Characterization and Association of Rips Repertoire to Host Range of Novel Ralstonia solanacearum Strains by In Silico Approaches

Ralstonia solanacearum species complex (RSSC) cause several phytobacteriosis in many economically important crops around the globe, especially in the tropics. In Brazil, phylotypes I and II cause bacterial wilt (BW) and are indistinguishable by classical microbiological and phytopathological methods, while Moko disease is caused only by phylotype II strains. Type III effectors of RSSC (Rips) are key molecular actors regarding pathogenesis and are associated with specificity to some hosts. In this study, we sequenced and characterized 14 newly RSSC isolates from Brazil’s Northern and Northeastern regions, including BW and Moko ecotypes. Virulence and resistance sequences were annotated, and the Rips repertoire was predicted. Confirming previous studies, RSSC pangenome is open as α≅0.77. Genomic information regarding these isolates matches those for R. solanacearum in NCBI. All of them fit in phylotype II with a similarity above 96%, with five isolates in phylotype IIB and nine in phylotype IIA. Almost all R. solanacearum genomes in NCBI are actually from other species in RSSC. Rips repertoire of Moko IIB was more homogeneous, except for isolate B4, which presented ten non-shared Rips. Rips repertoire of phylotype IIA was more diverse in both Moko and BW, with 43 common shared Rips among all 14 isolates. New BW isolates shared more Rips with Moko IIA and Moko IIB than with other public BW genome isolates from Brazil. Rips not shared with other isolates might contribute to individual virulence, but commonly shared Rips are good avirulence candidates. The high number of Rips shared by new Moko and BW isolates suggests they are actually Moko isolates infecting solanaceous hosts. Finally, infection assays and Rips expression on different hosts are needed to better elucidate the association between Rips repertoire and host specificities.


Introduction
Per year, around 20% of yield losses are due to infection by soil borne microbes [1,2]. Bacterial wilt, caused by Ralstonia solanacearum species complex (RSSC), is a cosmopolitan phytobacteriosis of difficult management and control in the field. It is responsible for significant yield losses in many crops in tropical regions and worldwide, affecting potato, tomato, eggplant, peppers, banana, eucalyptus, and ginger, among others [1, [3][4][5].
In the state of Pernambuco, Brazil, this bacterium was detected in all mesoregions, being responsible for total loss in crops where the disease was identified [6,7]. In the last decade, many phylogenetic studies proposed the reclassification of RSSC into three distinct species according to their phylotype position and center of origins: R. pseudosolanacearum (phylotypes I and III, from Asia and Africa), R. solanacearum (phylotypes IIA and IIB, from America), and R. syzygii (phylotype IV, from Indonesia) [8][9][10]. Due to the broad range of hosts of RSSC, they're commonly described in ecotypes according to the infected host and disease caused. All species in RSSC cause bacterial wilt (BW). Moko disease of Musa, brown rot of potato, and non-pathogenic to banana (NPB) are caused only by R. solanacearum, with brown rot and NPB being positioned in IIB as recent R. solanacearum strains derived from Moko. In turn, Sumatra disease of clove and blood disease bacterium (BDB) are caused by R. syzygii [8,11]. In Brazil, no occurrences of R. syzygii have been reported, whereas R. pseudosolanacearum and R. solanacearum are pointed out as being responsible for all BW cases [11]. Furthermore, Moko disease is highly prevalent in Latin America, considered an A2-level quarantine disease in the Northern (Amazonas, Pará, Rondônia, and Roraima) and Northeastern (Pernambuco and Sergipe) regions [12]. However, no phenotypic characteristics or symptoms displayed by infected plants enable distinguishing RSSC in phylotypes.
A relevant molecular mechanism in RSSC related to pathogenicity and virulence is their type III effectors (T3Es), commonly referred to as Ralstonia injected proteins (Rips). Those proteins are essential in pathogenicity success because they interfere with plant basal immunity and act on specific targets within cascade reactions in the cell, eliciting or attenuating hypersensitive responses [13]. Effectors triggering hypersensitive responses are related to avirulence traits, while effectors eliciting immune responses are related to virulence traits. Moreover, a series of studies has been carried out to identify the role of those effectors in host infection success. For instance, the absence of RipAA and RipP1 are linked to infection success in tobacco [14], RipS1 is linked to virulence contribution in African daisy and eggplant [15,16], and RipAZ1 is linked to avirulence in black nightshade plants [17]. Only 16% of the currently known Rips subfamilies compose the core effectome of RSSC [13,18]. Therefore, Rips repertoire tends to vary significantly with isolates' phylotype, ecotype, and local area of occurrence.
In this study, we aimed to apply in silico approaches to predict the pangenome and identify the exact taxonomy of RSSC isolates causing Moko disease and BW in Brazil's Northern and Northeastern regions, as well as investigate resistance and virulence genes, predict their Rips repertoire, and compare with previously identified Rip candidates related to host specificity.

