Rickettsial diseases are among the most common emerging neglected diseases in tropical areas and worldwide, posing the highest risks to those in poverty and marginalised communities. This prevalence is partly due to their arthropod vector-borne nature and the difficulty of diagnosing rickettsial diseases because of their nonspecific febrile clinical presentation. Considered one of the most common acute febrile illnesses after dengue and often underdiagnosed1,2,3, rickettsial diseases pose a significant disease burden worldwide, but current burden estimates are presumed to be inaccurate and understated due to limited data4.Historically delineated by mouse serotyping and antigenic profiling, the Rickettsiales as a whole have had a long-standing history of taxonomic and delineation inconsistencies. This ambiguity was addressed as early as 2001 by Dumler et al., playing a crucial role in the molecular reorganisation of Rickettsiales and Anaplasmataceae, whose work now serves as the basis of present-day taxonomy5,6. Recent whole-genome sequencing studies have revealed substantial ambiguity in current Rickettsia species definitions, raising questions about how species should be delineated in an obligate intracellular genus with limited phenotypic diversity7,8. At the same time, extensive historical antigenic profiling data, often regarded as obsolete with the advent of the genomic era, capture structured biological relationships that remain standard in serological assays such as IFA, the gold standard for Rickettsia diagnosis9,10.Recent years have seen renewed efforts to reconcile existing Rickettsia taxonomy with genomic methodologies; however, traditional bacterial species delineation guidelines yield drastic amendments to pre-existing taxonomy5,8. Two notable factors exacerbate divergence between pure WGS taxonomy and traditional taxonomy in Rickettsia: firstly, the disregard of crucial polyphasic characteristics such as geographical and phenotypic traits; secondly, the obligate intracellular nature of Rickettsia results in a smaller genome, reductive evolution and high synteny11. As such, current whole-genome sequencing-based classification of Rickettsia has exposed substantial taxonomic ambiguity that cannot be fully resolved by genomic metrics alone in an obligate intracellular genus with reduced phenotypic diversity.The historical antigenic profile remains an important facet of Rickettsiology and houses important phenotypic insights valued in the modern polyphasic species delineation approach. This review investigates historical serological evidence alongside contemporary genome-based taxonomy to assess their concordance and propose a polyphasic framework for resolving rickettsial classification and diagnostic challenges. Here, we synthesise historical antigenic and serological data with contemporary genome-based phylogeny, demonstrating that early antigenic groupings broadly correspond to many relationships recovered by modern WGS analyses, and argue that reintegrating these data is an important pillar for coherent taxonomy and rational serological diagnostic design.Overview of Rickettsia taxonomy and phylogenetic lineagesGeneral Rickettsia group classificationThe genus Rickettsia was named after Howard Taylor Ricketts for his 1906 research on the causative agent of Rocky Mountain spotted fever, Rickettsia rickettsii. Currently, Rickettsia bacteria are broadly classified as members of the genus Rickettsia, though colloquially the term “Rickettsia” encompasses the Order Rickettsiales, which includes a diverse group of obligate intracellular Gram-negative bacteria most commonly associated with arthropods such as ticks, lice, fleas, and mites. The Order Rickettsiales includes three main family lineages: Rickettsiaceae, Anaplasmataceae, and Candidatus Midichloriaceae. Outside of the genus Rickettsia, other genera found within Rickettsiales include Orientia, Wolbachia, Anaplasma, Ehrlichia, and Neorickettsia. Over 30 described Rickettsia spp. belong to the genus Rickettsia, alongside many candidate species and species known only as tick endosymbionts.As obligate intracellular arthropod-borne Gram-negative bacteria of the Order Rickettsiales, Rickettsia species can be classified into four primary groups: Spotted Fever Group (SFG), which includes Rickettsia conorii and R. rickettsii, Typhus Group (TG), which includes Rickettsia prowazekii and R. typhi, Transitional Group (TRG), which includes Rickettsia felis, and Ancestral Group (AG), which includes Rickettsia bellii (Fig. 1). Ongoing taxonomic confusion plagues the genus, particularly the question of what constitutes a Rickettsia species. Past serotyping was heavily utilised to