Origin of mitochondria

All analyses agree that mitochondria arose from within (or next to) the Alphaproteobacteria. Where exactly is contested, because the mitochondrial lineage — like several candidate relatives (Rickettsiales, SAR11) — has a fast-evolving, AT-rich, reduced genome, and such lineages tend to attract each other in phylogenies under simple models. The literature tab summarises the hypotheses and what each study did; the de novo tab shows a tree that is rebuilt automatically from current public data.

Hypotheses

Mitochondria sister to all (known) Alphaproteobacteria

When compositional heterogeneity is addressed — by removing/recoding compositionally biased sites, by site-heterogeneous mixture models (LG+C60, CAT-GTR, PMSF), and by adding metagenome-assembled genomes of uncultured marine alphaproteobacteria — mitochondria no longer nest inside Alphaproteobacteria but branch as the sister lineage of all sampled alphaproteobacteria (Martijn et al. 2018), or of all known Alphaproteobacteria including newer lineages (Muñoz-Gómez et al. 2022). The Rickettsiales-sister signal is interpreted as an artefact of shared AT-rich, fast-evolving genomes.

data & models: Martijn 2018: 24 mito-encoded proteins, ~100 taxa incl. 12 marine MAGs (Alpha-II, MarineAlpha), CAT-GTR + SR4/D6 recoding, compositionally biased sites removed. Muñoz-Gómez 2022: 108 markers, expanded taxon sampling incl. Alpha-II/MarineAlpha MAGs and novel free-living lineages, LG+C60+F+G, CAT-GTR, site-and-branch-heterogeneous (GHOST/heterotachy-aware) analyses.
status: Favoured by the most model-rich analyses to date; contested by Fan et al. 2020/2022 (see 'within Alphaproteobacteria').
key studies: Martijn et al. 2018; Muñoz-Gómez et al. 2022; Muñoz-Gómez et al. 2022; Roger et al. 2017

Loading…

Cladogram (branch lengths not meaningful). Orange nodes carry notes — hover them. Alphaproteobacterial backbone simplified after Muñoz-Gómez et al. 2019.

Mitochondria sister to Rickettsiales

The classical view: mitochondria are the closest relatives of the obligately intracellular Rickettsiales (Rickettsia, Anaplasma, Wolbachia…). First suggested from rRNA and later supported by the Rickettsia prowazekii genome and many concatenated-protein phylogenies. Rickettsiales share with mitochondria a reduced, AT-rich, fast-evolving genome — exactly the properties that make long-branch attraction and compositional bias a concern.

data & models: Andersson 1998: R. prowazekii genome, individual gene trees; Fitzpatrick 2006: genome-scale supertree/concatenation of α-proteobacterial genomes; Wang & Wu 2015: 29 slow-evolving proteins, site-heterogeneous CAT model, Rickettsiales-sister recovered.
status: Long-standing majority view; support tends to weaken with site-heterogeneous models, recoding and richer taxon sampling.
key studies: Andersson et al. 1998; Fitzpatrick et al. 2006; Wang & Wu 2015; Sassera et al. 2011

Loading…

Cladogram (branch lengths not meaningful). Orange nodes carry notes — hover them. Alphaproteobacterial backbone simplified after Muñoz-Gómez et al. 2019.

Mitochondria within Alphaproteobacteria, with Rickettsiales + marine Alpha-II lineages

Fan et al. (2020) argued that site-exclusion (compositional stripping) yields erratic results and instead reduced noise by systematic taxon sampling. Their trees place mitochondria robustly within Alphaproteobacteria, sharing an ancient common ancestor with Rickettsiales and the currently unclassified marine lineages (Alpha-II / MarineAlpha MAGs). Muñoz-Gómez et al. (2022) replied that this affiliation is itself an artefact of model misspecification; Fan et al. replied in turn.

data & models: Fan 2020: 3,000+ alphaproteobacterial genomes screened, systematic subsampling of taxa, cross-validation over tree series; homogeneous and mixture models without site exclusion. Geiger 2023: metabolic/gene-content approaches favouring an ancestry among free-living, aerobic Alphaproteobacteria.
status: Actively debated; the position of mitochondria relative to Rickettsiales and Alpha-II lineages remains the crux.
key studies: Fan et al. 2020; Muñoz-Gómez et al. 2022; Fan et al. 2022; Geiger et al. 2023

Loading…

Cladogram (branch lengths not meaningful). Orange nodes carry notes — hover them. Alphaproteobacterial backbone simplified after Muñoz-Gómez et al. 2019.

