Introduction

A study of molecular epidemiology and evolution of Mycobacterium tuberculosis has been greatly facilitated by the lack of horizontal gene transfer and strictly clonal population structure of this medically relevant biological species. The clonality implies that the population structure is hierarchical and, as we know, consists of large phylogenetic lineages, smaller genetic families or sublineages, and finally clonal clusters. Pathogenetically significant properties may be featured by any of these entities although clonal clusters of the closely related isolates are of particular epidemiological/clinical interest. This interest becomes even more pertinent if such drug resistance-associated and/or hypervirulent clusters demonstrate global or local population increase hence impact on the public health programs.

The M. tuberculosis Beijing genotype is globally spread lineage with important medical properties. The evolutionary history of the Beijing genotype is far from straightforward and was marked by some key turning points shaped by human migrations and demography. While Beijing itself likely emerged in the North of China, the ancestral lineage termed as proto-Beijing originated in the South of China [1]. Ancient or ancestral branch of the Beijing genotype is dominant in Japan, Korea, parts of China and Vietnam but extremely rare elsewhere in the world [2,3,4,5,6]. These strains have not been marked with particular clinically significant properties and show decreased transmission e.g., in Japan [7]. Although phylogenetic sublineages (ancient/ancestral and modern) of the Beijing genotype were first postulated in a Russian study [8], the ancient Beijing strains have been rarely found in Russia and did not attract any particular attention.

That being said, it was a surprise to find two clusters of exclusively MDR strains of the ancient Beijing sublineage in two locations in the Asian part of Russia [9] (see clusters 1071-32 and 14717-15 on Fig. 1). A murine model study demonstrated that one of these clusters 14717-15 that belongs to the RD181-intact sublineage, is not only MDR but highly lethal and hypervirulent [10]. In spite of this, this strain is prevalent only in one area in the Russian Far East, namely in Buryatia (16%). It was hypothesized that this situation is a result of the particular interplay of the human and bacterial genetics and long-term adaptation of these strains to the local human population.

Fig. 1
figure 1

Simplified evolutionary scenario of the M. tuberculosis Beijing genotype, including ancient Beijing 14717-15-cluster [47]. Main Russian clusters are in bold

Full size image

The strain was the most lethal of all Russian Beijing strains studied to date, including the notorious Beijing B0/W148 cluster, yet its phylo- and pathogenomics and geography were not studied in sufficient detail. In the present study, we aimed, based on analysis of the expanded strain/DNA collection and whole genome sequencing, to identify pathogenetically relevant genomic features of the Beijing 14717-15 cluster, to develop a simple method of its detection and to assess its geographic distribution in Eurasia.

Materials and methods

Study collections

The collection included DNA extracted from M. tuberculosis strains obtained between 1996 and 2020, within prospective or cross-sectional studies or collected as convenience samples, characterized in our previous studies [2, 8,9,10,11]. The study was approved by the Ethics Committees of St. Petersburg Pasteur Institute, St. Petersburg, Russian Federation (protocol 41 of 14 December 2017) and the Research Institute of Phthisiopulmonology, St. Petersburg, Russian Federation (protocol 31.2 of 27 February 2017). All methods were performed in accordance with the relevant guidelines and regulations.

Genotyping

DNA was extracted from cultured M. tuberculosis isolates using the CTAB-based method [12], DNA-Sorb-B kit (Interlabservis, Russia), or GenoLyse® kit (Hain Lifescience). One microliter of the DNA extracted using DNA-Sorb-B or GenoLyse® commercial kits and 10–20 ng of DNA extracted using CTAB method was used for PCR.

Spoligotyping and 24 loci MIRU-VNTR typing were performed according to standard protocols [13, 14]. The Beijing genotype was identified experimentally or in silico based on deletion RD207 (positions 3,120,521–3,127,920 in H37Rv genome, NC_000962.3). The main sublineages of the Beijing genotype were identified by the following molecular markers: (i) mutT4 codon 48 CGG > GGG mutation, (ii) mutT2 codon 58 GGA > CGA mutation, (iii) deletion RD181 (positions 2,535,429–2,536,140 in H37Rv genome, NC_000962.3). These three markers permit to differentiate between early ancient 1, early ancient 2, and classical ancient subgroups of the Beijing genotype (Fig. 1) [9]. Compared to some of the previous classifications summarized by Shitikov et al. [15], early ancient 1 and 2 correspond to Asia Ancestral 1 and 2 branches, respectively.