characterise Rickettsia species before gradually phasing out as advances in genetic research came to prominence. However, serotyping and subsequent advances in Rickettsia serology may still help reconcile our understanding of Rickettsia taxonomy as we know it.Fig. 1Full size imageSchematic cladogram of Rickettsia spp. categorised into groups based on prior understanding of Rickettsia phylogeny12,13,14,108,109,110,111,112. TRG is considered part of SFG in much of the literature, particularly antigenically (Blue font indicates inconclusive, debated, or unconfirmed placement within a group or debated group status).Spotted fever group (SFG)The largest and best-known group within the genus Rickettsia is SFG, which includes most Rickettsia spp. such as R. rickettsii, and R. conorii, the causative agent of Mediterranean spotted fever (Table 1). Pathogenic SFG bacteria cause febrile rashes or spotted fever, hence their namesake. Within SFG, phylogenetic clades have been established using multigene approaches to infer phylogenetic relationships, initially relying on several well-known rickettsial markers, including 16S rRNA, htrA, gltA, ompA, ompB, sca4, sca1, and sca212. Later, whole-genome analysis was utilised instead for a more comprehensive phylogenetic overview13. To summarise these informal subgroups broadly, three have been described: the Rickettsia massiliae subgroup, the Rickettsia helvetica subgroup, and the R. rickettsii subgroup12,13,14. Additionally, the Rickettsia japonica subgroup (referred to as R. heilongjiangnensis subgroup in this literature) has been briefly noted, though this classification remains under debate12,13.Table 1 Representative list of validated and unvalidated SFG species without formal selection criteria to illustrate the vast number of species within the group. An asterisk (*) after the species name indicates a species with limited data.Full size tableTransitional group (TRG)A fourth subgroup within SFG is historically the Rickettsia akari subgroup, with notable initial members including R. felis and R. akari. Through genetic analysis of this subgroup, the affinity between R. felis, R. akari, and AG rickettsiae was first elucidated, based on the close relationship between the putative conjugative plasmid (pRF) in R. felis and AG Rickettsia genomes, supporting the subsequent proposal of TRG as a separate clade of Rickettsia12,15. However, some authors may still prefer to group TRG within SFG because their differences are mainly genetic, and they, by and large, retain antigenic similarities with SFG16,17. In either classification, the TRG is generally considered valid. An analysis in 2022 also proposed a new classification of SFG and TRG into SFGI and SFGII, respectively and placed R. helvetica basal to TG but not to SFG entirely11.Typhus group (TG)TG was initially separated from SFG based on antigenic profile, serological cross-reactivity, clinical differences, and vector species12,18,19. Although the TG phylogenetic group includes only two species, R. prowazekii and R. typhi, members of this group have been clinically important to humans for centuries, particularly R. prowazekii, which has had a long impact throughout human history as causing the plague epidemic typhus that follows in the wake of wars, famine, and poor living conditions20.Ancestral group (AG)Lastly, AG is the basal lineage of all Rickettsia members as the earliest rickettsiae to have diverged from SFG and TG. Rickettsia bellii is the first to be characterised as neither a member of SFG nor TG, but rather a member of a group that diverged before the separation of SFG and TG Rickettsiae, basal to both groups21. AG includes only two members: R. bellii and Rickettsia canadensis (formerly Rickettsia canada). Notably, El Karkouri et al. proposed R. canadensis and R. bellii be re-classified as separate Canadensis and Bellii groups, respectively, due to their basal placement on separate external phylogenetic branches11, which may shape future classification of AG moving forward.Historical and contemporary approaches to rickettsial classificationHistorical methods of rickettsial classificationSpecies in the genus Rickettsia were first classified into antigenic groups based on serology, particularly mouse serotyping, leading to the SFG and TG groups22,23. Subsequent increased awareness of phenotypic characteristics, including epidemiological, antigenic, and immunological properties, as well as genetic characteristics, led to the current four groups of TG, SFG, TRG, and AG8,12,15,21.Transition to molecular phylogeneticsBefore the widespread adoption of whole genome sequencing (WGS) for elucidation of phylogeny and species designation, classification