Mitochondria sister to SAR11 (Pelagibacterales) — historical

Thrash et al. (2011) recovered mitochondria as sister to the abundant, streamlined marine SAR11 clade, suggesting a free-living, oligotrophic ancestor. Reanalyses showed the grouping to be a compositional/long-branch artefact: SAR11 and mitochondria (and Rickettsiales) share extremely AT-rich genomes; with better models and site stripping SAR11 falls elsewhere within Alphaproteobacteria and the SAR11–mitochondria clade disappears.

data & models: Thrash 2011: concatenated protein set incl. newly sequenced SAR11 genomes, homogeneous models. Rodríguez-Ezpeleta & Embley 2012 and Viklund 2012: CAT model, recoding, removal of fast/biased sites → SAR11 not related to mitochondria.
status: Refuted; kept here as a textbook example of compositional attraction.
key studies: Thrash et al. 2011; Rodríguez-Ezpeleta & Embley 2012; Viklund et al. 2012

Loading…

Cladogram (branch lengths not meaningful). Orange nodes carry notes — hover them. Alphaproteobacterial backbone simplified after Muñoz-Gómez et al. 2019.

What differs between studies

Study	Data	Model / treatment	Taxon sampling	Result
Andersson et al. 1998	R. prowazekii genome; single-gene and small concatenations	ML/NJ, homogeneous	few α-proteobacteria	Rickettsiales-sister
Fitzpatrick et al. 2006	genome-scale supertree + concatenation	ML, homogeneous	~50 α-proteobacterial genomes	Rickettsiales-sister
Thrash et al. 2011	concatenated proteins	ML/Bayesian, homogeneous	adds SAR11 genomes	SAR11-sister
Rodríguez-Ezpeleta & Embley 2012	reanalyses	CAT, recoding, site removal	as above	SAR11-sister rejected
Wang & Wu 2015	29 slow-evolving proteins	CAT-GTR	broad α-proteobacteria	Rickettsiales-sister
Martijn et al. 2018	24 mito-encoded proteins	CAT-GTR + SR4/D6 recoding, biased-site removal	+12 marine MAGs (Alpha-II, MarineAlpha)	sister to all sampled Alphaproteobacteria
Fan et al. 2020	conserved proteins, no site exclusion	ML incl. mixture models; taxon-subsampling series	systematic sampling from >3,000 genomes	within Alphaproteobacteria, with Rickettsiales + Alpha-II
Muñoz-Gómez et al. 2022	108 markers	LG+C60, CAT-GTR, PMSF, heterotachy-aware	expanded incl. novel free-living lineages + MAGs	sister to all known Alphaproteobacteria
Geiger et al. 2023	metabolic gene content	presence/absence, gene phylogenies	α-proteobacteria + MAGs	ancestry among free-living aerobic α-proteobacteria

Why the answer depends on the model

Mitochondrial, rickettsial and SAR11 genomes are all AT-rich, and their proteins are correspondingly enriched in amino acids encoded by AT-rich codons (F, I, K, N, Y). Site-homogeneous models (LG, WAG, JTT + Γ) assume the same amino-acid frequencies at every site and cannot distinguish shared composition from shared ancestry, so they tend to group these lineages together (long-branch attraction plus compositional attraction). Remedies used in the literature include site-heterogeneous mixture models (CAT-GTR, LG+C20–C60, PMSF), recoding amino acids into fewer classes (SR4, D6), removing the most compositionally heterogeneous sites, taxon sampling that breaks long branches (adding uncultured marine Alpha-II/MarineAlpha MAGs and free-living early-branching lineages) and heterotachy-aware branch-heterogeneous models. Different combinations lead to the different answers above; a fair reading is that the position of mitochondria at the base of, or just inside, Alphaproteobacteria remains genuinely open, with the most model-rich analyses currently favouring a sister-group relationship to all known Alphaproteobacteria (Muñoz-Gómez et al. 2022; review: Roger et al. 2017). Genome content and metabolism argue for a free-living, aerobic ancestor rather than a parasite like Rickettsia (Sassera et al. 2011; Geiger et al. 2023); host association evolved independently within Rickettsiales (Schön et al. 2022).

