Whole-genome analysis of carbapenem-resistant Acinetobacter baumannii from clinical isolates in Southern Thailand

Graphical abstract


Introduction
Multidrug resistance (MDR) in Gram-negative bacteria is a global public health concern as the treatment options are dramatically limited [1,2]. These pathogens have a high level of resistance to available antimicrobial classes, especially carbapenems and colistin, which are considered to be the last-line treatments [3][4][5]. Among them, carbapenem-resistant Acinetobacter baumannii (CRAB) is an important cause of nosocomial infections associated with high mortality rates [6]. It is commonly transmitted in intensive care units (ICUs ventilator-associated pneumonia, wound infections, urinary tract infections (UTIs), bloodstream infections, and meningitis [7,8].
CRAB can be resistant to carbapenems through various mechanisms. Carbapenemase production is the major mechanism of carbapenem resistance in A. baumannii as well as other Gram-negative bacteria. The carbapenemase enzymes have been classified by Ambler into three classes, class A, class B, and class D carbapenemases [9]. The carbapenemase-encoding genes are mainly located on mobile genetic elements (MGEs), plasmids, transposons, and integrons. Due to the presence of MGEs, many carbapenemase genes can be transferred between plasmids and chromosomes. Additionally, they can be horizontally transferred from one bacteria to another bacteria, leading to the rapid dissemination of carbapenemase genes [10]. Although the genetic basis associated with antimicrobial resistance (AMR) and bacterial pathogenesis among CRAB isolates has been characterized [11,12]. The genetic processes supporting the co-acquisition of multiple carbapenemase genes as well as other AMR genes still need to be elucidated further. Whole-genome sequencing (WGS) has become a powerful tool for rapidly analyzing the entire genomic DNA sequence of organisms. It has been used to characterize and understand the mechanisms of AMR and their spread through bacterial species, which is necessary for combating AMR-bacteria [13,14].
Previously, most of the studies have been reported the genetic characteristics of CRAB clinical isolates from many countries such as Korea, Thailand, Vietnam, Myanmar, Malaysia, and Brazil as well as in European countries [1,3,[15][16][17][18]. Their findings demonstrated that a high level of the CRAB isolates was assigned to an ST2 with the carriage of the bla OXA-23 gene and other AMR genes conferring resistance to many antibiotic classes, particularly aminoglycosides. In addition, the RecAci1 plasmid was predominantly found in the CRAB isolates, while the insertion sequence (IS) elements (e.g., ISAba1 and ISAba125) were also detected in the CRAB isolates.
Although several studies have provided WGS data of many CRAB isolates which are for understanding the distribution of AMR genes, virulence genes, and MGEs, there are to date few studies examining the genomic characteristics of CRAB isolates from Southern Thailand. Elucidation of the mechanisms of acquired AMR genes and virulence-associated genes and the genomic diversity among the CRAB isolates would help better understand their dissemination patterns in the regions. This is because CRAB isolates continuously evolve to survive in harsh environments, and they have been spreading throughout the world for a long time. Importantly, the AMR genes can be horizontally transferred (conjugation and transduction) to other related pathogenic bacteria, causing the rapid and global spread of AMR in Gram-negative bacteria. Thus, the objective of this study was to analyze the whole-genome sequence of CRAB, isolated from patients in 7 hospitals within lower Southern Thailand, to gain genomic insights into the clinical CRAB isolates of this area. The understanding and tracing of the rapid evolution of MDR in Gram-negative bacteria will play an important role in controlling these bacteria and slowing their spread until more effective treatments become available.

Bacterial isolates and clinical data
In this study, a total of 221 CRAB isolates were obtained from the Clinical Microbiology Laboratories (CMLs) of 7 hospitals located in lower Southern Thailand including Trang Hospital (n = 62; 28.05%), Songklanagarind Hospital (n = 54; 24.43%), Phatthalung Hospital (n = 45; 20.36%), Songkhla Hospital (n = 44; 19.91%), Satun Hospital (n = 7; 3.17%), Pattani Hospital (n = 7; 3.17%), and Yala Hospital (n = 2; 0.90%) (Fig. 1a). The 221 CRAB iso-lates were collected mostly from sputum but also from other clinical specimens (e.g., urine, pus, blood, body fluids, and tissue) of 221 patients who were admitted to the hospitals between March and August 2019 (Fig. 1b). In CMLs, the A. baumannii strains were identified by biochemical tests, according to Bergey's Manual of Systematic Bacteriology [19] and confirmed by Matrix-Assisted Laser Desorption/Ionization-Time Of Flight (MALDI-TOF) mass spectrometry (MS) [20,21]. The phenotypic resistance to carbapenem (imipenem and meropenem) in A. baumannii strains was evaluated by disk diffusion method. A. baumannii strains were defined as resistant to carbapenem, when the zone diameters were 18 mm for imipenem and/or 14 mm for meropenem, according to the CLSI guideline (2018) [22]. The inclusion criterion for CRAB isolates was the A. baumannii strains that were resistant to carbapenem, while A. baumannii strains without carbapenem resistance were excluded from the study.

