In a laboratory in Addis Ababa, researchers have taken a genetic census of one of medicine’s most feared adversaries, and the findings read like a field guide to the global superbug crisis. Scientists at the Ethiopian Public Health Institute sequenced the whole genomes of 30 Acinetobacter baumannii isolates held at the country’s National Clinical Bacteriology Reference Laboratory and found that virtually every strain was a multidrug-resistant organism capable of shrugging off the antibiotics doctors reach for when little else works. Twenty-nine of the 30 isolates were phenotypically resistant to carbapenems, the last-line drugs reserved for the gravest hospital infections, and nearly all of the resistant strains carried genes encoding NDM-1 or OXA-23, two enzymes at the top of antimicrobial-resistance watchlists worldwide. The dominant lineage belonged to International Clone 2, a bacterial pedigree that has spread through intensive care units on every inhabited continent. Published in BMC Genomics on 29 August 2026, the study delivers one of the most detailed genomic portraits yet of drug-resistant A. baumannii in Ethiopia.

Acinetobacter baumannii is an opportunistic Gram-negative bacterium that has made hospital environments its evolutionary home. It survives on dry surfaces for weeks, tolerates many common disinfectants, and colonizes ventilators, catheters and sinks, from which it can invade the lungs, bloodstream, wounds and urinary tract of vulnerable patients. Infections cluster in intensive care units, where the bacterium exploits breaks in the skin and defenses of the airway. What makes it formidable is not a single toxin but an unrivaled capacity to accumulate resistance: the species swaps genes with neighbors, mutates under antibiotic pressure and shields itself behind a polysaccharide capsule. When strains become resistant to carbapenems — broad-spectrum beta-lactam antibiotics long considered the final reliable option for serious A. baumannii infections — clinicians are left with combination therapies whose toxicity often rivals their benefit. The World Health Organization has therefore placed carbapenem-resistant A. baumannii at the top of its critical-priority list, the bacteria for which new treatments are most urgently needed.

The Ethiopian study began not at a bedside but in a biobank. The 30 isolates had been collected and archived by the National Clinical Bacteriology Reference Laboratory, the country’s central facility for confirming difficult clinical cultures, and the analysis received a waiver of individual informed consent from the Ethiopian Public Health Institute’s Institutional Review Board because the de-identified archival strains involved no direct patient contact. Each isolate underwent conventional antimicrobial susceptibility testing interpreted against Clinical and Laboratory Standards Institute breakpoints, followed by whole-genome sequencing, which reads the organism’s complete DNA blueprint. The team, led by Abebe Aseffa Negeri of the Ethiopian Public Health Institute and spanning the institute’s bacterial disease, genomics and bioinformatics divisions, set out to address a conspicuous gap: although drug-resistant A. baumannii is a growing problem in East African hospitals, high-resolution genomic data from the region have been scarce, leaving officials largely blind to which lineages circulate locally and which resistance mechanisms they carry.

The phenotypic results were stark. Every one of the 30 isolates qualified as multidrug-resistant, a formal designation meaning the bacterium is nonsusceptible to at least one agent in three or more antimicrobial classes. Twenty-nine of the 30 — 96.7 percent — were resistant to carbapenems, placing them in the notorious category abbreviated CRAB, for carbapenem-resistant A. baumannii. For clinicians, such numbers translate directly into constrained choices: infections caused by these strains cannot be treated with standard carbapenem regimens and must instead be managed with older, more toxic drugs such as polymyxins, or with newer agents that are expensive and often unavailable in resource-limited settings. The near-universal resistance observed in a single reference laboratory collection suggests that carbapenem-resistant lineages are no longer sporadic arrivals in Ethiopian healthcare facilities but established residents, quietly reproducing and evolving within hospital wards.

Whole-genome sequencing explained why. Twenty-eight of the 29 carbapenem-resistant isolates — 96.6 percent — carried acquired carbapenemase genes, the blueprints for enzymes that chemically destroy carbapenem antibiotics. The most prevalent was bla_NDM-1, found in 20 of the 29 resistant isolates. NDM-1 is a metallo-beta-lactamase, a class B enzyme that uses zinc ions at its active site to hydrolyze the beta-lactam ring at the heart of penicillins, cephalosporins and carbapenems alike; because its catalytic strategy differs from that of other beta-lactamases, most conventional inhibitor drugs are powerless against it. Seventeen isolates carried bla_OXA-23, a class D oxacillinase that inactivates carbapenems through a covalent acyl-enzyme intermediate and is considered a hallmark of International Clone 2. Many strains harbored both genes, stacking redundant enzymatic defenses. Beyond the carbapenemases, the genomes packed a diverse arsenal of additional resistance determinants, genes that erode susceptibility to other antibiotic classes and together account for the uniformly multidrug-resistant phenotypes observed in the laboratory.