Whole genome sequencing

Whole genome sequencing was performed at the HiSeq platform (Illumina). DNA libraries were prepared using ultrasound DNA fragmentation and NEBNext Ultra DNA Library Prep Kit for Illumina (New England Biolabs). Data for the M. tuberculosis sequenced genomes were deposited in the NCBI Sequence Read Archive (project number PRJNA822891).

TB Profiler database (http://tbdr.lshtm.ac.uk/) was used for genotypic detection of drug resistance. MDR, pre-XDR and XDR phenotypes were defined according to the updated World Health Organization definitions: MDR are strains resistant to isoniazid and rifampicin; pre-XDR - resistant to isoniazid, rifampicin, fluoroquinolone; XDR - resistant to isoniazid, rifampicin, fluoroquinolone plus bedaquiline and/or linezolid [16].

Bioinformatics and phylogenetic analysis

A dataset comprising Mycobacterium tuberculosis lineage 2 isolates with intact RD181 (n = 618) and one H37Rv isolate was retrieved from NCBI database (https://www.ncbi.nlm.nih.gov/sra) using SRA Toolkit v3.0.0 (https://github.com/ncbi/sra-tools) and parallel-fastq-dump v0.6.7 (https://github.com/rvalieris/parallel-fastq-dump). Quality of downloaded FASTQ files was assessed with FastQC v0.11.9 (https://github.com/s-andrews/FastQC).

These 618 genomes included 8 Russian genomes and 610 genomes from 23 other countries [1, 15, 17,18,19,20,21] (see Table S1 with accession numbers). The TBvar v1.1.5 workflow (https://github.com/dbespiatykh/TBvar) was used for mapping and variant calling. In brief, FASTQ reads were mapped to the reference M. tuberculosis H37Rv genome (RefSeq accession no. NC_000962.3) using BWA MEM v0.7.17 [22] algorithm. Mapped reads were sorted by coordinates, converted to BAM format and indexed using SAMtools v1.16.1 [23]. Subsequently, duplicate reads were removed with Sambamba v1.0 [24]. Mapping quality was assessed with SAMtools stats and mosdepth v0.3.3 [25]. All the following variant calling steps were performed with GATK4 v4.3.0.0 package [26]. All reports were aggregated with MultiQC v1.10.1 [27]. Variants effects were annotated with SIFT4G v19.0.2 [28] and SnpEff v5.1d [29].

Lineages from called SNPs were assigned with TbLG v0.1.5 (https://github.com/dbespiatykh/tblg). TB-Profiler v4.4.2 was used to discover resistance mutations and for spoligotyping [30]. To construct the phylogenies, the SNP alignment was extracted from the tab-delimited output of GATK VariantsToTable. Repetitive regions were excluded using a mask from a previously published study (available at https://github.com/mbhall88/head_to_head_pipeline/blob/master/analysis/baseline_variants/resources/compass-mask.bed) [31]. Recombinant regions from the SNP alignment were filtered out using Gubbins v3.2.1 [32]. The resulting alignment was cleaned with SNP-sites v2.5.1 [33]. Maximum likelihood (ML) phylogeny was inferred from 619 sequences with 16 220 nucleotide sites using IQ-TREE 2 v2.2.0.3 [34]. Support values were inferred from 1 000 ultrafast bootstrap replicates (UFBoot [35]) with the “-bnni” argument and from 1 000 replicates for Shimodaira-Hasegawa (SH) approximate likelihood ratio test with the “-altr” argument. Best-fit model was determined by ModelFinder [36] with the “-m MFP” argument, best-fit model according to Bayesian information criterion (BIC) was K3Pu + F + ASC + R7. M. tuberculosis H37Rv1998 (SRR20082811) was used as an outgroup. ML phylogeny was visualized with the ggtree v3.7.1.002 [37], ggtreeExtra v1.4.2 [38], ggplot2 v3.3.6 (https://ggplot2-book.org/), ggstar v1.0.4 (https://github.com/xiangpin/ggstar), ggplotify v0.1.0 (https://github.com/GuangchuangYu/ggplotify), ggnewscale v0.4.7 (https://github.com/eliocamp/ggnewscale), randomcoloR v1.1.0.1 (https://github.com/ronammar/randomcoloR), and tidytree v0.4.2 (https://github.com/YuLab-SMU/tidytree) packages for R v4.1.2 [39]. To construct minimum spanning tree (MST) SNP distance matrix was created using Seqtk v1.3-r106 (https://github.com/lh3/seqtk) and snp-dists v0.8.2 (https://github.com/tseemann/snp-dists). MST tree was inferred and visualized using ape v5.7 [40], igraph v1.4.1 (https://github.com/igraph/rigraph), ggnetwork v0.5.12 (https://github.com/briatte/ggnetwork), and ggplot2 v3.4.1 R packages.

