Accurate identification of clinically important Acinetobacter spp.: an update

Acinetobacter species have emerged as one of the most clinically important pathogens. The phenotypic techniques which are currently available are insufficient in accurately identifying and differentiating the closely related and clinically important Acinetobacter species. Here, we discuss the advantages and limitations of the conventional phenotypic methods, automated identification systems, molecular methods and MALDI-TOF in the precise identification and differentiation of Acinetobacter species. More specifically, several species of this genus are increasingly reported to be of high clinical importance. Molecular characterization such as of blaOXA-51-like PCR together with rpoB sequencing has high discriminatory power over the conventional methods for Acinetobacter species identification, especially within the Acinetobacter calcoaceticus–Acinetobacter baumannii complex.

relevant species, responsible for causing 80% of infections [7,8]. A. pittii and A. nosocomialis are also considered clinically important and have been implicated in causing community and hospital-acquired infections [4,5,8]. The two novel species mentioned previously (A. seifertii and A. dijkshoorniae) were also isolated from human clinical specimens [4,5]. A. calcoaceticus is considered to be a nonpathogenic environmental organism isolated from soil and is rarely involved in causing diseases [4,5].
Acinetobacter species were found to be associated with increased mortality [9]. The ACB complex has emerged as a high priority nosocomial pathogen, especially in intensive care units, thereby posing a challenge to infection management practices [10].
However, identification of Acinetobacter to the species level remains complicated. Clear differentiation among various Acinetobacter species using standard biochemical methods and automated systems are challenging, thus leading to inaccurate identification [9]. Although various molecular methods are available, they are not regularly employed in routine diagnostic laboratories [11].
Accurate identification of Acinetobacter is essential as these nosocomial pathogens, A. baumannii, A. pittii and A. nosocomialis, differ in their biological and pathological characteristics [12,13]. In addition, differences have also been observed among these pathogens in their ability to colonize the skin, antimicrobial susceptibility and mechanisms of antimicrobial resistance [12,13]. A. baumannii is more predominantly isolated from patients with poor outcome and increased resistance to antimicrobial agents when compared with other pathogens within the ACB complex [12,13]. A. baumannii is responsible for causing bacteremia, ventilator-associated pneumonia, urinary tract infections and skin and soft tissue infections. In 2017, Liu et al. reported A. baumannii to have a high mortality rate, of 29.8-36.9% [14]. Jain et al. reported that A. pittii as the most frequently isolated species with a lower rate of mortality in children [9]. Recent studies from other countries have shown that A. pittii and A. nosocomialis were responsible for 29% of ACB complex bacteremia in the USA, 24-45% in Taiwan, 50% in Korea and 66% in Norway [14].
A study by Lee et al. in 2013 clearly demonstrates important differences in the clinical features and outcome of pneumonia caused by A. baumannii and A. nosocomialis [15]. Patients infected with A. baumannii were more likely to have severe illness, increased antimicrobial resistance and higher mortality when compared with A. nosocomialis, supporting the concept that the pneumonia caused by A. baumannii and A. nosocomialis should be considered as two different clinical entities [15]. Therefore, it is important to have species-level identification of the ACB complex due to differences in epidemiology, antimicrobial resistance patterns and clinical outcomes, and reporting of infection from the ACB complex without further speciation may be considered inadequate, leading to inferior clinical care and inappropriate use of antibiotics [16].

Role of phenotypic methods
The identification scheme for Acinetobacter species, first proposed in 1986 by Bouvet and Grimont, was based on 28 phenotypic tests -physiological, nutritional and enzymatic tests [17]. Further refinement came in 1987, which includes growth at different temperatures (37, 41 and 44 • C), gelatin hydrolysis, assimilation of 14 carbon sources and acid production from glucose [18]. Based on the growth characteristics of Acinetobacter spp. at different temperatures, the scheme facilitated the differentiation of most clinically relevant isolates as A. baumannii , A. nosocomialis, A. pitti and A. calcoaceticus. [18,19].
The species-level identification of Acinetobacter is not consistent and reliable with the existing phenotypic methods [6]. Recently, the extension of the Bouvet and Grimont phenotypic system was proposed by Nemec, and includes growth at different temperatures (32,35,41 and 44 • C), glucose acidification, hemolysis, gelatinase production and growth on 36 single carbon and energy sources (http://apps.szu.cz/anemec/anemec.htm). A list of phenotypic tests for the identification of ACB complex is detailed in Table 1.