References

Andersson SGE, Zomorodipour A, Andersson JO, Sicheritz-Pontén T, Alsmark UCM, Podowski RM, Näslund AK, Eriksson A-S, Winkler HH, Kurland CG (1998). The genome sequence of Rickettsia prowazekii and the origin of mitochondria. Nature 396: 133–140. doi:10.1038/24094

Fitzpatrick DA, Creevey CJ, McInerney JO (2006). Genome phylogenies indicate a meaningful α-proteobacterial phylogeny and support a grouping of the mitochondria with the Rickettsiales. Molecular Biology and Evolution 23: 74–85. doi:10.1093/molbev/msj009

Esser C, Ahmadinejad N, Wiegand C, Rotte C, Sebastiani F, Gelius-Dietrich G, Henze K, Kretschmann E, Richly E, Leister D, Bryant D, Steel MA, Lockhart PJ, Penny D, Martin W (2004). A genome phylogeny for mitochondria among α-proteobacteria and a predominantly eubacterial ancestry of yeast nuclear genes. Molecular Biology and Evolution 21: 1643–1660. doi:10.1093/molbev/msh160

Thrash JC, Boyd A, Huggett MJ, Grote J, Carini P, Yoder RJ, Robbertse B, Spatafora JW, Rappé MS, Giovannoni SJ (2011). Phylogenomic evidence for a common ancestor of mitochondria and the SAR11 clade. Scientific Reports 1: 13. doi:10.1038/srep00013

Rodríguez-Ezpeleta N, Embley TM (2012). The SAR11 group of alpha-proteobacteria is not related to the origin of mitochondria. PLoS ONE 7: e30520. doi:10.1371/journal.pone.0030520

Viklund J, Ettema TJG, Andersson SGE (2012). Independent genome reduction and phylogenetic reclassification of the oceanic SAR11 clade. Molecular Biology and Evolution 29: 599–615. doi:10.1093/molbev/msr203

Sassera D, Lo N, Epis S, D'Auria G, Montagna M, Comandatore F, Horner D, Peretó J, Luciano AM, Franciosi F, Ferri E, Crotti E, Bazzocchi C, Daffonchio D, Sacchi L, Moya A, Latorre A, Bandi C (2011). Phylogenomic evidence for the presence of a flagellum and cbb3 oxidase in the free-living mitochondrial ancestor. Molecular Biology and Evolution 28: 3285–3296. doi:10.1093/molbev/msr159

Wang Z, Wu M (2015). An integrated phylogenomic approach toward pinpointing the origin of mitochondria. Scientific Reports 5: 7949. doi:10.1038/srep07949

Martijn J, Vosseberg J, Guy L, Offre P, Ettema TJG (2018). Deep mitochondrial origin outside the sampled alphaproteobacteria. Nature 557: 101–105. doi:10.1038/s41586-018-0059-5

Muñoz-Gómez SA, Hess S, Burger G, Lang BF, Susko E, Slamovits CH, Roger AJ (2019). An updated phylogeny of the Alphaproteobacteria reveals that the parasitic Rickettsiales and Holosporales have independent origins. eLife 8: e42535. doi:10.7554/eLife.42535

Fan L, Wu D, Goremykin V, Xiao J, Xu Y, Garg S, Zhang C, Martin WF, Zhu R (2020). Phylogenetic analyses with systematic taxon sampling show that mitochondria branch within Alphaproteobacteria. Nature Ecology & Evolution 4: 1213–1219. doi:10.1038/s41559-020-1239-x