Genome library preparation and sequencing
Genomic DNA of all the CRAB isolates was extracted using the TIANamp Bacterial DNA Kit (Tiangen, Beijing, China), following the manufacturer's instructions. The extracted DNA was sent to the Beijing Genomics Institute (BGI) in China for short-read WGS. For testing sample qualification, the DNA concentrations were measured by Qubit Fluorometer (Invitrogen), while DNA integrity and purity were investigated by Agarose Gel Electrophoresis. Then, 1 mg of the qualified genomic DNA (23 kbp) was randomly fragmented by Covaris. Fragmented sequences with a size of 800 bp were selected using the Agencourt AMPure XP-Medium kit. End-repair and 3 0 -adenylation were performed on the fragments, and adaptors were ligated to the ends of these 3 0adenylated fragments to amplify the fragments. The PCR products were then purified using an Agencourt AMPure XP-Medium kit. The double-stranded PCR products were heat-denatured and circularized by the splint oligo sequence. The single-strand circle DNA (ssCir DNA) was formatted as the final library. The quality of the whole-genome library was checked by quality control (QC). The qualified libraries were sequenced by BGISEQ-500 (BGI, China). Finally, 150-bp paired-end reads were received by combinatorial Probe-Anchor Synthesis (cPAS).

Genome assembly and annotation
De novo assemblies of our 221 CRAB genomes were generated using SPAdes v3.12 [23]. The quality and completeness of the genome assemblies were assessed by Quast v5.0.2 [24] and Busco v5.1.2 [25,26], respectively. According to the exploration of reported A. baumannii genomes in the National Center for Biotechnology Information (NCBI) (https://www.ncbi.nlm.nih.gov/), we predicted that the highest length of A. baumannii genomes is approximate 4.4 Mbp. Thus, assembled sequences containing a read length of > 4.4 Mbp were initially excluded from the study, because they might contain contaminant sequences from other species. The genomes were then annotated using Prokka v1.12 [27], and finally, the tRNAs and rRNAs were identified by tRNAscan-SE v2.0 [28,29] and RNAmmer v1.2 [30], respectively.

Pan-genome and phylogenetic analysis
The pan-genome of our 221 CRAB isolates was analyzed using Roary v3.13.0 [48], with a 95% minimum blastp identity and a 99% core definition threshold. Then, we called SNPs of core genes to reduce the computational complexity for phylogenetic tree construction using SNP-sites v2.4.1 [49]. Afterward, a phylogenetic tree was then built by raxmlHPC-PTHREADS v8.2.12 with the neighbor-joining method using 1000 bootstraps [50]. Visualization of the phylogenetic tree was performed using Geneious R10.26 [51] and Phandango website (https://jameshadfield.github.io/phandango/) [52]. A pan-genome frequency plot, a piechart of the pangenome, and a presence and absence matrix against a phylogenetic tree were created using roary_plots script (https://github.com/san-ger-pathogens/Roary/tree/master/contrib/roary_plots). In addition, we also performed the pan-genome analysis of our CRAB genomes compared to previously published genomes from Thailand [16]. The phylogenetic trees were constructed based on the SNPs of core genes and accessory genes, respectively, with the neighbor-joining method using 1000 bootstraps.

Genome assembly quality
In the 221 CRAB isolates, de novo assembly yielded genome lengths from 3,777,937 bp to 4,319,283 bp, with an average of 3,930,367 bp. The number of contigs ranged from 26 to 193, with an average of 68. The GC-content varied from 38.68% to 39.10%, with an average of 38.91%. The N50 and L50 values of the 221 assembled genomes ranged from 45,197 bp to 444,207 bp having an average of 175,941 bp and 3 to 28 having an average of 9, respectively. The details of the assembly quality are given in Table S2.