Role of automated bacterial identification systems
Various commercial automated systems for identification are available and widely used in clinical microbiology laboratories [20,21]. These automated systems are based on different technologies such as measurement of growth of the microorganisms using biochemical or physical parameters, CO 2 production, color, turbidity, changes in fluorescence due to microbial consumption of biochemical and carbohydrate substrates in media, production or consumption of heat, changes in pressure due to microbial consumption and bacterial detection by viability staining and metabolic activity [20,21]. The commonly used automated commercial systems for the species-level identification Glucose acidification − + V (+) + + +

Assimilation of:
Adipate of Acinetobacter in the clinical laboratories includes API 20NE, VITEK 2, Phoenix, Biolog, MicroScan WalkAway and Accelerate Pheno™ system [18]. The API 20NE is available with microtubes and has dehydrated media and substrates. The end products of the test are detected by an indicator or addition of reagents. The API 20NE, manual method from bioMerieux compared with conventional biochemicals in the identification of 198 isolates has reported an overall agreement of 92% [19,20]. With the updated database of API 20NE (version 5.1), Bernards et al. achieved 87% accurate identification [2]. The Biolog system for automated identification of bacteria is based on the redox system [22]. Bernard and his colleagues compared strain groupings of Acinetobacter by Biolog cluster analysis with DNA-DNA hybridization (DDH) studies and found 84.5% of the strains were correctly identified at the genus level by the Biolog system [23]. The BD Phoenix system involves the use of an oxidation-reduction indicator and turbidometric growth detection, whereas the MicroScan WalkAway system uses photometer or fluorometer to determine the growth development. The VITEK 2 system utilizes the growth based technology for microbial identification. Wang and colleagues have reported that VITEK MS method was able to accurately identify three species (A. junii, A. hemolyticus and A. johnsonii) but failed to discriminate the species within the ACB complex [24]. Such results are expected because the clinical identification system depends on the database already exists in the automated systems that are not comprehensive [18]. In addition, the substrates used for species identification have not been designed specifically for the identification of Acinetobacter [18].
Other the US FDA-cleared automated identification systems designed to identify the organisms directly from positive blood culture bottles include the Verigene system (Nanosphere) gram negative blood culture (BC-GN) nucleic acid tests and the FilmArray blood culture identification panel (Biofire diagnostics) [25,26]. The Verigene system involves purification of nucleic acids and hybridization to oligonucleotide gold nanoparticles on microarray [21]. In a study by Kim et al., Gram-negative bacteria were cultured from 150 samples, of which 22 were Acinetobacter spp and all were accurately identified by the Verigene [26]. The filmArray identification system involves extraction of DNA followed by PCR and melt curve analysis [21]. Phenotypically, A. baumannii was identified as a member of the ACB complex followed by bidirectional sequencing of the 16S rRNA gene. After comparison, the FilmArray blood culture identification detected all the 51 culture isolates as A. baumannii [27]. The Accelerate Pheno system, a fully automated system, has the capability of performing identification and antimicrobial susceptibility testing directly from positive blood cultures [28]. The Accelerate Pheno system combines gel electrofiltration and a broad panel of fluorescence in situ hybridization for bacterial identification and morphokinetic cellular analysis, which measures the response of individual cells and colonies to antibiotics over time using time-lapse imaging for antimicrobial susceptibility testing [29]. The overall identification performance has been reported with the sensitivity level of 98.6% and specificity of 99.7% with a positive predictive value and a negative predictive value of 100% for A. baumannii [29].
Snyder and colleagues have compared the BD Phoenix system with MicroScan WalkAway and showed 100% concordance for the identification (ID) of nonfermentative Gram-negative bacilli (NFGNB) [30].