The NGS data (fastq files) were used for in silico spoligotyping using SpoTyping program [41].

For the enrichment analysis Clusters of Orthologous Genes (COG) categories were annotated using eggNOG-mapper v2.1.9 [42], gene ontology (GO) categories with PANNZER tool [43] using Positive Predictive Value (PPV) cutoff of 0.5, KEGG pathways were annotated using BioServices v1.11.2 [44] Python library. Additionally, functional categories from the TubercuList database were also tested for enrichment [45]. All enrichment analyses were performed in R using Fisher’s exact test.

The significance of amino acid substitutions was assessed using PAM1 (Point Accepted Mutation 1) values calculated by PhyResSE online tool. The SIFT tool was used to predict whether an amino acid substitution affects protein function based on sequence homology and the physical properties of amino acids (https://sift.bii.a-star.edu.sg/index.html).

PCR-RFLP analysis of Beijing 14717-15-clusters SNPs

Two SNPs at genome positions 2,423,040 and 1,448,330 were tested by HhaI PCR-RFLP assays.

The first SNP is at genome position 2,423,040 A > G and concerns gene Rv2161c (amino acid change in codon 33 Val > Ala [GTG/GCG], gene position 98T > C). This A > G mutation creates an additional site for HhaI (GCGC). Two primers are used for PCR of this gene region: 2423040F 5’-GTCCGGCAGCTCTCCACCG and 2423040R 5’-TGCAGTTCGTCACCGACCTGACC. PCR conditions: 95 °C, 5 min; 35 cycles of 95 °C, 30 s, 67 °C, 20 s, 72 °C, 20 s, and final extension 72 °C, 3 min. PCR product size was 146 bp. After HhaI digestion at 37 °C for 3 h, the fragments were separated in 1.4% standard agarose gel. The profile for wild type allele consists of two fragments 87 and 59 bp, and in case of mutation, of three fragments 65, 22, and 59 bp.

The second SNP is at genome position 1,448,330 G > T and concerns gene Rv1293 (lysA) (silent mutation in codon 101-Ala). This mutation inactivates the single HhaI site in this gene fragment. Two primers are used for PCR of this gene region: 1448330F 5’-TGGAAGTGGGGCGAACGTGC and 1448330R 5’-TTGACCGCAGCGGTCAACTCTGA. PCR conditions were the same as above, and PCR product size was 201 bp. After HhaI digestion at 37 °C for 3 h, the fragments were separated in 1.4% standard agarose gel. The profile for wild type allele consists of two fragments 121 and 80 bp, and in case of mutation, the PCR product remains undigested 201 bp.

Results and discussion

Phylogenomic position of Beijing 14717-15-cluster

Phylogenomic analysis of the Beijing isolates with intact RD181 (early ancient 1 sublineage of the Beijing genotype) was performed on 8 Russian genomes (2 from Omsk, West Siberia and 6 from Buryatia, Far East) and 610 genomes from 23 countries, mostly from East and Southeast Asia (Table S1, Fig. 2a). All Russian isolates clustered in a separate branch on the tree (see the uppermost branch on Fig. 2b). Four isolates from Europe were also found within this cluster and included three from the Netherlands and one from Spain. VNTR typing of the Russian isolates assigned all of them to the Mlva type 14717-15 and related profiles. For this reason, we term this branch the Beijing 14717-15-cluster. All isolates of this cluster had in silico deduced spoligotype SIT269 (Table S2) that is a derived profile from the classical Beijing SIT1 by deletion of spacers 35 and 36. The experimental spoligotyping profiles were available for the Russian isolates and were concordant with their in silico spoligoprofiles.