of bacteria initially relied on DNA G + C content and DNA-DNA hybridisation (DDH) in the 1960s, followed by the widespread popularisation of 16S rRNA taxonomic classification, which was considered a gold standard in phylogenetic reconstructions24,25 for a long time. Currently, improvements in 16S rRNA classification are achieved by typing multiple conserved genes in Rickettsia and classifying species based on sequence similarity26,27. The most used scheme combines pan-bacterial 16S rRNA, gltA, and SFG rickettsial-specific OmpA (sca0), pan-rickettsial OmpB (sca5), and pan-rickettsial gene D (sca4), with the rationale that OmpA and OmpB are large, immunodominant surface proteins with substantial genetic variation between species28,29,30,31,32,33,34. Under this scheme, a putative Rickettsia candidate must have ≥ 98.1% homology with 16S rRNA and > 86.5% with gltA to an existing Rickettsia species. Generally, multiple gene schemes provide greater phylogenetic resolution than 16S rRNA alone, owing to the presence of multiple metabolic genes. Within the Rickettsia genus, both 16S rRNA and multiple-gene-derived phylogenies have been shown to strongly agree with the WGS phylogeny8,35. A ribosomal multilocus sequence typing (rMLST) scheme36 that uses 53 conserved ribosomal genes to distinguish between Rickettsia species has not been widely used.Whole genome sequencing and the taxonomic realignment of RickettsiaIntroduction of whole genome metrics and rickettsial complexityRecent literature utilising WGS5,8,35 provides further clarification on the Rickettsia phylogeny as well as argues for species designation revision, finding several taxonomic heterotypic synonyms and making the case that several species should instead be designated subspecies to other existing species, while others expand into Rickettsia genetic evolution and diversity, proposing new phylogenetic groupings11. We have summarised this observed relationship with a conceptual schematic in Fig. 2. The reason for this renewed scrutiny is the lack of an appropriate taxonomic system for Rickettsia and the lack of significant clinical and phenotypic data5. A noteworthy issue repeatedly raised in the existing WGS literature is the problematic and inconsistent nature of Rickettsia delineation and nomenclature, whilst Rickettsia monophyly and taxonomy are supported. In fact, the topic of rickettsial taxonomy criteria has long been contentious within the field5,7,37. Outside the field, large-scale analysis of bacterial whole genomes is used to define genus and species boundaries, and these automated classifications, based on thresholds across the whole bacterial taxonomy, are often inconsistent with historical classifications or with those defined by whole-genome sequence analysis within the Rickettsial genus5,35. Genomic data is currently available for 86 species or subspecies as recognised by NCBI, including eight candidatus species. A summary of public Rickettsia genome sequences deposited in NCBI Genbank is presented in Table S1.Fig. 2Full size imageConceptual schematic summarising broad phylogenetic and antigenic relationships among major Rickettsia groups discussed in this review. The figure integrates published whole-genome phylogenies and historical antigenic relationships. Branch lengths, node positions, and subgroup boundaries are illustrative only and do not represent evolutionary distance or a formal phylogenetic reconstruction. (Blue font indicates inconclusive, debated, or unconfirmed placement within a group.)Breakdown of current WGS-based taxonomic tools and thresholdsWGS-based taxonomic metrics rely on various methods to quantify the overall genome-relatedness index (OGRI). The most well-known and widely adopted bacterial species delineation methods are DDH/digital DDH (dDDH), average nucleotide identity (ANI), and genome blast distance phylogeny (GBDP)38. DDH traditionally measures genetic distance between two organisms and is used extensively in bacterial species delineation and phylogeny, with dDDH subsequently developed as a bioinformatic in silico analogue39, whereas ANI measures nucleotide similarity between genomes, with its later derivative orthoANI, which modifies ANI to use orthologous fragments and be reciprocally consistent40. Together, dDDH and ANI are typically used to delineate bacterial species. Other metrics, such as average amino acid identity (AAI) and core genome alignment sequence identity (CGASI), have also been implemented. CGASI uses the sequence identity of a concatenated core genome alignment to infer phylogenetic relationships. AAI determines genomic relatedness through amino acid identity and is especially preferred beyond species rank, as