Molecular identification of Acinetobacter species
For the molecular identification of Acinetobacter species, several methods have been developed, studied and shown to be adequate. This includes DNA-DNA hybridization (DDH), DNA sequence-based analysis and restriction analysis of PCR amplified DNA sequences (PCR-restriction fragment length polymorphism [PCR-RFLP]). The following sections briefly describe each of the methods.
DNA-DNA hybridization DDH is a gold standard method where the whole genome of the organisms is compared for delineation of bacterial species. Johnson and colleagues assigned Acinetobacter strains to five distinct DNA groups with four strains ungrouped [31]. Bouvet and Grimont distinguished 12 genospecies [32]. Bouvet and Jeanjean described genomic species 13-17 [33]. A parallel study to Bouvet and Jeanjean describing three new genomic species 13-15 was reported by Tjernberg and colleagues [34]. In 1988, the genomic sp. 12 was described as A. radioresistens [35].
Although DDH has been considered a gold standard for bacterial species differentiation over the past 50 years, due to several disadvantages like a labor-intensive and error prone nature, genotype-based technology was needed [36]. As whole-genome sequencing is available, numerous overall genome relatedness indices (OGRI) were developed to replace DDH [37].
The OGRI is digital DDH-like similarity indices based on computational comparisons of two genome sequences [36]. Various OGRI are available, including average nucleotide identity (ANI), genome BLAST distance phylogeny and maximal unique matches index [36]. Among these, ANI is most widely used and considered as the gold standard method for species demarcation [36,37]. Generally, ANI values of 95-96% are accepted for species delineation, corresponding to the traditional 70% DDH threshold [36]. To calculate the reciprocal ANI values, a new algorithm called OrthoANI (ANI by orthology) is available [37].
DNA sequence-based analysis DNA sequence-based analysis is used to understand the relatedness of organisms from their DNA sequence similarity. Various DNA sequences like 16S rRNA, rpoB, 16S-23S intergenic spacer (ITS) have been studied in classification and identification of Acinetobacter.
16S rRNA & rpoB sequencing 16S rRNA sequencing is the most commonly used gold standard technique for unknown identification of bacterial species [41]. The 16S rRNA gene of bacteria contains nine hypervariable regions that show sequence diversity among the different bacterial species and are therefore used for species-level identification [42]. The V8 hypervariable region of 16S rRNA is useful in the identification of Acinetobacter at the genus level [43]. Misbah et al. amplified 1500 bp 16S rRNA region using three primer pairs to identify clinical isolates of Acinetobacter species and all were identified as A. baumannii [44] 5 and 97.1%, respectively. Although 16S rRNA sequencing was reported to be polymorphic for the identification of other bacterial genera, it is not effective in differentiating Acinetobacter at the species level due to its extremely low polymorphic nature of the variable region [45]. Due to this limitation, protein encoding genes like rpoB, gyrB and recA, which evolve much faster, were investigated as diagnostic biomarkers. The rpoB gene is 4089 base pair (bp) long and only one copy is present [45]. The rpoB gene has four variable regions which includes two polymorphic regions and two flanking spacers enclosed by conserved sequences [45]. The intraspecies similarity of rpoB for A. baumannii was 98.7-100%, A. nosocomialis 98.7-100%, A. pittii 98.6-100%, A. seifertii 98.5-100%, A. dijkshoorniae 98.6-100% and A. calcoaceticus 97.7-99.7% [5]. The interspecies similarity of rpoB for A. baumannii with A. pittii, A.nosocomialis, A. calcoaceticus, A. seifertii and A, dijkshoorniae were 92-93%, 94-95%, 91-92%, 94-95% and 93-94%, respectively [5].
Recently, Lee and colleagues reported, 93.4 and 98.2% accuracy for the species-level identification of Acinetobacter by 16S rRNA and rpoB sequencing, respectively [46]. Similarly, Khosravi's group revealed that 16S rRNA sequencing of A. grimontii and A. junii, was ambiguous due to high similarity between these two species, whereas the rpoB gene sequencing showed 100 and 99% similarity to A. grimontii and A. junii, respectively [47]. Initially, Vaneechoutte et al. proposed that A. grimontii as a heterotypic synonym of A. junii by DDH and AFLP [48]. Wang et al. reported that the similarity of rpoB sequences of A. grimontii and A. junii is greater than 99%, thereby suggesting their combination into one species [49]. La Scola's group investigated the efficacy of rpoB gene sequencing for the differentiation and identification of 17 reference strains of Acinetobacter spp. and seven unnamed genome species. The percentage similarities of rpoB gene sequences of different strains varied between 83 and 99% [45]. Furthermore, Gundi and colleague studied the use of partial rpoB sequence analysis (PRBS) on a collection of Acinetobacter spp. by analyzing residues spanning the 350 base pair (bp) hypervariable zone of the rpoB gene [50]. With partial rpoB sequence analysis, it was found that while interspecies similarity were well separated (ranging from 88.3 to 96.9%), the intraspecies similarity of the rpoB gene sequences was found to be 98-100% for A. baumannii, 98-100% for A. nosocomialis, 98-100% for A. pitti and 98.6-100% for A. calcoaceticus [50].
In summary, while 16S rRNA sequencing is highly reliable at the genus level, it has poor discriminatory ability at the species level. In contrast, rpoB genes among Acinetobacter spp. are highly variable and considered to be appropriate for species identification [49].