Genomes Database
In total, 120 complete genomes were used for this study, 118 of R. solanacearum and 2 genomes of R. pseudosolanacearum; 106 of them were retrieved from the public genome repository of the National Center for Biotechnology (NCBI) and 14 unpublished private genomes were isolated from the Northern and Northeastern regions of Brazil, from which 12 cause Moko disease and 2 cause BW in tomato. These 14 isolates were previously sequenced on the Illumina Hi-Seq 2500 platform in a paired-end library of 2 × 150 bp at the University of Göttingen (Germany).

Quality Control and Assembly
The sequencing quality was estimated using the FastQC (v0.11.8) metrics report [19]. Subsequently, we used SPAdes (v3.14) [20] for genome assembly with default parameters. For genome completeness verification and assembly parameters, assembled genomes underwent evaluation through BUSCO (v4.1.2) (Benchmarking Universal Single-Copy Orthologs) [21] against the bacteria_odb10 database. Assemblies with completeness below 90% were discarded. In parallel, we also used QUAST (v5.2.0) (Quality Assessment Tool for Genome Assemblies) [22] with default parameters. The information regarding the strain name, collection site, host, and disease caused for all genomes used is available in Table A1.

Prediction of the RSSC Pangenome
To identify clusters of similar genomes, the RSSC pangenome was estimated with Roary (v3.13.0) [29] with subsequent visualization of the matrix and phylogenetic tree on Phandango (https://jameshadfield.github.io/phandango/#/main, accessed on 6 June 2022) [30] and Roary's built-in R script. A phylogenomic tree inferred on single-copy orthologs was plotted using OrthoFinder (v2.5.4) and iTOL (v6.0) [31,32]. The pangenome's alpha value was calculated with an in-house script using OrthoFinder's outputs with a formula based on the Heap's Law model, in which α < 1 indicates an open pangenome [33,34]: In which: n = number of genes, N = number of genomes, and k and γ are constants defined to fit the specific curve. Following, γ can be calculated as: By that, α < 1 indicates an open pangenome, in which the more genomes are sequenced and added to the analysis, the more genes will be discovered. On the other hand, α > 1 indicates a closed pangenome, meaning despite more genomes being added, no significant increase in new genes would be discovered. Additionally, we also used the Least Squares Fit Principle to predict the number of singletons added to each genome and a probable number of genes for core genome stabilization, following: In which: n = number of genes, x = number of genomes, e is the Euler number, and k, τand tgΘ are constants.

Whole-Genome Methods for Taxonomy Insights
Furthermore, all genomes underwent two distinct approaches for species classification: first, an Average Nucleotide Identity (ANI) analysis was conducted through the MUMmer alignment method using pyANI (v3.0) [35], considering a ≥96% similarity criteria for different genomes belonging to the same species. Afterward, an in silico DDH (DNA-DNA hybridization) analysis was performed in the GGDC web server (v3.0) [36,37] (https: //ggdc.dsmz.de/ggdc.php#, accessed on 30 June 2022), with subsequent visualization in Morpheus (https://software.broadinstitute.org/morpheus/, accessed on 8 January 2022). Due to the limitation in the number of genomes allowed in GGDC web server, we only used 77 of 120 genomes from the database, including the 14 newly sequenced ones. A ≥70% similarity criterion was considered for different genomes belonging to the same species, and ≥79-80% similarity criterion for subspecies classification [38].