Sequence types (STs) and antimicrobial resistance (AMR) determinants
In the MLST results of the 221 CRAB isolates (Table S3), ST2 had the largest frequency having been identified in 119 (53.85%) isolates, followed by ST164 (n = 29; 13.12%), ST374 (n = 18; 8.14%), ST16 (n = 13; 5.88%), ST215 (n = 12; 5.43%), and so forth. However, STs could not be assigned for 2 (0.90%) isolates, PSU043 and PSU114. Among the 7 housekeeping genes of A. baumannii, using MLST (Pasteur), the alignment of PSU043 genome with allele 8 of the recA gene showed 99.73% identity and 100% coverage. This PSU043 contained 1 nucleotide substitution (A to G) at position 99 of the recA gene sequence, resulting in a non-identified ST, while PSU114 provided 100% identity and 100% coverage in alignment with all 7 genes. However, when the locus combination of allele 181 of the gltA gene and the alleles of the 6 other genes were analyzed, according to the MLST allelic profile, PSU114 could not be assigned to an ST.

Virulence-associated genes
The investigation of the virulence factors of Acinetobacter spp. in the CRAB isolates revealed the presence of virulence genes associated with adherence, biofilm formation, enzyme, immune evasion, iron uptake, regulation, and serum resistance. All isolates possessed genes encoding OmpA (outer membrane protein A), AdeFGH efflux pump, PANG (poly-N-acetylglucosamine), LPS (lipopolysaccharide), BfmRS (regulation of biofilm formation), and PbpG (penicillin-binding protein). A total of 220/221 (99.55%) isolates (except SK066) harbored all acinetobactin genes. Twenty-six to 29 virulence genes were identified in 220 of the isolates, while only 10 virulence genes were detected in SK066. Among the capsule-encoding genes that are responsible for immune evasion, all isolates carried ACICU_0071 (ATPase gene), ACICU_0092 (phosphomannomutase gene), and pgi genes.  Table S8.

Bacteriocin-encoding gene and bacteriophage genomes
In this study, we also explored the presence of bacteriocin and bacteriophages. For bacteriocin identification, the study showed that all CRAB isolates carried the zooA gene encoding zoocin A. In The bacteriocin-encoding gene and bacteriophage genome information is given in Fig. S2, Table S9.

Pan-genome and phylogenetic analysis
The results of pan-genome analysis among the study 221 CRAB isolates showed that 9,318 (68.68%), 2,412 (17.78%), 1,615 (11.90%), and 222 (1.64%) of 13,567 pan genes were identified as cloud, core, shell, and soft-core genes, respectively. These CRAB isolates contained various genes associated with transcriptional regulators and transporters in 3.96% and 4.19% of the pangenome, respectively. Notably, we also found genes encoding transposases and bacteriophage proteins (e.g., heads, capsids, and tails) in 1.30% and 0.51% of the accessory genome. Furthermore, genes encoding hypothetical proteins were also observed in 5.97% and 66.44% of core and accessory genomes, respectively. Overall, the phylogenetic tree could be divided into many clades according to STs, as illustrated in Fig. 6 and S3. Among these clades, ST1479 was in the same clade as ST164. For two isolates that could not be assigned to an STs, one isolate was grouped into the ST374 clade, while the other isolate was located in the ST16 clade. Interestingly, we noticed that two isolates, SK066 and SK070, which were very different in terms of gene presence and absence, belonged to the different clades, as shown in the red box of Fig. S4. According to the gene presence and absence matrix (Fig. S4) and pan-genome graph (Fig. S5), the pan-genome profiles of these CRAB isolates seem to provide an open pan-genome with vast genomic diversity (Fig. S5).