Fig. 2
figure 2

Phylogenomic analysis of the Beijing genotype isolates with intact RD181. (A) Global dataset (n = 618); (B) Beijing 14717-15-cluster and neighboring branches; (C) Minimum spanning tree of the Beijing 14717-15-cluster with information on the region of origin of Russian strains and year of isolation of the isolates (when available)

Full size image

On the phylogenetic tree, the relatively nearest neighbors of the Beijing 14717-15-cluster were isolates from Korea and Japan (Fig. 2b). All Beijing 14717-15-cluster isolates and some Korean isolates had spoligotype SIT269 however, given a limited number of spacers in the CRISPR locus of the Beijing genotype, this profile may be a result of convergent evolution and does not necessarily indicate a common origin.

The phylogenetic network (Fig. 2c) shows that 115 SNPs separated Russian cluster from the most recent common ancestor with isolates from Korea which implies only very distant relation of these isolates. It may be noted that Korean isolates were separated by even more SNPs between them (mostly 130–200 SNPs [not shown]) that likely correlates with their long-term evolution in a country of the endemic high prevalence of the RD181-intact ancient Beijing sublineage, i.e. Korea.

Resistance mutations were detected in silico based on the WGS data (Fig. 2b, Table S3). One should keep in mind a certain bias of the studied isolates from some countries, in particular a collection from Korea included mainly drug resistant isolates. On the other hand, all Russian isolates came from the population-based studies and were not preselected in any way. In this view, MDR/pre-XDR status of Russian isolates is noteworthy. All isolates of this cluster harbored two-mutation signature of the high-confidence resistance mutations katG Ser315Thr and rpsL Lys88Arg [46]. Interestingly, two isolates from the Netherlands, 1998, harbored only two first-line drugs resistance mutations (in rpsL88 and katG315) and thus were likely brought to the Netherlands during the early dissemination of this strain. katG Ser315Thr is the most frequent INH-resistance associated mutation and its presence is expected. The other mutation rpsL Lys88Arg is also a well-known high-confidence mutation associated with STR resistance but it is less frequent than rpsL43 mutation [46]. Together these two mutations katG Ser315Thr and rpsL Lys88Arg may be considered as a characteristic marker of this cluster although they alone cannot be used for its identification.

We further identified polymorphisms specific of the Beijing 14717-15-cluster (Table S4). They included 55 SNPs in CDS (35 non-synonymous, 20 synonymous) and 10 SNPs in intergenic regions. Some of the SNPs were in the genes related to mycobacterial virulence and adaptation and could hypothetically influence an increased virulence and lethality of this cluster which was demonstrated previously in both murine model and in TB patients [10, 47]. For example, PPE18 is known to be related to immune evasion [48,49,50]. Some other genes (fadE17, mmpS3, pks7) are related to adaptation and virulence [51, 52]. Nevertheless, gene function enrichment analysis revealed that the genes with nonsynonymous mutations were only enriched in lipid metabolism category according to Tuberculist.

Based on the PAM 1 values, it is possible to hypothesize a significant influence of the amino acid change and such SNPs with PAM1 below 5 were identified in 10 genes including pks7, fadE17, hpx (Table S4).

In addition, SIFT P values were calculated for 35 nonsynonymous mutations. As a result, 12 SNPs in genes of different categories (Lipid metabolism, Regulatory proteins, Intermediary metabolism and respiration, PE/PPE, Cell wall and cell processes) were found to significantly affect protein function (P < 0.05) (Table 1). Information on these 12 genes was searched in Pubmed but only few of them were found and without relation to pathobiological properties. However, at least some of these genes such as, polyketide synthase Pks7, methyltransferase Rv0567, conserved transmembrane protein Rv0064, transcriptional regulatory protein Rv0823c and two PE/PPE genes deserve particular attention. In particular, Pks genes encoding the polyketide synthases are involved in the lipopolysaccharide and complex lipids biosynthesis [53]. Mutations in the pks genes were also suggested to have a compensatory role in drug resistance [51, 52].

Table 1 Twelve in silico predicted significant mutations characteristic of the Beijing 14717-15-cluster
Full size table

PCR-RFLP assay for detection of Beijing 14717-15-cluster

Among cluster-specific SNPs identified above, we selected two functionally neutral SNPs (PAM1 = 9867) and designed PCR-RFLP assays to detect them. These SNPs were in genome positions 1,448,330 G > T (Rv1293 Ala101Ala) and 2,423,040 A > G (Rv2161c Val(s)33Ala). The neutral SNPs reflect a neutral evolution non-influenced by selection pressure and unlikely to independently occur in different and unrelated phylogenetic groups. The use of two SNPs enhances the reliability of detection of this cluster.