Genome Sequencing and Characterization of New Brazilian RSSC Isolates
The QUAST report revealed that the isolates B4 and CCRMRs121 had the largest and smallest genomes, with 5,858,492 and 5,364,378 base pairs, respectively. All genome sizes were similar to the average genome size for R. solanacearum at NCBI (5,059,182 bp). Other quality metrics for all isolates are available in Table 1. Furthermore, the BUSCO report revealed all genomes were complete considering single-copy orthologous genes. None of the assembled genomes had missing genes from the database. However, most had at least one fragmented or duplicated gene, which did not interfere with further analyses (Figure 1). Genomes presented between 5034 and 4592 coding sequences and 61 to 66 non-coding RNAs (including tRNAs, rRNAs, tmRNAs, and others). According to NCBI, the average count of coding sequences in R. solanacearum is 4774. No direct correlation was found between CDS-ncRNA amounts and the type of disease caused by each isolate. Overall, 19 virulence genes (adeG, cheA, cheB, cheD, cheW, cheY, cyaB, flgG, fliA, fliM, fliP, htpB, icl, katG, motA, pilT, sodB, tsr, tuf A, and motA) and 29 metal resistance genes (adeB, adeG, bcrC, chrA1, chrB1, chrF, cnrA, cnrT, copA, copR, cueA, czcA, dpsA, mdtB, mdtB/yegN, merA, merP, merR, merT, mexK, oprJ, pbrA, pstA, pstB, pstC, rcnA/yohM, ruvB, silA, and smrA) were predicted for the isolates. All predicted genes had a similarity of ≥60%. The most abundant genes were adeG, cheY, htpB, and pilT, present in 118 of 120 genomes. Similarly, for metal resistance genes, the most abundant were adeG, bcrC, czcA, dpsA, pstB, and ruvB. Strains UW386, T110, and SL3022, isolated from soil and infected potato, exhibited the most quantity of metal resistance genes, with 17 and 16, respectively. In parallel, strains T110 and SL3730, exhibited the highest amounts of virulence genes, although all genomes presented 10 to 7 virulence genes. The newly sequenced genomes presented 9 to 7 virulence genes and 10 to 8 metal resistance genes. Neither host/source nor isolate origins seemed to have a correlation to genetic virulence of resistance. The clustermaps for VFDB and BacMet are available in Figures 2 and 3.  By that, we predict that at each new genome added to the RSSC pangenome, ∼ =0.1905 new ortholog genes would be discovered, and the core genome would stabilize in around 1562 genes. Overall, the RSSC pangenome tends to stabilize once a steady low number of conserved genes was maintained through all genomes ( Figure 5).
In ANI analysis, three groups of genomes were formed, separated by similarity below 90% ( Figure 6). First, R. solanacearum genomes from public databases isolated in China and Japan clustered with R. pseudosolanacearum genomes isolated in Brazil, indicating they all belong to phylotype I. Next, in phylotype III, three R. solanacearum genomes isolated in Africa composed a much smaller cluster relatively similar to the previous one, indicating they're also R. pseudosolanacearum genomes. The following cluster was composed of R. solanacearum genomes isolated in South Korea, but since they were not similar to the previous or next cluster, they make up phylotype IV or R. syzygii. None of the new isolates clustered within the clusters mentioned up to now. The reminiscent cluster was composed of American isolates, except for CFBP strains from Iran. Among this last cluster, there was a clear division of more similar genomes: isolates CCRMRs283, CCRMRs286, CCRMRs294, and CCRMRs317 formed a smaller cluster (Cluster 1), and B75, B106, CCRMRs91, CCRMRs121, CCRMRs223, CCRMRs279, CCRMRs302, CCRMRs314, and CCRMRs339 formed a more significant cluster (Cluster 2). The genome from isolate B4 was the most distinct of all 14 but still more similar to Cluster 1 than Cluster 2 genomes. The new isolates present in Cluster 1 caused only Moko disease, while public genomes in it were obtained from isolates of other ecotypes, including NPB and brown rot. New isolates present in Cluster 2 caused Moko disease or BW on tomato only, indicating they are, respectively, phylotypes IIB and IIA, and part of the actual R. solanacearum species. This distribution of strains was also true for the phylogenomic tree based on gene family conservation found on OrthoFinder ( Figure 7); however, it is possible to notice that isolate B4 was more distant from other IIB isolates than in ANI profile, clustering a clade with CFBP8695, CFBP8697, RS488, UY031, and UW163. That being the case, all sequenced isolates causing BW in this study are phylotype II strains.   FJAT91F8  SL3103  KACC10709  T42  SL3730  SL2729  SL3882  T60  T117  SL3822  T78  SL3300  FJAT1303F8  FJAT15353F1  FJAT1303F50  FJAT15353F8  FJAT15353F50  FJAT15244F8  FJAT15244F1  FJAT15244F50  FJAT1303F1  FJAT15304F1  FJAT15304F6  FJAT15304F50  FJAT15340F50  FJAT15340F6  FJAT15340F1  CQPS1  202  203  204  RSCM  EP1  MAFF301560  OE11  T110  SL2330  T25  SL3755  HA41  362200  B2  YC45  MAFF211491  MAFF241648  MAFF211479  MAFF311693  SEPPX05  UW763  GM1000  UW386  UW773  UW774  PSI07  T98  SL3175  T95  SL2064  KACC10722  T51  T11  SL3022  T101  SL2312  T82 T12  FJAT1458F1  FJAT1458  FJAT1458F50  FJAT1463F50  FJAT1463F51  FJAT15249F1  FJAT15249F50  FJAT15252F1  FJAT15252F50  FJAT448F1  FJAT448F50  FJAT454F1  FJAT454F501  FJAT91  FJAT91F1  FJAT91F50  FJAT91F8  KACC10709  MAFF211472  MAFF211471  FJAT1452F1  FJAT1452F50  FJAT442F1  FJAT442F50  FJAT445F1  FJAT445F50  MAFF241647  202  203  204  CQPS1  EP1  FJAT1303F1  FJAT1303F50  FJAT1303F8  FJAT15244F50  FJAT15244F1  FJAT15244F8  FJAT15304F1  FJAT15304F6  FJAT15304F50  FJAT15340F50  FJAT15340F1  FJAT15340F6  FJAT15353F50  FJAT15353F1  FJAT15353F8  MAFF301560  OE11  RSCM CCRMRs91  RS489  UW576  SL2064  KACC10722  T11  T51  T95  SL3022  T82  T12  T101  SL2312  PSI07  SL3175 T98 Burkholderia cenocepacia MSMB384WGS Figure 7. Phylogenomic tree of R. solanacearum species complex (RSSC) strains used in this work. Strains' names are colored according to Figure 6 pattern: phylotypes I in orange, phylotypes III in brown, phylotypes IV in blue, and phylotypes II in black. Newly sequenced genomes are also highlighted according to their respective ecotypes.
Even though the ANI analysis and phylogenomic tree evidence two distinct clades within the newly sequenced isolates, the in silico DDH values varied significantly from pairwise comparison, and it was not possible to find a consensus that separated IIA and IIB isolates in subspecies (Figure 8).  Figure 8. Heatmap representing the in silico DNA-DNA hybridization (isDDH) of genomes from R. solanacearum species complex (RSSC). The value for each DDH is available inside each dot.