Restriction analysis of PCR amplified DNA sequences
This method involves amplification and restriction of conserved gene sequences using various restriction enzymes. Several gene sequences such as 16S rRNA, 16S-23S ITS and recA are widely used [51].
In the amplified ribosomal DNA restriction analysis (ARDRA) method, amplification of the 16S rRNA gene followed by RFLP using different restriction enzymes enabled identification of all Acinetobacter species except A. haemolyticus (genomic species 4) and A. johnsonii (genomic species 7) [1]. However, when Nowak et al. used the 16S-23S ITS region, identical PCR and RFLP patterns were observed [52]. Multiplex PCR targeting recA and 16S rDNA genes followed by RFLP yielded reliable identification of Acinetobacter geno species five, seven and ten including the closely related species within ACB complex [53].
Although the above methods are helpful in understanding the clinical relevance and epidemiology of Acinetobacter species, they are too laborious for day-to-day application in the laboratory or field setting. As these techniques involve skilled labor, they are restricted only to reference laboratories.

Sequence-based analysis of other gene targets
The other sequence targets available for the identification and discrimination of Acinetobacter species are summarized in Table 2. A multiplex PCR targeting the recA and 16S-23S rRNA ITS region for the identification of Acinetobacter species within the ACB complex was developed using 22 reference strains and 138 clinical isolates of Acinetobacter spp. The multiplex PCR was found to be 100% accurate in identification of Acinetobacter at the species level, which was later confirmed by sequence analysis [54]. In 2013, Soo and colleagues developed a loop-mediated isothermal amplification (LAMP) method for identification of A. baumannii using the 16S-23S rRNA ITS region. The sensitivity of the LAMP assay was found to be 100-fold superior to the simple multiplex PCR method in identifying Acinetobacter species [55]. However, the main drawback is that the LAMP method has not been studied on different clinical isolates.
Another reliable marker for identification of A. baumannii is the bla OXA-51 -like gene, which is intrinsic to A. baumannii. The utility of bla OXA-51 -like gene-specific PCR has been studied by Turton and colleagues on 106 A. baumannii isolates and 23 different Acinetobacter spp, which were previously identified by amplified 16S rDNA restriction analysis (ARDRA). All 106 A. baumannii clinical isolates, as well as the ATCC 19606 strain, flagged positive for bla OXA-51 -like gene, whereas isolates that showed negative were identified as another species of Acinetobacter. So far, 95 variants of bla OXA-51-like gene have been described and the possible limitation of this PCR method could be due to its inability to detect all the described variants of the bla OXA-51 gene [56,57]. However, false-positive and false-negative results for detection of the bla OXA-51 gene were reported by a few studies, as mentioned below. False-positive results were observed by Chen's group in two different studies involving 676 and 74 isolates of non-baumannii Acinetobacter species, which showed 0.9 and 5.5% prevalence of the bla OXA-51 -like gene, respectively. The bla OXA-51 -like gene has weak carbapenem hydrolyzing activity. Insertion sequences (ISs) such as ISAba1 could be inserted upstream, leading to overexpression of the bla OXA-51 -like genes [58]. It was reported that there is presence of the bla OXA-51 -like gene in non-baumannii species including A. nosocomialis and other genomic species close to A. nosocomialis. All these isolates contain the upstream insertion of ISAba1 on plasmids, which might have been transferred from A. baumannii by the horizontal gene transfer method [59].
The false negative issue was also observed during an outbreak situation, where approximately 14 isolates generated a 1.7-kb PCR product instead of 353 bp for this marker. Sequencing of the 1.7-kb PCR band revealed disruption of the bla OXA-51 -like gene by the insertion element ISAba19, thereby suggesting the bla OXA-51 -like gene to be an unsuitable molecular marker for A. baumannii identification [60]. A separate study also reported disruption of the bla OXA-51 -like gene by insertion elements like ISAba15 and ISAba19 [61].
Notably, presence of such plasmid-borne insertion elements along with bla OXA-51 -like gene confers high level carbapenem resistance, thereby limiting the use of bla OXA-51 -like gene as a target for differentiating A. baumannii from other Acinetobacter species. Therefore, it is essential to include additional gene targets like rpoB sequencing along with the bla OXA-51 -like PCR for the differentiation of A. baumannii from ACB complex.
Multiplex PCR based on species-specific genes is considered simple, rapid and reliable for identification of different species within the ACB complex. This multiplex PCR is robust, reproducible and cost effective and yields results in less than 2 h. A multiplex PCR based on gyrB primers was proposed for the identification of A. baumannii and A. nosocomialis [62]. This multiplex PCR was expanded to include identification of A. pittii and A. calcoaceticus [63]. A total of 495 clinical isolates of Acinetobacter were subjected to gyrB multiplex PCR and VITEK2 analysis. The accuracy rates were found to be 90.5 and 76.6% for gyrB multiplex PCR and VITEK2, respectively [46]. The gyrB multiplex PCR showed 100% concordance in identifying species within the ACB complex whereas VITEK2 was able to identify only A. baumannii and failed to identify other species within the ACB complex [46].
Although various target regions like recA and gyrB are available for the identification and differentiation of Acinetobacter species within the ACB complex, none of them has been validated for the recently included novel species A. seifertii and A. dijkshoorniae, which can be considered as a limitation.