To map the lineages, the researchers applied multilocus sequence typing, a method that reads internal fragments of several housekeeping genes and assigns each isolate a sequence type based on its combination of allelic variants. Using the Pasteur scheme, the predominant type was ST2, found in 14 of the 30 isolates, or 46.7 percent — the sequence type that defines International Clone 2, one of a handful of globally disseminated lineages behind hospital outbreaks on multiple continents. ST85 accounted for five isolates and ST1, the marker of International Clone 1, for three. The Oxford scheme, which resolves relationships at finer scale, split the collection into ST1697, present in nine isolates; ST1089 and ST451, with five each; and ST405, with three. Strikingly, three isolates carried allele combinations matching no sequence type in the MLST database, hinting at locally evolved genetic profiles not yet catalogued anywhere in the world. The double-scheme approach shows why genomic surveillance matters: broad clonal labels reveal global connections, while finer typing exposes local diversity.

The core-genome analysis pushed resolution further still. Rather than sampling a handful of genes, the team aligned the DNA sequences shared by all 30 isolates and catalogued the single-nucleotide polymorphisms — individual DNA letter changes — that distinguish them. The resulting phylogeny showed isolates generally clustering according to their sequence types, confirming the internal consistency of the two typing systems. Within that broad pattern, the fine structure told two stories. Closely related clusters, separated by only tiny numbers of mutations, pointed to clonal relatedness: single strains that had spread through healthcare settings, leaving genetically near-identical descendants. Meanwhile, genetically distinct isolates sharing the same sequence type suggested independent diversification, in which bacteria descended from a common ancestor accumulated mutations and recombined over time while circulating locally. Distinguishing between these scenarios matters operationally, because a tight cluster may signal an ongoing transmission chain that infection control teams can interrupt, whereas deep diversity indicates long-term endemic circulation within the hospital system.

The study also catalogued the bacteria’s protective architecture. Typing of the capsular polysaccharide locus — the K-locus, a cluster of genes that manufactures the sugar armor surrounding each cell — identified nine different KL types across the collection, with KL152 dominating at 46.7 percent. The capsule is far more than decoration: it masks surface molecules from immune recognition, helps the bacterium survive desiccation on hospital surfaces, promotes biofilm formation on catheters and ventilator tubing, and blocks infection by bacterial viruses being explored as therapeutics. Analysis of the outer core locus, the genetic region encoding the outer core of the lipooligosaccharide that decorates the cell surface, revealed three major OCL types: OCL1 in half the isolates, OCL16 in 20 percent and OCL15 in 16.7 percent. Together, these surface signatures provide a molecular fingerprint of the Ethiopian population and a reference point for therapies now in development, several of which target specific capsular types.

Virulence potential was equally conspicuous. All isolates carried conserved repertoires of genes governing biofilm formation — the construction of slimy, antibiotic-impermeable communities on surfaces — and quorum sensing, the chemical communication system bacteria use to coordinate group behaviors such as biofilm maturation and toxin production. Genes for iron acquisition were universal, enabling the bacteria to strip this essential nutrient away from host proteins, as were components of secretion systems that deliver effector proteins and type IV pili, the retractable molecular harpoons that mediate surface attachment, twitching motility and DNA uptake. The plasmid profile added a final layer of concern. The isolates harbored plasmids of the Rep_3 superfamily, extrachromosomal DNA molecules that replicate independently of the chromosome, representing 14 distinct plasmid types across the collection. Plasmids are the classic vehicles of horizontal gene transfer, and their abundance in a carbapenem-resistant population raises the possibility that resistance determinants could hop between lineages, reshaping the local gene pool into new and potentially more formidable combinations.

For the researchers, the takeaway is that whole-genome sequencing has moved from academic luxury to public health necessity. The technique, they conclude, provides high-resolution insight into population structure and genetic diversity, and its routine application can strengthen antimicrobial-resistance surveillance while informing infection prevention and control strategies — identifying which clones are circulating, flagging transmission clusters and revealing when new resistance mechanisms arrive. The team has deposited all 30 genomes in a publicly accessible NCBI BioProject, PRJNA1447996, allowing researchers anywhere to place the Ethiopian isolates on the global map of A. baumannii diversity. As International Clone 2 and its cargo of NDM-1 and OXA-23 demonstrate, resistance genes are not contained by borders; they ride with patients, equipment and trade routes into any hospital that lacks the surveillance to detect them. In that sense, the sequenced isolates from Addis Ababa function as an early-warning system — proof that the genetic machinery of the superbug era is already entrenched, and that watching it, base pair by base pair, is now a core instrument of defense.