Both SNPs can be detected by HhaI-RFLP analysis of the amplified PCR regions (Fig. 3). PCR conditions are the same for both genes, and both PCR products are digested (separately) by the same HhaI endonuclease. Both PCR-RFLP assays were optimized with isolates with known WGS sequences and VNTR profiles. Both mutations were found only in isolates of the Beijing 14717-15 cluster.

Fig. 3
figure 3

PCR HhaI-RFLP detection of Beijing 14717-15-cluster based on: (A) SNP at 2,423,040 A > G (Rv2161c Val33Ala) and (B) SNP at 1,448,330 G > T (Rv1293 Ala101Ala). Lanes 1–5 – Beijing 14717-15-cluster. Lanes 6–7 – other genotypes. М – molecular weight marker 100 bp ladder (Fermentas). The raw gel image is shown in Figure S1

Full size image

These two phylogenetic SNPs of the Beijing 14717-15 cluster were screened for specificity in the proprietary Beijing global genome databases (> 6000 genomes [Dr. Joao Perdigao, Universidade de Lisboa, Portugal, and > 10,000 genomes [Dr Egor Shitikov, Lopukhin Federal Research and Clinical Center of Physical-Chemical Medicine, Russia]). As a result, these two SNPs were found robustly specific and unique for the Beijing 14717-15-cluster isolates.

A strategy to target a limited number of SNPs (at least two) was recommended and applied to identify specific strains or clones by PCR based assays [54,55,56] which increases the robustness. Thus, analysis of both targeted SNPs is the most robust method to detect the Beijing 14717-15-cluster. Nonetheless, detection of particular clusters/genotypes based on use of a single marker is an acceptable and parsimonious approach, provided that such marker was proven specific and sensitive in the validation studies and this concerns both detection of the particular clusters and families and the development of the SNP-barcode system [15, 57]. In this view, since analysis of the two SNPs showed completely concordant results, testing of any of them appears the most practical and time-saving approach to trace this clinically significant MDR Beijing 14717-15-cluster.

It should be noted that the strains with the intact RD181 locus belong to the early ancient sublineage of the Beijing genotype, which is very heterogeneous and includes strains with diverse VNTR profiles. In this sense, the SNPs identified by us are markers only of the Beijing cluster specific to Russia (primarily Buryatia), but not markers of the entire heterogeneous RD181-intact branch within deeply-rooted ancestral Beijing sublineage.

Geographic screening of Beijing 14717-15-cluster

The two PCR-RFLP assays were further applied to the Beijing genotype isolates that represented different Beijing sublineages and had different VNTR profiles. These validation collections included isolates from Europe, Russia, Central and East Asia. The PCR-RFLP analysis of two SNPs correctly assigned all isolates with known Mlva 14717-15 and related profiles to the Beijing 14717-15-cluster. The method has 100% sensitivity and 100% specificity to detect Beijing 14717-15-cluster.

We further applied these PCR-RFLP assays to screen the available DNA collections from Russian regions and other countries. Results summarizing the above validation and screening analysis are shown in Table 2; Fig. 4 and demonstrate the clear peak of the Beijing 14717-15-cluster in Buryatia, Far East.

Table 2 Detection of Beijing 14717-15 in retrospective local collections
Full size table
Fig. 4
figure 4

Geographic distribution of the Beijing 14717-15-cluster isolates in Russian regions, based on results of this study and previous publication [10, 57,58,59]. Circle size is roughly proportional to the percent of these isolates in the local M. tuberculosis population. Gray shade in the circle means absence of these isolates in the analyzed collection (irrespective of the circle size)

Full size image

Analysis of the available archival strains isolated in 1996–2002 in northwestern Russia (St. Petersburg and other regions) did not reveal the isolates of the Beijing 14717-15-cluster. However, two isolates of this cluster were detected in the Netherlands in 1998. One possibly related isolate (based on 12-MIRU-VNTR typing) was described in Lithuania, and was isolated in 2007 [9].