Discussion
The rapid increase of CRAB infections seriously threatens the global population. Since most CRAB isolates resist many potential antimicrobial classes, it is currently difficult, if not impossible, to manage CRAB infections, especially through the use of suitable antibiotics. Faced with these challenges, many scientists across the world are trying to characterize and understand the mechanisms of antimicrobial resistance in CRAB isolates. We therefore used WGS as a comprehensive method for studying and exploring the genetic basis of 221 CRAB isolates collected at hospitals in Southern Thailand.
According to the results of this study, the largest proportion of CRAB isolates was identified in ST2, which belonged to international clone 2 (IC2) [53,54]. ST2 is the most predominant type and the most widespread in many parts of the world, especially Thailand [55][56][57]. The prevalences of other STs such as ST164, ST374, ST16, ST215, ST25 (IC7), ST129 (IC2), and ST1 (IC1) are different in each region [55,56,[58][59][60]. Hamidian and Nigro (2019) reported the level ST of 3575 CRAB isolates showing ST2 as the most prevalent ST, followed most prominently by ST1 (IC1), ST76, SLV2 (a single-locus variant of ST2), ST25 (IC7), ST10, and then others with lower prevalences [56]. Here, we also noticed that 6 isolates belonged to ST1479, a new ST that was recently discovered in a clinical isolate of extensively drug-resistant A. baumannii (XDRAB) from Thailand [61]. Unfortunately, two isolates could not be assigned to an ST by PubMLST in this study. PSU043 contained one nucleotide substitution (A99G) in the recA gene, which is a silent mutation that does not cause an alteration of the amino acid. Therefore, we predicted that this mutation may not change the function of the RecA protein, and the nearest ST identified by the PubMLST was ST16. For PSU114, although there were no mutations in all seven housekeeping genes, we still could not identify the ST.
In the detection of AMR genes, a majority of the CRAB isolates possessed the bla OXA-23 gene, which is the distinctive class D carbapenemase-encoding gene in A. baumannii. It has been reported in many countries all over the world including the USA, Australia, Germany, Brazil, China, Korea, Thailand, Vietnam, Malaysia, Pakistan, and Egypt [1,3,[15][16][17][18]55,56,62,63]. Normally, oxacillinase (OXA) enzymes have a weak hydrolyzing activity and are poorly expressed, resulting in a low level of carbapenem resistance [56,64]. Nevertheless, the bla OXA expression can be enhanced by an IS located upstream of the bla OXA genes, leading to a high level of carbapenem resistance [56,65]. As to the results of our study, ISAba1 was identified in 56.04% of the bla OXA-23 -positive CRAB isolates. Many earlier reports have demonstrated that ISAba1 is generally located upstream of the bla OXA-23-like gene, and it provides a strong promoter that drives the expression of the bla OXA-23-like gene [63,[66][67][68]. In addition, we found other bla OXA variants such as the bla OXA-58 gene. The OXA-58 enzyme has been reported with highlevel resistance to carbapenem in A. baumannii [69]. More than half of our CRAB isolates also carried the bla OXA-66 gene, and some of the isolates harbored the bla OXA-91 , bla OXA-259 , and bla OXA-402 genes, and so forth. These genes are member of the bla OXA-51-like genes, the intrinsic oxacillinase genes with low-level carbapenemase activity that naturally occur and are located on the chromosome of A. baumannii [62,65,70,71].
Computational and Structural Biotechnology Journal 20 (2022) 545-558 bla PER-7 ) and other b-lactamase genes (e.g., bla AmpC , bla ADC-6 , bla ADC-30 , bla ADC-73 , bla ADC-79 , bla CARB-16 , ble, etc.). The bla AmpC and bla ADC genes are the class C b-lactamase genes, while the bla CARB gene is a class A b-lactamase gene. Overproduction of AmpC b-lactamase in combination with ESBLs, efflux pumps, and/or porin loss has been associated with carbapenem resistance in Gram-negative bacteria [75,76]. Besides b-lactam resistance genes, the other genes that probably confer resistance to aminoglycoside, fosfomycin, tetracycline, chloramphenicol, trimethoprim, sulfonamide, lincosamide, and macrolide were found in these CRAB isolates and they might be classified as multidrug-resistant (MDR) isolates accordingly. Many studies have reported on the retained susceptibility to aminoglycosides (e.g., amikacin and gentamicin) and tetracyclines (e.g., doxycycline and tigecycline) in carbapenem-resistant Gram-negative bacteria (CR-GNB) [74,[77][78][79]. Unfortunately, although aminoglycosides and tetracyclines are used as a monotherapy, or in combination with other antimicrobial agents against CR-GNB infection [77,80], aminoglycosides seem more effective in carbapenem-resistant Enterobacteriaceae (CRE) than in CRAB, while tetracyclines are normally used to combat both CRE and CRAB [80]. The tigecycline resistance is continuously found in CRAB [78]. Notably, although low-level resistance to plazomicin, a novel aminoglycoside, has been reported in CRAB [80], we found that over a haft of the CRAB isolates carried the armA gene, which has been reported as confer high-level resistance to