosynth., plants and bacteria, octanoyl-ACP => dihydrolipoyl-E2/H (M00881)\nPimeloyl-ACP biosynth., BioC-BioH pathway, malonyl-ACP => pimeloyl-ACP (M00572)\nSiroheme biosynth., glutamyl-tRNA => siroheme (M00846)\nAdenine ribonucleotide biosynth., IMP => ADP,ATP (M00049)\nAdenine ribonucleotide degradation, AMP => Urate (M00958)\nDeoxyribonucleotide biosynth., ADP/GDP/CDP/UDP => dATP/dGTP/dCTP/dUTP (M00053)\nGuanine ribonucleotide biosynth., IMP => GDP,GTP (M00050)\nPyrimidine deoxyribonucleotide biosynth., UDP => dTTP (M00938)\nPyrimidine ribonucleotide biosynth., UMP => UDP/UTP,CDP/CTP (M00052)\nB2\nH10\nG10\nA8\nA10\nB7\nF10\nCa. B. v.\n10\n8\n0\n15\n0\n10\n7\n0\n13\n0\n10\n7\n12\n15\n0\n9\n7\n9\n10\n2\n10\n7\n8\n10\n2\n11\n8\n2\n12\n2\n10\n7\n7\n8\n2\n11\n7\n4\n8\n2\nSEC-SRP\nType III - Flagellar export apparatus\nType IV - Conjugal transfer pilus assembly protein\nType VI\nTAT\n14\n12\n21\n16\n4\n Amino acid\n metabolism\n Carbohydrate\nmetabolism\n Energy\nmetabolism\n Lipid\nmetabolism\n Metabolism\nof cofactors\nand vitamins\n Nucleotide\nmetabolism\n Secretion\n systems\nKEGG Module Completeness (%)\n0\n10\n20\n30\n40\n50\n60\n70\n80\n90\n100\nFigure 4\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\nLeishmania\nOrhogroups in genus  = 8,237\nOrthogroups unique = 475\nTrypanosoma\nOrthogroups in genus = 11,183\nOrthogroups unique = 3,877\nBodo\nOrthogroups in genus = 13,816\nOrthogroups unique = 7,321\nPerkinsela sp.\nBodo saltans\nBodo B7\nBodo F10\nBodo A10\nBodo A8\nBodo H10\nBodo B2\nBodo G10\nTrypanosoma vivax\nTrypanosoma brucei gambiense\nTrypanosoma brucei equiperdum\nTrypanosoma equiperdum\nTrypanosoma brucei brucei\nTrypanosoma grayi\nTrypanosoma melophagium\nTrypanosoma theileri\nTrypanosoma conorhini\nTrypanosoma rangeli\nTrypanosoma cruzi marinkellei\nTrypanosoma cruzi cruzi\nTrypanosoma cruzi Brazil\nAngomonas deanei\nStrigomonas culicis\nPhytomonas sp. EM1\nPhytomonas sp. Hart1\nLeptomonas pyrrhocoris\nLeptomonas seymouri\nNovymonas esmeraldas\nPorcisia hertigi\nLeishmania martiniquensis\nLeishmania sp. Namibia\nLeishmania sp. Ghana\nLeishmania orientalis\nLeishmania tarentolae\nLeishmania mexicana\nLeishmani major\nLeishmania donovani\nLeishmania infantum\nLeishmania naiffi\nLeishmania lindenbergi\nLeishmania braziliensis\nLeishmania utingensis\nLeishmania shawi\nLeishmania panamensis\n0 5000 10,000 0 200 400 600\nSupport\n0.8 - 1\n0.5 - 0.79\n< 0.5\nTree scale : 0.5\nPie charts show No. \northogroups in each genus\nOrthogroups \nshared with other genera\nOrthogroups unique to genus\nOrthogroups\nin genome\nOrthogroups\nunique to genome\nFigure 5\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint \n\nL. donovani\nL. infantum\nL. major\nanacixe\nm .L\nsisneilizarb .L\nsisne\nmanap .L\nanahG .ps .L\naibi\nmaN .ps .L\nsilatneiro .L\nL. martiniquensis\nP. hertigi\nT. brucei brucei\nesneib\nmag iecurb .T\nizurc .T\nBodo B2\n01G odoB\nBodo H10\n8A odoB\nBodo A10\nBodo F10\nBodo B7\nB. saltans\nKEGG module completeness (%)\n0\n10\n20\n30\n40\n50\n60\n70\n80\n90\n100\nArginine biosynthesis, ornithine => arginine (M00844)\nProline biosynthesis, glutamate => proline (M00015)\nProline degradation, proline => glutamate (M00970)\nProline metabolism (M00972)\nTryptophan metabolism, tryptophan => kynurenine => 2-aminomuconate (M00038)\nTyrosine degradation, tyrosine => homogentisate (M00044)\nLeucine degradation, leucine => acetoacetate + acetyl-CoA (M00036)\nCysteine biosynthesis, serine => cysteine (M00021)\nMethionine degradation (M00035)\nMethionine salvage pathway (M00034)\nHistidine degradation, histidine => N-formiminoglutamate => glutamate (M00045)\nLysine degradation, lysine => saccharopine => acetoacetyl-CoA (M00032)\nGlutathione biosynthesis, glutamate => glutathione (M00118)\nGABA biosynthesis, eukaryotes, putrescine => GABA (M00135)\nPolyamine biosynthesis, arginine => agmatine => putrescine => spermidine (M00133)\nPolyamine biosynthesis, arginine => ornithine => putrescine (M00134)\nGlycine cleavage system (M00621)\nThreonine biosynthesis, aspartate => homoserine => threonine (M00018)\nC10-C20 isoprenoid biosynthesis, non-plant eukaryotes (M00367)\nC5 isoprenoid biosynthesis, mevalonate pathway (M00095)\nCitrate cycle (TCA cycle, Krebs cycle) (M00009)\nCitrate cycle, first carbon oxidation, oxaloacetate => 2-oxoglutarate (M00010)\nCitrate cycle, second carbon oxidation, 2-oxoglutarate => oxaloacetate (M00011)\nGluconeogenesis, oxaloacetate => fructose-6P (M00003)\nGlycolysis (Embden-Meyerhof pathway), glucose => pyruvate (M00001)\nGlycolysis, core module involving three-carbon compounds (M00002)\nPentose phosphate pathway (Pentose phosphate cycle) (M00004)\nPentose phosphate pathway, non-oxidative phase, fructose 6P => ribose 5P (M00007)\nPentose phosphate pathway, oxidative phase, glucose 6P => ribulose 5P (M00006)\nPRPP biosynthesis, ribose 5P => PRPP (M00005)\nPyruvate oxidation, pyruvate => acetyl-CoA (M00307)\nGalactose degradation, Leloir pathway, galactose => alpha-D-glucose-1P (M00632)\nGlyoxylate cycle (M00012)\nInositol phosphate metabolism, PI=> PIP2 => Ins(1,4,5)P3 => Ins(1,3,4,5)P4 (M00130)\nMalonate semialdehyde pathway, propanoyl-CoA => acetyl-CoA (M00013)\nNucleotide sugar biosynthesis, glucose => UDP-glucose (M00549)\nPropanoyl-CoA metabolism, propanoyl-CoA => succinyl-CoA (M00741)\nUDP-N-acetyl-D-glucosamine biosynthesis, eukaryotes, glucose => UDP-GlcNAc (M00892)\nUDP-N-acetyl-D-glucosamine biosynthesis, prokaryotes, glucose => UDP-GlcNAc (M00909)\nCytochrome bc1 complex respiratory unit (M00151)\nV-type ATPase, eukaryotes (M00160)\nC4-dicarboxylic acid cycle, NAD - malic enzyme type (M00171)\nC4-dicarboxylic acid cycle, NADP - malic enzyme type (M00172)\nC4-dicarboxylic acid cycle, phosphoenolpyruvate carboxykinase type (M00170)\nCAM (Crassulacean acid metabolism), dark (M00168)\nCAM (Crassulacean acid metabolism), light (M00169)\nReductive pentose phosphate cycle (Calvin cycle) (M00165)\nN-glycan precursor biosynthesis (M00055)\nbeta-Oxidation (M00087)\nbeta-Oxidation, acyl-CoA synthesis (M00086)\nFatty acid biosynthesis, elongation (M00083)\nFatty acid elongation in endoplasmic reticulum (M00415)\nAcylglycerol degradation (M00098)\nCeramide biosynthesis (M00094)\nKetone body biosynthesis, acetyl-CoA => acetoacetate/3-hydroxybutyrate/acetone (M00088)\nPhosphatidylcholine (PC) biosynthesis, PE => PC (M00091)\nPhosphatidylethanolamine (PE) biosynthesis, ethanolamine => PE (M00092)\nPhosphatidylethanolamine (PE) biosynthesis, PA => PS => PE (M00093)\nSphingosine biosynthesis (M00099)\nTriacylglycerol biosynthesis (M00089)\nCholesterol biosynthesis, FPP => cholesterol (M00101)\nErgocalciferol biosynthesis, FPP => ergosterol/ergocalciferol (M00102)\nC1-unit interconversion, eukaryotes (M00141)\nC1-unit interconversion, prokaryotes (M00140)\nCoenzyme A biosynthesis, pantothenate => CoA (M00120)\nLipoic acid biosynthesis, plants and bacteria, octanoyl-ACP => dihydrolipoyl-E2/H (M00881)\nMolybdenum cofactor biosynthesis, GTP => molybdenum cofactor (M00880)\nNAD biosynthesis, aspartate => quinolinate => NAD (M00115)\nNAD biosynthesis, tryptophan => quinolinate => NAD (M00912)\nRiboflavin biosynthesis, plants and bacteria, GTP => riboflavin/FMN/FAD (M00125)\nAdenine ribonucleotide biosynthesis, IMP => ADP,ATP (M00049)\nDeoxyribonucleotide biosynthesis, ADP/GDP/CDP/UDP => dATP/dGTP/dCTP/dUTP (M00053)\nGuanine ribonucleotide biosynthesis, IMP => GDP,GTP (M00050)\nDe novo pyrimidine biosynthesis, glutamine (+ PRPP) => UMP (M00051)\nPyrimidine degradation, uracil => beta-alanine, thymine => 3-aminoisobutanoate (M00046)\nPyrimidine deoxyribonucleotide biosynthesis, UDP => dTTP (M00938)\nPyrimidine ribonucleotide biosynthesis, UMP => UDP/UTP,CDP/CTP (M00052)\nAmino acid\nmetabolism\nCarbohydrate\nmetabolism\nEnergy\nmetabolism\nBiosynthesis of terpenoids and polyketides\nGlycan metabolism\nLipid\nmetabolism\nVitamine and\ncofactor metabolism\nNucleotide\nmetabolism\nFigure 6\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted October 3, 2025. ; https://doi.org/10.1101/2025.10.03.678719doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}