methods such as ANI are less useful for more distantly related genomes.Rickettsia-specific WGS threshold revisions and species merger proposalsMajor taxonomic resources used in bacterial genetics do not match the species groupings used in rickettsial research. The National Center for Biotechnology Information (NCBI) Taxonomy database is used to classify nucleic acids submitted to any International Nucleotide Sequence Database Collaboration (INSDC) database. Under the NCBI taxonomy4142, all TRG species are grouped into the SFG, R. raoultii and R. heilongjiangensis are subspecies of R. conorii, and R. buchneri is a subspecies of R. tamurae. The Genome Taxonomy Database43, designed to provide a standardised taxonomy for improved classification of novel bacterial species, uses a 95% ANI cutoff, which reclassifies most SFG species as R. rickettsii.Genus-specific analyses have proposed alternative criteria for species delineation in Rickettsia (Table 2). Diop et al.8. proposed a modified criterion of both ANI and dDDH, as the current ANI of ≥95–96% and a dDDH value of ≥70% is inapplicable to Rickettsia spp. based on Rickettsia’s few phenotypic properties and low genetic heterogeneity, coupled with its obligate intracellular lifestyle. Doing so would merge several species into a single species, such as R. canadensis with R. bellii, R. typhi with R. prowazekii, and the entire TRG. Chung et al.5. also demonstrated that the equivalent CGASI will classify the majority of SFG as the same species, except for R. monacensis, which is classified into a single species. Hördt et al.35 utilised a dDDH cutoff of 70%, and the same issue arises: R. gravesii, R. heilongjiangensis, R. japonica, R. sibirica, and R. slovaca are classified as R. conorii subspecies, and R. buchneri is classified as an R. tamurae species. As such, Diop et al. proposed a new species demarcation criterion of >92.3% dDDH and >99.19% orthoANI, respectively. Under these revised Rickettsia-specific criteria, only R. heilongjiangensis and R. japonica are to be assigned together at 92.4% dDDH according to Hördt et al. According to Diop et al., R. argasii should be reassigned to R. heilongjiangensis, and R. rickettsii subsp. californica (formerly R. philipii; previously known as Rickettsia sp. 364D) should be reassigned to R. rickettsii. Additionally, other Rickettsia, such as R. sibirica and R. parkeri, with notably high dDDH and orthoANI values of 92.3% and 99.19%, respectively, are observed and serve as the dDDH and orthoANI cutoffs in the paper. There is also the possibility that this criterion could paradoxically be too lax, as Chung et al. note in their own CGASI calculation: the genus-species delineation for Rickettsia is currently problematic and should be excluded from the analysis. In which case, Diop et al.’s decision to base species demarcation cutoffs on the current list of validly published species as the gold standard serves as a pragmatic solution for the time being. However, using high dDDH and orthoANI cutoffs chosen to separate R. sibirica and R. parkeri raises the question of whether they are sufficiently distantly related to constitute separate, distinct species. Across all three WGS datasets, however, a consistent phylogenetic tree pattern is observed with consistent groupings, especially in SFG (Figs. 3 and 4, Supplementary Table S2).Table 2 Table summarising current WGS-based taxonomic thresholds in Rickettsia species delineation.Full size tableFig. 3Full size imageA heatmap showing serological cross-reactivity between different Rickettsia species, grouped by our proposed clade and subgroup as shown in Fig. 2. Data taken from papers cited in this review, alongside additional data104,113,114,115,116,117,118,119,120,121,122,123,124,125. “No data” is shown in grey where no serological cross-reactivity data source was available.Fig. 4Full size imageA heatmap showing average nucleotide identity (ANI) between different Rickettsia species, grouped by our proposed clade and subgroup as shown in Fig. 2. Reference genomes for each strain are listed in Table S2, and ANI was calculated using FastANI, using the matrix values for each comparison.Antigenic structure and group identity of Rickettsia spp.The taxonomic uncertainty revealed by WGS raises the question of whether historical antigenic data, used to define rickettsial groups long before genomic tools were available, capture biologically meaningful relationships. While WGS-based taxonomic techniques are powerful and discriminatory, current species characterisation emphasises a comprehensive polyphasic approach, considering not only genetic similarity but also bacterial characteristics such as geographical, chemotaxonomic, and