Muñoz-Gómez SA, Susko E, Williamson K, Eme L, Slamovits CH, Moreira D, López-García P, Roger AJ (2022). Site-and-branch-heterogeneous analyses of an expanded dataset favour mitochondria as sister to known Alphaproteobacteria. Nature Ecology & Evolution 6: 253–262. doi:10.1038/s41559-021-01638-2

Muñoz-Gómez SA, Susko E, Williamson K, Eme L, Slamovits CH, Moreira D, López-García P, Roger AJ (2022). Phylogenetic affiliation of mitochondria with Alpha-II and Rickettsiales is an artefact. Nature Ecology & Evolution 6: 1829–1831. doi:10.1038/s41559-022-01871-3

Fan L, Wu D, Goremykin V, Trost K, Martin WF, Zhu R (2022). Reply to: Phylogenetic affiliation of mitochondria with Alpha-II and Rickettsiales is an artefact. Nature Ecology & Evolution 6: 1832–1835. doi:10.1038/s41559-022-01896-8

Schön ME, Martijn J, Vosseberg J, Köstlbacher S, Ettema TJG (2022). The evolutionary origin of host association in the Rickettsiales. Nature Microbiology 7: 1189–1199. doi:10.1038/s41564-022-01169-x

Geiger O, Sanchez-Flores A, Padilla-Gomez J, Degli Esposti M (2023). Multiple approaches of cellular metabolism define the bacterial ancestry of mitochondria. Science Advances 9: eadh0066. doi:10.1126/sciadv.adh0066

Roger AJ, Muñoz-Gómez SA, Kamikawa R (2017). The origin and diversification of mitochondria. Current Biology 27: R1177–R1192. doi:10.1016/j.cub.2017.09.015

Automated concatenated-protein tree

Current tree

103 taxa

10,199 aligned sites · 35 markers

method: IQ-TREE LG+F+G4
support: SH-aLRT
log-likelihood: -821,164.9
rooted on: outgroup (Betaproteobacteria, Gammaproteobacteria, Bacteria)
run: 2026-08-16 · profile ci · 2 min

Where do mitochondria branch?

Mitochondria (n = 42) are monophyletic (support 100). Sister group: Rickettsiales (8 genomes); support at the joining node 98.9.

Interpretation is computed automatically from the tree (see Methods) and updates with every run.

Downloads

tree.nwk (rooted Newick, labels = accessions) · tree.json (with metadata)
concat.faa.gz · partitions.nex · occupancy.tsv
IQ-TREE report · manifest.json

Read with care. This automated tree uses a single concatenated matrix (35 mitochondrion-encoded proteins and their bacterial homologs) under a site-homogeneous LG+F+G4 model with IQ-TREE's fast search, without recoding, site stripping or metagenome-derived taxa. That is exactly the setting under which mitochondria are expected to be attracted to Rickettsiales (see the Literature tab). It is a reproducible, always-current baseline — not the last word. A deeper analysis (LG+C20+F+G4 with PMSF, full search, UFBoot) can be run locally with the deep profile — see below.

Loading tree…

Tip colours follow the taxonomic groups in the legend; mitochondrial genomes are labelled with the eukaryote species. Node dots: filled black ≥ 95, grey ≥ 80, hollow < 80 (SH-aLRT); node labels show the support values. Click nodes to collapse, legend entries to collapse groups, tips to open the NCBI record.

Marker occupancy

Which of the 35 markers each taxon contributes (103 taxa kept with ≥ 4 markers; markers kept if present in ≥ 40% of taxa). Gene-rich protist mtDNAs (jakobids, Malawimonas, red algae, cryptophytes, stramenopiles) contribute ribosomal proteins and complex I subunits absent from animal/fungal mtDNAs.

Deep analysis (manual, site-heterogeneous model)

Deep tree

103 taxa

10,206 aligned sites · 35 markers

method: IQ-TREE LG+C20+F+G4
support: SH-aLRT / UFBoot
log-likelihood: -785,123.9
rooted on: outgroup (Betaproteobacteria, Gammaproteobacteria, Bacteria)
run: 2026-08-15 · profile deep · 157 min

Where do mitochondria branch?

Monophyletic: True. Sister group: Rickettsiales (8); support SH-aLRT 93.6 / UFBoot 96.

Loading tree…