Role of MALDI-TOF
Matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF) is increasingly used in diagnostic microbiology for the routine identification of bacteria at the genus, species or subspecies level [64,65]. Generally, MALDI-TOF is prepared by coating the sample with an energy absorbent solution called matrix. The matrix is crystallized while drying and the sample within the matrix is also crystallized [66]. The sample is then subjected to a laser beam. Desorption and ionization with the laser beam produces protonated ions from analytes such as protein in the sample, which are accelerated and separated from each other based on their m/z [66,67]. Afterward, the charged analytes are detected and measured using a TOF analyzer. A characteristic peptide-mass-fingerprint (PMF) is generated for analytes in the sample based on the TOF information [66,67]. Microbial identification by MALDI-TOF is performed by comparing the PMF of unknown organism with PMFs available in the database [66].
Bacterial cultures subjected to ethanol-formic acid extraction as well as directly from bacterial colonies were used for identification [68]. Table 3 summarizes various studies carried out for the identification of Acinetobacter spp. using MALDI-TOF.
Espinal's group tested the previously characterized 60 Acinetobacter isolates (18 A. nosocomialis, 17 A. pittii and 18 A. baumannii) and seven reference strains of different Acinetobacter species using MALDI-TOF with Bruker database. It was found that 85% (n = 6) accuracy in identification at species level for the reference strains. Specieslevel identification was accurate and achieved for all A. baumannii (100%, n = 18) isolates and all A. pittii (94%, n = 17), except one. None of the A. nosocomialis were identified by MALDI-TOF as the Bruker database lacks a signature specific for A. nosocomialis. A set of reference spectra was created using three strains each of A. baumannii, A. pittii and A. nosocomialis to supplement the Bruker database. Reanalysis of spectra from all 60 isolates using a local database incorporated to MALDI-TOF was performed with reference signatures of A. nosocomialis, which provided unequivocal identification for all the tested strains [69].
Sedo et al. found an inability of MALDI-TOF to discriminate between A. baumannii and A. nosocomialis within the closely related ACB complex. In their study with a taxonomically clearly defined set of 105 strains representing the four species of the ACB complex, they evaluated a protocol with change in the matrix composition without changing the other steps of the sample processing. With the modified protocol, differentiation between A. baumannii and A. nosocomialis was improved and 13 A. nosocomialis strains which were wrongly assigned were correctly identified [70].
Hsueh et al. evaluated the ability of MALDI-TOF (Bruker Biotyper) in the correct identification of species within the ACB complex as well as species other than the ACB complex. The overall accurate identification rate of species within the ACB complex was only 85.3%.
Although the current database included A. nosocomialis, the correct identification rate was 72.4%. Additionally, accurate identification of non-ACB complex species like A. junii, A. ursingii, A. johnsonii and A. radioresistens was achieved using the Bruker Biotyper [71].
Kishii and colleagues evaluated MALDI-TOF for the identification of Acinetobacter to the species level by comparison with sequence analysis of rpoB using 123 isolates of Acinetobacter species [72]. MALDI-TOF identified 86.2% of Acinetobacter isolates to species level and 13% of the isolates to genus level. With rpoB sequence analysis, 84% were confirmed to the correct species level and discordant results were obtained for 16% isolates. The Table 3. Various studies using MALDI-TOF for the identification of Acinetobacter spp. rpoB sequencing, which is better for species delineation, was able to assign 123 isolates to 13 species whereas MALDI-TOF assigns only to 9 species indicating the less discriminatory power of MALDI-TOF [72]. MALDI-TOF combined with chemometric tools like partial least squares discriminant analysis and hierarchical cluster analysis was reported to have (100%) accuracy in the identification within the ACB complex [73]. Overall the diagnostic efficiency of MALDI-TOF from different studies, showed 100% sensitivity, 91.7-99% specificity and 70-100% concordance with other methods for identification of A. baumannii.