phenotypic traits4445. One such important characterisation, specifically within the genus Rickettsia, with a long historical basis, is the antigenic and serological profile, which was used to initially group TG and SFG2223. The antigens of TG and SFG were identified early on, starting in the early 70s, particularly the large, immunodominant surface proteins pan-Rickettsia OmpB and SFG, and the TRG-specific OmpA, which together make up the main Rickettsial S-layer 46,47,48,49,50,51. Congruent with phylogenetic taxonomy, the cell surface antigenic composition of SFG and TRG is the most closely related, with TG diverging before the split of TRG and SFG, hence lesser antigen identity, and lastly AG having split off in the early stages of rickettsial evolution, thus the reason for their lack of shared antigenic profile and non-pathogenicity. The primary proteins involved in immunogenicity, virulence, and infection belong to the Sca family and are found in most Rickettsia spp. such as Sca1, Sca2, Sca3, and Sca452,53,54. Other surface-exposed proteins with reports of antigenicity include GroEL5255,56,57,58, GroES58, Adr1 and Adr2, OmpW, Porin-458,59,60, while other proteins with disputed or unknown antigenicity include TolC, TRG and SFG-specific RickA566162. Apart from intra-genus antigenic composition variation, intra-species variation is also observed in strains and subspecies with different virulence, where specific epitopes are unrecognisable by monoclonal antibodies, suggesting that, despite shared antigenic composition, different strains and subspecies of the same species may not be antigenically identical and may be structurally different63. In which case, empirical confirmation is crucial for determining the antigenic profile shared between species, given new taxonomic resolution and existing antigenic or serologic research. This process is necessary to reconcile our previous serologic understanding with modern genomic data and taxonomic classification.Serological diagnosis and genomic correlationHistorical overview of serological diagnosis in rickettsiaeHistorically, the now-obsolete Weil-Felix test was the first rickettsial diagnostic technique used to diagnose typhus fever in 1916, utilising antigenic cross-reactivity between Proteus and Rickettsia species64. As diagnostic methods for rickettsial infections improved, techniques such as complement fixation and microagglutination replaced the Weil-Felix test965. With the advent of modern serological diagnosis, the immunofluorescence assay (IFA) emerged as the gold standard for rickettsial diagnosis due to its high accuracy910. Standard practice is to compare convalescent and acute-phase samples to detect a fourfold rise in antibody titer and confirm positivity66,67,68,69. Other serological techniques are still used in clinical settings, including enzyme-linked immunosorbent assay (ELISA), indirect immunoperoxidase (IIP), indirect hemagglutination (IHA), microagglutination (MA), and microimmunofluorescence (MIF).Antigenic profiling and early rickettsial classificationSerotyping, antigenic profiling, and cross-reactivity determination were largely conducted using MIF, IHA, MA, and IFA, which were important components of historical rickettsial classification. Early on, TG and SFG were classified using such methods, with a clear divide in antigenic distance, which forms the basis of modern group-based serological diagnosis2270. A large part of rickettsial serology is the widespread, strong in-group cross-reactivity and demonstrable cross-reactivity across groups, particularly between TRG and SFG, TG and SFG, and within SFG71,72,73,74,75,76,77. As such, species-level identification is only possible with specialised techniques available only in a reference laboratory910.Genomic-antigenic correlations within the spotted fever group (SFG)Within the spotted fever group, historical antigenic profiling shows strong concordance with modern WGS-based phylogenetic clustering. A side-by-side comparison between past antigenic relatedness and our current understanding of WGS rickettsial taxonomy provides a comprehensive overview that combines both fields into a cohesive big picture. Congruent with established WGS-based taxonomy, antigenic profile and serologic cross-reactivity show clear TG, TRG, and SFG separation, with SFG demonstrating clusters of distinct subgroups. The core SFG R. rickettsii subgroup is characterised by extensive antigenic similarity and cross-reactivity, including R. rickettsii, R. conorii, R. africae, R. slovaca, R. parkeri, and R. sibirica (Table 3). R. sibirica and R. parkeri are antigenically very similar, with a specificity difference (SPD) of 3.5, where