Year: 2026
Type: article
Abstract: Antimicrobial therapy, developed over the past century, has historically relied on broad-spectrum agents that are effective against a wide range of bacterial pathogens (Muteeb et al. 2023). Although the approach restored global health, the rapid rise in antimicrobial resistance (AMR) and the shift in commensal microbiota show the shortcomings of treating diverse pathogens with uniform chemical pressures (Huang et al. 2025). Each dose of a broad-spectrum antibiotic reshapes microbial communities, disrupts host-microbiome interactions, and exerts selective pressure that accelerates the emergence and spread of resistance genes (Li et al. 2025). As AMR rates increase across major clinical pathogens, the medical community is confronting a stark reality: what once worked reliably can no longer be assumed safe, sustainable, or effective (Ahmed et al. 2024).Precision antimicrobials are an alternative to broad-spectrum antibiotics. Precision antimicrobials act in two ways: either they directly kill the pathogenic organism with minimal side effects, or they specifically inhibit a crucial node in pathogenesis, thereby interfering with the pathogen's maintenance and/or persistence in the host (Paharik et al., 2017). Precision antimicrobial agents that selectively target specific pathogens, molecular pathways, or resistance mechanisms offer a tailored strategy to combat AMR (de la Fuente-Nunez et al. 2017). Advances in whole-genome sequencing, systems biology, and laboratory automation reveal a pathogen diversity at unprecedented resolution, showing that strains within the same species can differ dramatically in virulence, resistance determinants, and metabolic dependencies (Tedersoo et al. 2019). However, new computational tools enable rapid interpretation and linking genomic signatures to functional vulnerabilities (Clark and Lillard Jr 2024). Precision therapeutics aims to design pathogen-specific, mechanism-aware, and adaptable antimicrobial interventions that are responsive to evolving resistance landscapes (Spaulding et al. 2018).Such transition shifts are seen in both oncology and immunology, where personalized therapy has superseded one-size-fits-all approaches (Tomasello, Szefler, and Cahill 2026). While precision antimicrobial strategies borrow conceptual inspiration from oncology precision medicine, infectious disease care is fundamentally different because infections may be polymicrobial, pathogens can evolve rapidly, and many patients require immediate empiric therapy before diagnostic results are available (Bissonnette and Bergeron 2012). Accordingly, precision antimicrobials should be presented as a complement to, rather than a replacement for, timely empiric treatment in acute care settings. However, realizing this vision requires more than biological insight. It demands computational/algorithmic innovations that are capable of predicting effective targets, identifying optimal compounds, as well as navigating the complexity of host-pathogen interactions (Wong, de la Fuente-Nunez, and Collins 2023). This opinion article outlines how precision antimicrobials can reshape the landscape of therapeutic development (Figure 1). We first describe the limitations of broadspectrum therapy, then explore how multi-omics data can pinpoint actionable molecular targets and discuss the centerpiece of this transition: the rise of artificial intelligence (AI) and computational modeling as engines for rational antimicrobial design. Finally, we highlight how digital predictions inform clinical decision-making, forming an iterative pipeline that links molecular data to actionable therapies. Together, these components represent a transformative framework for the next-generation precision antimicrobial discovery.Broad-spectrum antibiotics provide empirical coverage against a wide range of Gram-positive and Gram-negative bacteria, including AMR strains, while awaiting culture results. They are particularly important in settings without rapid diagnostics and in intensive care units (ICUs), where delays in antimicrobial therapy are associated with increased mortality (Rhee et al. 2020;Timsit, Depuydt, and Kanj 2024). Despite these advantages, indiscriminate reliance on broad-spectrum coverage is increasingly being questioned because it can drive resistance selection and microbiome disruption even when narrower options would have sufficed. However, this critique should be tempered by the fact that broad-spectrum therapy remains essential in septic shock, ICU empiric treatment, polymicrobial infections, and settings of substantial diagnostic uncertainty, where delayed or overly narrow therapy can worsen outcomes (Strich et al., 2020).A study by Webb et al. involving 1,995 patients with community-acquired pneumonia reported that 39.7% received broad-spectrum antibiotics, whereas only 3% of cases were caused by multidrug-resistant (MDR) bacteria, highlighting