Prediction of Rips Repertoire of R. solanacearum Strains and Ecotype Correlation
In total, 88 subfamilies of Rips were present in the new 14 isolates. Overall, B4 and B76 had the lowest and the highest number of predicted Rips, with 67 and 76 out of 88, respectively. The Rips repertoire of each isolate is available in Figure 9.   Cluster IIB Rips repertoire was more homogeneous than Cluster IIA's, despite a higher number of isolates causing Moko in the latter. Considering both clusters, only six events of Rip duplication occurred: RipA5, RipEI1, RipE2, RipS1, RS_T3E_Hyp7, and RS_T3E_Hyp8. There was no duplication event in common for both Clusters, and there was no Rip absent in all 14 isolates besides the hypothetical ones. Congruently with the profile observed in ANI, B4 also exhibited the most distinct pattern of Rips presence-absence-duplication compared to the other 13 isolates. Starting with Cluster IIB, they commonly shared 55 Rips, and B4 presented ten exclusive Rips: RipAQ, RipAW, RipBD, RipF1, RipM, RipS6, RipV2, RS_T3E_Hyp3, RS_T3E_Hyp4, and RS_T3E_Hyp7. All isolates shared a duplication of RipA5; however, only B4 did not share a duplication of RipE1, and B4 and CCRMRs294 shared an absence of RipA4. In Cluster IIA, isolates causing Moko disease commonly shared 51 Rips. The isolates CCRMRs279, CCRMRs302, and CCRMRs314 had remarkably similar repertoires, except for the absence of RipAT and RipBC in CCRMRs302, and the absence of RipF1 in CCRMRs279. Finally, the new BW isolates commonly shared 63 Rips. Both isolates have almost the same repertoire, except for the absence of RipAR, RipAX2, RipH3, RipP3, and RS_T3E_Hyp7 in CCRMRs121, and the absence of RipH2 in CCRMRs223. Moko IIA, Moko IIB, and BW isolates commonly shared 43 Rips. These comparisons also revealed that very few Rips were ecotype-specific; the 12 isolates causing Moko disease only commonly shared RipH3. Moko IIA isolates exclusively shared RipAR, while Moko IIB isolates exclusively shared four Rips (RipAA, RipJ, and RS_T3E_Hyp10), and BW isolates exclusively shared six Rips (RipA4, RipAX1, RipK, RipS7, RipT, and RipV2). More Rips were exclusively shared among Moko IIB and BW isolates (RipAT, RipE2, RipN, RipTPS, RipU, RipZ, RS_T3E_Hyp8, and RS_T3E_Hyp9) than among Moko IIA and BW (RipAQ, RipAW, RipAZ1, RipF2, RipM, and RipY) ( Figure 10). Of the 22 candidate Rips for Moko disease suggested by Ailloud et al. [28], only 4 were not commonly shared by the 12 Moko isolates: RipAA, RipE2, and RipF1 in Moko IIA and RipF1 and RipH2 in Moko IIB. However, it is important to highlight that only three of those Moko candidate Rips were not commonly shared by BW isolates: RipF1, RipH2, and RipAA. In contrast to public genomes of other Brazilian R. solanacearum infecting tomato, RS488 and RS489 (BW2) shared 44 Rips with CCRMRs121 and CCRMRs223 (BW1), with 16 Rips exclusively shared by BW1. Interestingly, isolates in BW1 shared more Rips with Moko IIA and IIB isolates than BW2, resting only 29 Rips shared by all four groups from the 43 early found ( Figure 11). The Rips repertoire comparison of BW1 and BW2 is available in Figure A1.