aminoglycosides, particularly gentamicin, amikacin, tobramycin, and plazomicin [81]. Our CRAB isolates also harbored other aminoglycoside resistance genes (aph(6)-Id, aph(3 0 )-Ia, aadA1, aac(3)-IId, ant(2 00 )-Ia, aac(6 0 )-lb, armA, etc.) and tetracycline resistance genes (tet(B) and tet(39)), a finding concordant with previous studies [82,83]. As well, previous studies demonstrated good efficacy of fosfomycin in combination with other antibiotics against CR-GNB [74,84,85]. However, our findings showed that the abaF gene, a fosfomycin resistance gene, was seen in almost all CRAB isolates. This would indicate that aminoglycosides, tetracyclines, and fosfomycin should be used with caution for the treatment of CRAB infections. In addition, the mphE and msrE genes were present in a high number of the CRAB isolates. These genes commonly associated with MDR in A. baumannii, which confer resistance to macrolide by inactivation (mphE) and modification of the target site (msrE) [86,87]. In our study on the dissemination mechanisms of the AMR gene among Gram-negative bacteria, we investigated the MGEs (plasmids, IS elements, and integrons). The findings showed that the largest number of CRAB isolates harbored the RepAci1 and RepAci7 plasmids. The RepAci1 plasmid is one of the most widespread plasmids in Acinetobacter spp. [88]. Towner et al. (2011) reported that the bla OXA-23-like and bla OXA-58-like genes were associated with the carriage of the repAci1 replicase gene located on the RepAci1 plasmid [71]. Surprisingly, here, we found a high prevalence of RepAci7 plasmid in these CRAB isolates, which has not been reported in any other countries. Loraine et al. (2020) previously analyzed the genomics of A. baumannii isolated from three hospitals located in Central and Southern Thailand. They found the RepAci1, RepAci6, and RepApAB49 plasmids to have a high frequencies, while the RepAci7 plasmid was not detected in any A. baumannii isolates. Additionally, plasmid pA297-3 was present in all ST25 isolates. This plasmid has been associated with the AMR spread in Acinetobacter spp., particularly in ST25 isolates [89]. Among the 61 ISs detected in our study, we found that ISAba22 was the most prevalent IS, followed by ISEc29, ISAba26, ISVsa3, ISAba1, and ISAba24. Besides functioning as the promoters for the expression of many AMR genes, the ISs generally provide for a cut-and-paste mechanism of transposition [90]. Previous studies have shown that the bla OXA-23 gene is mostly located in a composite transposon Tn2006 that is bracketed by two copies of ISAba1 [88,90]. In terms of integrons, we found an integron-integrase gene (IntI1) in 34 CRAB isolates. The AMR genes, especially aminoglycoside resistance genes (aac(6 0 )-Ib and aac(3)-Ia), were present in the integron-encoded IntI1 integrase. The SK066 isolate carried the bla -IMP-14 gene in the class 1 integron, similar to many other reports [91,92]. According to the MGE results, we indicate that if the AMR genes locate on the MGEs, these particular genes might be horizontally transferred (conjugation) to other Acinetobacter spp. as well as other species of Gram-negative pathogenic bacteria. Thus, the MGEs are a significant factor in the acquisition and spread of AMR genes [93].
In addition to AMR, pathogenic bacteria have evolved and developed virulence to host-defense mechanisms [94,95]. We found the virulence-associated genes encoding for many virulence factors (e.g., adherence, biofilm formation, enzyme, immune evasion, iron uptake, regulation, and serum resistance) in all CRAB isolates. This finding could imply that the presence of these virulence genes may increase the pathogenicity of these CRAB isolates and the severity of infection [95]. Importantly, they might be spread to other bacteria through horizontal gene transfer, similar to the AMR genes [94]. Furthermore, we found the zooA gene encoding zoocin A in all isolates. Zoocin A, a bacteriocin-like inhibitory substance (BLIS), was first identified in Streptococcus equi subsp. zooepidemicus strain 4881 [96][97][98]. It is a peptidoglycan hydrolase that is responsible for inhibiting the peptidoglycan synthesis in many other streptococcal species, especially S. mutans, S. sobrinus, and S. cricetus [96][97][98]. Importantly, it is also classified as a penicillin-binding protein (PBP) [96], which provides weak blactamase activity against penicillin. This might be one of the factors causing high-level b-lactam resistance in the CRAB isolates. In the investigation of the bacteriophage genome, a very high level of CRAB isolates showed sequence alignment to Siphoviridae phage, while some isolates also harbored Myoviridae and/or Podoviridae phages. These three phage families belong to the order Caudovirales, which contains double-stranded DNA (dsDNA) genomes [99]. We thus postulate that the CRAB isolates might have been previously infected by these particular phages. More importantly, the DNA phages can drive the bacterial genes, especially AMR genes, to other bacteria by generalized transduction [100,101]. Thus, the discovery of these bacteriophage genomes within the CRAB genomes may be indicated as one of the factors causing the spread of AMR genes or other genes.