the potential overuse of broad empirical antimicrobial therapy (Webb et al. 2019). After adjusting for confounders, broad-spectrum antibiotic use was associated with higher mortality and a 17.5% rate of antibiotic-related adverse events. In cases with negative bacteriological results, alternative diagnoses may have been missed (Webb et al. 2019). In another cohort of 15,183 adults with bacteriologically confirmed community-acquired sepsis, 81.6% received appropriate antibiotics (Rhee et al. 2020). Less than 30% of cases were attributed to MDR pathogens, yet unnecessary broad-spectrum antibiotics were administered in 67.8% of patients. This inappropriate use was linked to increased hospital mortality (adjusted OR 1.27, 95% CI 1.06-1.40) and higher rates of acute kidney injury and Clostridioides difficile infections. Furthermore, studies on nosocomial infections suggest that dual antibiotic therapy offers no significant advantage over single-agent treatment while increasing overall antibiotic exposure (Barbier et al. 2024).One major disadvantage of broad-spectrum antibiotics is their disruption of the microbiome. By non-selectively eliminating both pathogenic and beneficial bacteria, broad-spectrum agents can cause dysbiosis, increasing the risk of opportunistic infections such as C. difficile colitis (Dahiya and Nigam 2023). Dysbiosis impairs the balance of the gut microbial ecosystem, reducing its intrinsic resistance and protective capacity (Levy et al. 2017). Changes to gut microbiota are increasingly implicated to several noncommunicable diseases. Persistent dysbiosis may drive the progression of manageable conditions into chronic disorders and has been associated with gastrointestinal diseases, liver disorders, colorectal cancer, diabetes, and dermatological conditions such as eczema. Furthermore, antibiotic-induced dysbiosis can lead to diarrhoea and recurrent infections caused by C. difficile (Duvallet et al. 2017;Wirbel et al. 2019). A 2025 study using 16S rRNA gene sequencing examined the short-and long-term effects of commonly used antibiotics ampicillin, vancomycin, metronidazole, and neomycin on murine gut microbiota showed significant antibiotic-induced alterations in intestinal microbial composition, leading to gut dysbiosis and competition among bacterial populations (Huang et al. 2025). The study also reported that combination antibiotic therapy, particularly involving quinolones and metronidazole, caused more persistent dysbiosis than treatment with a single antibiotic (Huang et al.Precision antimicrobial strategies are only as effective as the diagnostics that guide them. Conventional culture-based identification and susceptibility testing can take 48 to 96 hours, a delay incompatible with conditions such as sepsis that demand therapy within the first hour, which is why empiric broad-spectrum treatment remains the default practice (Marin et al. 2025). Rapid multiplex PCR panels (e.g., GeneXpert Carba-R, BioFire) now return pathogen and resistance-gene results within 45 minutes to a few hours, while culture-independent metagenomic and nanopore sequencing extend this to polymicrobial and culture-negative infections in near real time (Alatawi et al. 2025). However, cost, infrastructure, and trained-personnel requirements still limit access to these tools, and portable nanopore platforms, though cheaper than short-read sequencing, remain constrained in low-resource settings by reagent supply chains and the lack of standardized workflows (Thornval et al. 2025). Closing this diagnostic gap is a prerequisite for precision antimicrobials to move from concept to clinically actionable, globally equitable practice.The precision antimicrobial paradigm anchors drug discovery to critical molecular targets of the bacterial pathogen, structurally absent or sufficiently divergent in the human host, and accessible to pharmacological intervention (Lewis 2013;Tommasi et al. 2015). The strategy demands a systematic framework integrating genomic essentiality data, structural biology, and translational feasibility into a coherent prioritization logic (Brown and Wright 2016;Tommasi et al. 2015). The first non-negotiable criterion is essentiality: gene loss or protein inhibition must be a critical target that impairs bacterial viability in the host (Payne et al. 2007). Genome-wide screens using transposon insertion sequencing (Tn-seq) and CRISPRi have systematically arranged essential genes across priority pathogens including Staphylococcus aureus, Mycobacterium tuberculosis, Klebsiella pneumoniae, and Acenetobacter baumannii (Tommasi et al. 2015;Stokes et al. 2020). Essentiality must be paired with selectivity, exploiting fundamental prokaryoticeukaryotic discontinuities: 70S versus 80S ribosome architecture, type II versus type I fatty acid synthase organization, and the absence of two-component signal transduction in mammalian physiology (Brown and