Pangenome and Nucleotide Identity Analysis Reveal Global Misclassification of RSSC Isolates in Public Databases and Genetic Diversity of New Brazilian Isolates
Our study used a large dataset of high-quality publicly available R. solanacearum complete genomes, elucidating their taxonomy via robust in silico whole-genome approaches confirming many previous findings [9,10,35,39,40]. The misclassification of older R. solanacearum genomes has recently been addressed by Sharma et al. [41], who also pointed out the discrepancy in representative genomes from African and South Asian isolates available at NCBI. As the sequevar/biovar classification has been shown to fail at represent the diversity of highly recombinogenic isolates [41], whole genome methods, such as ANI, isDDH and phylogenomic inferences are more prone to accurately provide the genetic diversity on RSSC and other bacterial phytopathogens with controversial taxonomy [35]. The open pangenome profile observed through our analysis corroborates up-to-date studies, with similar values found for core, accessory, and unique genomes [42] (Geng et al., 2022). As soil borne microbes, the resistance of Ralstonia strains to heavy metals was described a long time ago [5,43,44], but no recent analyses have included Brazilian isolates. Considering that pesticides and fertilizers commonly used in high-production crops typically comprise heavy metals in their composition [45,46], these findings raise an alert for small and big producers in countries such as Brazil that struggle with R. solanacearum infestation. As for the new RSSC isolates from Brazil's Northern and Northeastern regions, they fit in R. solanacearum, but are from separate phylotypes. The fact that 12 isolates infecting Musa in close geographic spots still differ in phylotype sublevel only shows how diverse Brazilian RSSC isolates are, which corroborates with the hypothesis of the Amazon region being the diversity center of phylotypes IIA and IIB [7,47]. At first look, most of the newly sequenced isolates fitting in phylotype IIA might seem a surprise, as most of the Brazilian isolates are actually included in phylotype IIB, followed by phylotypes IIA and I. However, it has also been reported that phylotype IIA isolates have a higher proportional presence in Brazil's North and Northeastern regions. In contrast, phylotype IIB has a higher abundance in Central, Southeastern and South regions [7,48]. Since phylotype IIA isolates have been characterized as more genetically diverse and recombinant than IIB [47], a less diverse repertoire of Rips was expected for the latter, and also because at least two different ecotypes were suspected for IIA isolates. Analyzing the Rips repertoire is important because each subfamily of Rip plays distinct roles throughout the infection process, depending on the environment, tissue, and signals recognized within the host [49]. On that thought, we suggest here that Rips that were not commonly shared by all isolates of their respective ecotype might contribute to their individual virulence when infecting the host. In this sense, a good indicator is that B4 was isolated from a banana plant with more severe symptoms on roots, while all other Moko isolates were isolated from banana plants with wilted leaves and healthy roots. The presence of more than one copy of Rips and paralog subfamilies in RSSC is largely documented. Even though it has been described as genetic redundancy, it is also seen as a general strategy for giving bacterial virulence robustness via acting on similar targets, participating in the same molecular functions and biological processes [18,50].