We additionally looked at the geographic distribution of the main Russian clusters of the modern sublineage of the Beijing genotype (B0/W148 and Central Asian Russian) and two main clusters of the ancient sublineage (Beijing 14717-15 and Beijing 1071-32), based on results of this study and previous publication [9, 58,59,60]. This comparison showed the overall prevalence of the modern Beijing clusters across Russia and presence of Beijing 1071-32 at low prevalence but also indifferent parts of European Russia and Western Siberia. In this view, the high 18% prevalence of the Beijing 14717-15-cluster in Buryatia is in the striking contrast with its almost complete absence elsewhere.

We note that percent of this cluster roughly correlates with proportion of the Buryat ethnic group in Buryatia itself and its neighbors. Due to human influx from European and Siberian parts of Russia since the 1930s, the proportion of Buryats decreased from 44% to 1926 to 19% in 1970 but remains stable in the last 20–30 years and makes up to 28–30% of the total population of this region (https://en.wikipedia.org/wiki/Buryatia#Demographics). Buryats also live in the neighboring provinces in Far East and Siberia. Thus, presently, the percent of Buryats in Buryatia is 30%, in Zabaykalie 8% and in Irkutsk 3%. In turn, percent of the Beijing 14717-15-cluster in these areas is 18%, 8%, and 2%, respectively. We do not have information on the ethnic background on the patients in the previous studies but the above figures are suggestive of some correlation.

Interestingly, no Beijing 14717-15 strains were found in the neighboring Mongolia [59]. Buryat and Mongol languages are related, and Y chromosome and mtDNA based study identified common genetic components for Buryats and Mongols [61, 62] but Buryats were separated form Mongols very long ago, and definitely long before emergence of this particular M. tuberculosis strain. The noticeable decrease in frequency of N1c1 haplogroup in western direction and the presence of a significant proportion of unique haplotypes in Buryats indicate the absence of the intensive gene drift from Buryats to Mongols [61]. Based on mtDNA graphs, Buryats are very heterogeneous and only one of their subgroups is close to Mongols [62]. A relatively mass migration of Buryat people to Mongolia took place 90 years ago when they flew Red Army. However, no significant human movement took place since the 1930s and the two countries are separated by the state borders. This could be the reason why this strain was not brought to Mongolia form the neighboring Buryatia.

Conclusions

Important strains may unexpectedly emerge among minor genotype lineages as was shown for genotypes of the Euro-American lineage, such as drug-resistant clones within Haarlem, LAM, Ural, NEW-1 families [63,64,65,66]. Herein described the Beijing 14717-15-cluster is the other relevant example. The strain was shown concordantly lethal and virulent in mice and human studies [10, 47]. Its elevated prevalence only in one region was linked to some hypothetical interplay of human immune system and the genetic background of this strain during local coevolution and long-term coadaptation. Further studies including GWAS-based may eventually shed more light.

Cluster-specific SNPs that significantly affect protein function were identified in 12 genes of different categories (Lipid metabolism, Regulatory proteins, Intermediary metabolism and respiration, PE/PPE, Cell wall and cell processes). Most of these genes were previously unreported and could potentially be associated with increased pathogenic properties of these strains.

Furthermore, when the entire bacterial genome is considered, not only SNPs but also insertions and deletions could be cluster-specific and functionally significant. A further study of such alterations and their association with pathogenic properties of the isolates is warranted through more complete genome sequencing (including de-novo assembly and long-read sequencing), and experimental allelic exchange approach.

The Russian isolates of the cluster 14717-15 were from the Asian part of the country. They had two common resistance mutations rpsL Lys88Arg and katG Ser315Thr. Phylogenetically, their neighbors were isolates from Korea, while the Russian isolates from both Omsk and Buryatia and some of the Korean isolates had a characteristic spoligoprofile SIT269 (derived from the classic spoligo profile Beijing - SIT1). However, the distance between Russian and the closest Korean isolates was at least 115 SNPs (corresponding to ~ 230 years, based on generally assumed mutation rate of 0.5 SNPs/genome/year) and SIT269 may well result from convergent evolution. In this view, the hypothesis of the Korean distant descent of this medically significant Russian cluster remains a speculation. Availability of more genomes from East Asia should hopefully permit more robust reconstruction of its evolutionary history while omics studies may help to reach a more informed view on pathobiological relevance of its genetic variation.