Wright 2016; Farha and Brown 2019). These intrinsic differences define selectivity windows through which pathogen-targeted agents operate with minimal host toxicity (Payne et al. 2007;Tommasi et al. 2015).Table 1 integrates twelve priority target categories organized into four functional groups (cell envelope, information processing, regulatory/adaptive, and proteostasis/resistance modulators) alongside their selectivity basis, key mechanistic advances, and the validation technologies enabling their identification. Rapid advances in structural modeling from AlphaFold2, chemoproteomics (including activity-based protein profiling (ABPP) and thermal proteome profiling (TPP)), and pan-genomic studies on clinical isolate collections have fundamentally altered the pace of target discovery, revealing polymorphisms that distinguish patient populations by predicted inhibitor susceptibility, a diagnostic approach similar to oncology biomarker projects (Boucher et al. 2009;Stokes et al. 2020;Tommasi et al. 2015). The four functional groups in Table 1 are arranged to reflect a hierarchy of selectivity, moving from targets that are structurally absent from human cells to those that are merely divergent in sequence or regulation. Group A (cell envelope) and Group B (information processing) targets offer the cleanest selectivity windows, since structures such as FtsZ, the bacterial 70S ribosome, and the type II fatty acid synthase system have no direct human counterpart, which is why several agents derived from this group, including teixobactin, fidaxomicin, and afabicin, have reached clinical evaluation. Group C (regulatory and adaptive) targets such as two-component signal transduction systems are attractive precisely because they are entirely absent in mammalian physiology, but most inhibitors remain at an earlier discovery stage because these systems are less biochemically tractable than classical enzyme targets. Group D (proteostasis and resistance modulators) illustrates a complementary strategy: rather than killing bacteria outright, agents such as efflux pump inhibitors and ClpP dysregulators restore the efficacy of existing antibiotics or eliminate persister cells that evade conventional bactericidal mechanisms. Read together, the table indicates that the validation technologies are not interchangeable, but cumulative: Tn-seq and CRISPRi typically establish essentiality first, structural methods such as cryo-EM and AlphaFold2 then resolve the selectivity basis at atomic resolution, and chemoproteomic platforms (ABPP, TPP) subsequently confirm target engagement in cells, together forming the prioritization pipeline that determines which targets are mature enough for medicinal chemistry investment.Artificial intelligence is increasingly useful for prioritizing targets, assessing candidate compounds, and integrating genomic, phenotypic, and structural data in antimicrobial research (Zavaleta-Monestel et al. 2025). However, its output must be interpreted cautiously because model performance can be limited by training-set bias, sparse representation of certain pathogens, uneven annotation quality, and weak external validation across laboratories, species, and clinical settings (Serrano et al. 2024;Sakagianni et al., 2024). In addition, AI-generated structure-function relationships may appear plausible but remain unverified, and some predictions can fail when transferred to new datasets or real biological systems. These limitations mean that AI should be treated as a decision-support tool that helps generate hypotheses, rather than as a substitute for experimental validation (Aithani et al. 2023;Bakhtiyar et al. 2025;Mamoudou and Mune 2025).Machine learning models trained on genomic and phenotypic data can now infer pathogen vulnerabilities directly from mutational signatures or resistance trajectories (Hossain & Yousefi, 2026). Neural networks integrate comparative genomics, essential gene datasets, and multi-omics profiles to identify promising targets that differentiate between strains or species, to develop antibiotics that do not affect beneficial bacteria in the environment (Przymus et al. 2024;Wang, Zhu, and Li 2025). Predictive modeling also enables researchers to map resistance pathways before they emerge experimentally, facilitating the selection of compounds that are less prone to rapid resistance evolution (Jacobs et al. 2024). The use of such technologies is consistent with recent developments that indicate that it is possible to predict mechanism of action (MoA) with the help of deep learning solely based on variant signatures (Krentzel et al. 2026).Rational drug design has been further revolutionized by advances in structural prediction and computational modeling (Mozaffari et al. 2025). Tools such as AlphaFoldderived structural models, physics-informed scoring functions, and generative molecular design frameworks enable the rapid prediction of ligand-target