To demonstrate the genomic diversity, we analyzed the pangenome among our 221 CRAB isolates. The pan-genome profiles showed that these CRAB isolates shared 17.78% core genes, and contained a very high percentage (68.68%) of cloud genes. In a presence and absence matrix of pan-genome, the accessory genome demonstrated the presence of genes encoding transposases and bacteriophage proteins (e.g., head, capsid, and tails), and the high level of genes encoding hypothetical proteins in these CRAB isolates. We hypothesized that these genes are probably involved in adaptation mechanisms, particularly the acquisition of AMR and virulence genes and the ability to persist in some changing environments [102,103]. Based on our analysis, the pan-genome graph could be possibly considered as an open pan-genome since our dataset contained only 221 isolates. However, in the analysis of a larger dataset of A. baumannii isolates, the result also exhibited that exponential pan-genome growth was observed when increasing a great number of pan-genome [103][104][105][106]. Furthermore, the high proportion of accessory genes demonstrated a high genomic diversity among these isolates. This discovery could be an indicator of a useful path for researchers to explore and perhaps find new genes and to study genomic diversity in the A. baumannii strains, especially carbapenem-resistant and MDR isolates. Notably, two isolates (SK066 and SK070) in our study were uncommon and very different from the other CRAB isolates. However, after submitting the genome sequences of 221 isolates into the NCBI server, the submission details reported that only 219 isolates were CRAB, whereas the other 2 isolates belonged to other Acinetobacter species. The average nucleotide identity (ANI) results from NCBI revealed that the SK066 and SK077 isolates were identified as Acinetobacter pittii (97.42% identity) and Acinetobacter nosocomialis (97.85% identity), respectively. These two isolates were previously confirmed as A. baumannii by both in vitro and in silico methods and also carried carbapenem resistance genes as well as other AMR genes. This could indicate that these genomes are biologically related among Acinetobacter spp. To further elucidate the genomic data of these two isolates, additional tools and technologies such as long-read WGS should be considered to provide more information in the future. In addition, the pan-genome analysis of our CRAB genomes compared to previously published genomes demonstrated that the specific features could not be observed among the accessory genomes. These findings could be indicated that genetic features of the A. baumannii clinical isolates from Thailand are closely related.
WGS provides more accurate details and more precise information than traditional microbiological methods; for example, in the comparison of antimicrobial susceptibility patterns obtained from antimicrobial susceptibility testing (AST) and bacterial DNA fingerprints obtained from pulse-field gel electrophoresis (PFGE) [107]. These strengths of this method allow scientists to compare entire genome sequences within the bacterial cell, for a better understanding of the epidemiology of the pathogens such as identifying the environmental source of an outbreak, transmission events, and the mechanism for spreading antimicrobial resistance [108][109][110].
Since the presence or absence of MGEs carrying AMR genes results in genetic variation and changes in antimicrobial susceptibility patterns, WGS can rapidly provide crucial data during an acute outbreak [111]. Therefore, using WGS to study the genomic insights into the pathogens, especially antibiotic-resistant strains, could be beneficial for outbreak investigations and surveillance as well as infection control and prevention.

Conclusions
This study revealed significant information from a short-read WGS analysis. All CRAB isolates possessed various AMR genes, MGEs (especially plasmids and ISs), and virulence-associated genes, which can be horizontally transferred to other pathogenic bacteria causing widespread carbapenem resistance. The bacteriocin gene and the bacteriophage genomes were present at the highest frequency in these CRAB isolates. Finally, a way to rapidly identify and characterize the genomic features of the CR-GNB strains is necessary before we will be able to finally control the spread of these pathogens in the future.

Data availability statement
The assembled genomes of all 221 CRAB isolates have been deposited in the NCBI GenBank under BioProject number PRJNA752484 with BioSample numbers SAMN20599216 to SAMN20599436.

Funding
This research was funded by the National Science and Technology Development Agency (NSTDA), Thailand (project number: P-20-51325) and the Graduate Scholarship, Faculty of Medicine, Prince of Songkla University, Thailand (grant number: 62-012). This study was also supported by the Faculty of Science, Prince of Songkla University, Thailand (grant number: SCI64040135).

Ethical approval
The study was approved by the Human Research Ethics Committee (HREC) of Prince of Songkla University (reference number: 64-284-14-1, date of approval: 9 June 2021).

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.