Number of specimens tested
MALDI-TOF technique was unable to identify the two novel species, A. seifertii and A. dijkshoorniae and several studies have shown that the estimated identification rate for A. nosocomialis range at 70% suggesting to update and improve the database for accurate identification of species within the ACB complex [68]. A total of 78 Acinetobacter isolates representing A. baumannii, A. nosocomialis, A. pittii, A. dijkshoorniae and A. seifertii were used to calculate the species-specific average spectra. Mean spectra derived from A. seifertii, A. dijkshoorniae and A. nososcomialis were added to the database [68]. Finally, when spectra from all the 78 isolates were analyzed using the custom database which included both the mean spectra from all the Acinetobacter species in the default taxonomy with the new reference signatures of A. nosocomialis, A. seifertii and A. dijkshoorniae allowed the identification of Acinetobacter to the correct species level within the ACB complex [68].
Variations in the sensitivity and specificity in identifying ACB complex by MALDI-TOF have been reported [68]. This is due to the methodological differences followed in each of the study. High rates of >98% of sensitivity and specificity were obtained when the bacterial colonies were subjected to a modified extraction protocol using formic acid and acetonitrile, compared with the conventional method which yielded a sensitivity and specificity of 91 and 98%, respectively [68,69]. Though both the protocols have used total mass for identification, the methodological variation of sample preparation impacts on the overall sensitivity and specificity rates.

Conclusion
Species within the ACB complex are difficult to distinguish. The differences in clinical relevance of the ACB complex and their antimicrobial susceptibility profile warrant species-level identification. The traditional phenotypic methods such as culture and biochemical identification are slow, unreliable and less efficient at differentiating the most clinically important A. baumannii within the ACB complex. PCR targets such as the bla OXA-51 -like gene, 16S-23S rRNA ITS, recA and gyrB; and PCR-RFLP of 16S rRNA, 16S-23S ITS and recA were considered as suitable markers for Acinetobacter speciation. However, these methods are mostly performed by reference laboratories. With an accurate database, MALDI-TOF can be helpful for species-level identification. To conclude, molecular characterization using bla OXA-51 -like PCR together with rpoB sequencing has high discriminatory power over the conventional methods for Acinetobacter species identification, especially within the ACB complex.

Future perspective
Identification of Acinetobacter species within the ACB complex is still challenging. Currently, available phenotypic methods are insufficient. Although various commercially available automated identification systems are widely used in clinical laboratories, such systems are still not effective for identification. Incorporation of more diverse mass spectra of different species within the ACB complex will improve the MALDI-TOF database, which eventually will be useful in accurate identification of the ACB complex at the species level. For molecular characterization, a single gene target as a biomarker is not sufficient for accurate identification of ACB complex. Whole-genome sequencing can be considered as a promising tool; however, its usage is limited in lower resource settings. Furthermore, comparative genomic studies will help us to identify the species-specific targets that can be used as a marker for molecular identification.

Financial & competing interests disclosure
The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.
No writing assistance was utilized in the production of this manuscript.

Author contributions
S Vijayakumar reviewed the literature and drafted the manuscript. B Veeraraghavan and I Biswas reviewed and edited the manuscript. All authors read and approved the manuscript.
future science group www.future-science.com

Summary points
• Identification of Acinetobacter species within the closely related Acinetobacter calcoaceticus-Acinetobacter baumannii complex is crucial.
• The existing phenotypic methods and commercial automated systems are not reliable for the species-level identification of Acinetobacter.
• At the genus level, 16S rRNA sequencing is reliable whereas rpoB sequencing is considered to be appropriate for species level identification.
• bla OXA-51 -like gene is considered to be the species-specific biomarker for Acinetobacter baumannii.
• Presence of plasmid-borne insertion elements along with bla OXA-51 -like gene confers carbapenem resistance, thereby limiting the use of bla OXA-51 -like gene as a target for differentiating Acinetobacter baumannii from other Acinetobacter species.
• MALDI-TOF with an accurate database will be useful in the species-level identification within the A. calcoaceticus-A. baumannii complex.

Open access
This work is licensed under the Creative Commons Attribution 4.0 License. To view a copy of this license, visit http://creativecomm ons.org/licenses/by/4.0/ 0:bold /0:bold