Rips Repertoire of Brazilian Isolates Are More Correlated to Genomic Similarity Rather Than Ecotype
As for the duplicated Rips, RipA5 (AWR5), and RipE act as typical avirulence factors eliciting hypersensitive responses on Arabidopsis thaliana and Nicotiana benthamiana and suppressing the expression of jasmonic-acid-dependent genes and salicylic acid synthesis [51,52]; however, RipAC and RipAY inhibit RipE1-mediated HR [18]. In contrast, RipS1 acts as a virulence factor that inhibits key targets on reactive oxygen species (ROS) pathways [51], considering that alone, CCRMRs279, CCRMRs302, and CCRMRs314 would have a higher potential for more virulent behavior.
A few Rips have been correlated with host specificity in South Asian RSSC strains infecting solanaceous hosts, with RipAS3 and RipH3 linked to pathogenicity in tomato and RipAC linked to pathogenicity in eggplants [53]. However, in RS488 no copy for RipS3 was predicted (see Figure A1), even though it caused BW in tomato. Indeed, this reinforces the hypothesis that a repertoire of Rips is keener to the success of pathogenicity in some hosts than in a few groups of Rips. We suggest that the presence of previous Moko candidates was more accurate for IIB isolates because it compared different ecotypes present only in phylotype IIB, such as NPB and brown rot, and also due to their clonal behavior. Therefore, the greater difference observed in Moko IIA isolates might indicate different selective pressure on those strains derived from the higher genetic diversity observed in this phylotype, making Rips gain or loss more probable. The broad presence of Moko candidates in the new BW isolates and their broadly shared Rips repertoire can be explained by two main arguments: the first one is that the most common recent ancestor of Ralstonia isolates was already capable of infecting banana and host-adapted polymorphisms (HAP) would be present in the Rips derived from it, making them functionally specialized either for solanaceous and musaceous hosts respective defense mechanisms [28].
Moreover, it has been shown that NPB and Moko disease strains have minimal genomic differences and still have high gene expression differences when infecting their respective hosts [54]. Thus, even if Moko and BW isolates have no significant differences in their Rip sequences, their gene expression would still differ when infecting different hosts. The second argument is that these BW isolates are actually from Moko ecotype infecting solanaceous hosts due to the optimal environmental conditions found in Brazil's Northeastern region, as it has been previously reported in environments with high temperatures and humidity conditions [55,56]. This argument gains strength when considering that the pan-effectome of R. solanacearum is clearly diverse with a small core effectome of 16 Rips [13,18] contrasting with the 43 Rips present in all 14 isolates of two different ecotypes. Moreover, their Rips repertoire was more similar with Moko isolates than with RS488 and RS489. Hence, based on what we found and considering the second argument, only 14 Rips would be eligible candidates for Moko disease: RipA2, RipAS, RipAU, RipG3, RipG4, RipG6, RipH1, RipL, RipS1, RipS2, RipS3, RipS4, and RS_T3E_Hyp12.

Conclusions
The present study is the first to include Brazilian isolates of Ralstonia and use a robust effector database to characterize the effectome of Brazilian isolates. The commonly shared Rips by isolates in different ecotypes might aid in further phytopathology studies by providing target avirulence proteins in hosts when searching for breeds resistant to bacterial wilt, Moko, and so on. It is important to note that further research efforts, preferably with in vitro and in vivo data on gene expression and infection essays on different hosts are required to determine whether the Rips identified here are essential candidates for ecotype specificity. In addition, more phylotype IIA isolates causing bacterial wilt in Solanaceae to identify commonly shared Rips in this ecotype are needed. Finally, even though Rips presence/absence is an excellent indication for host specificity association, it is not the final determinant. The whole-genome approaches were essential in correctly identifying these isolates' taxonomy, proving their potential for solving complicated bacterial species complexes, such as RSSC. Efforts to characterize hypothetical and redundant Rips are essential to elucidate missing roles on molecular pathways linked to triggered and innate immunity in plants.  Table A1.

Conflicts of Interest:
The authors declare no conflicts of interest.

Abbreviations
The following abbreviations are used in this manuscript:     Figure A1. Heatmap of Rips repertoire of newly sequenced (Bac. Wilt 1) and publicly available (Bac. Wilt 2) genomes from isolates causing bacterial wilt in Brazil. The genomes on the right have notably fewer Rips than the newly sequenced ones, despite all causing diseases in tomato.