interactions (Sim et al. 2025). Large-scale parallel docking and simulation studies conducted on GPU-based systems virtually test millions of potential drugs, identifying only high-affinity compounds while discarding those that might cause adverse side effects due to interactions with nontargeted sites (Blanes-Mira et al. 2022). The application of reinforcement learning methods to optimize lead molecules considers toxicity and specificity constraints (Gow et al. 2022).Network-based computational strategies add another dimension by integrating host-pathogen interactions, metabolic dependencies, and regulatory circuitry (Durmuş et al. 2015). The approach allows for the creation of targeted antimicrobial combinations capable of killing microbes as well as interfering with their adaptation to induce tolerance and persistence (Cantrell et al. 2022). Significantly, AI-enabled drug repurposing can leverage existing drugs against new targets in microbes -a critical factor for AMR-related research (Yakobi and Nwodo 2025).Together, these computational approaches establish a foundation for precision antimicrobial development: algorithmic filtering of chemical space, target-specific compound optimization, and predictive modeling of resistance evolution. AI elevates antimicrobial design from trial and error to a rational, data-driven discipline, an essential shift for the post-broad-spectrum era. Considering AI's potential advantages in drug research, there are a number of obstacles and limitations to consider. The availability of appropriate data is one of the main obstacles (Vamathevan et al., 2021). To address the limits of AI-based drug development by treating AI as part of a hybrid pipeline: improve the data, retain people in the loop, and require every prediction to pass experimental and translational validation (Blanco-González et al., 2023).The rapid emergence of AMR has accelerated the development of novel broadspectrum antibacterial agents. BWC0977 is a novel inhibitor of bacterial topoisomerase that targets DNA gyrase and topoisomerase IV, thereby hindering bacterial DNA synthesis. The drug is extremely effective against multidrug-resistant Gram-positive and Gram-negative bacteria, owing to its low minimum inhibitory concentrations (MICs). Preliminary clinical trial results have suggested the safety and effectiveness of BWC0977 in treating serious infections caused by multidrug-resistant bacteria, such as bacterial pneumonia (Hameed et al. 2024).The translation of broad-spectrum antibacterial activity into clinical application has also been demonstrated in phase III clinical studies, such as the EAGLE-2 and EAGLE-3 trials, which evaluated the efficacy of the novel antibiotic Gepotidacin compared with Nitrofurantoin for the treatment of uncomplicated urinary tract infections (uUTIs). The findings revealed that Gepotidacin was non-inferior to nitrofurantoin, with superiority in some analyses of EAGLE-3. Overall, the an with being the most side adverse effects were to and no safety were et al. with antibacterial are another of broad-spectrum agents for therapeutic is a broad-spectrum activity against bacteria, and with activity against Staphylococcus The compound phase II clinical as a treatment for infections, disease and as an for and and promising is a with activity against Gram-positive bacteria and a low of resistance by the drug a phase II clinical trial its to of Staphylococcus in patients at risk of infections et al. the and in of the bacterial In the drug demonstrated antimicrobial activity against with selective efficacy and phase I clinical have and safety data and parallel with these broad-spectrum several antibiotics how the logic is the in Table 1 as a is a antibiotic that has through phase II for and infections, including while gut since its target enzyme is to fatty acid et al. 2020). A more is a inhibitor that a metabolic specific to its to the compound to selectively eliminate this pathogen from while commensal species et al. 2024). These agents that precision antimicrobials are not essential pathways, once using the genomic and structural approaches in can be into selective compounds with even though most remain earlier in clinical development than their broad-spectrum antimicrobial are the key to AMR and the microbiome while clinical effectiveness et 2024). The of functional genomics, and structural can generate targets that enable targeted therapeutic interventions for more and more effective and 2024). intelligence and predictive diagnostics rapid compound selection and the prediction of resistance mechanisms. However, precision antimicrobial therapy on rapid and diagnostics that can identify both the pathogen and key resistance within a clinically actionable et al., et al., 2026). clinical trials, antimicrobial and are crucial for the effective use of targeted and are to the development of and broad-spectrum as an and most remains to alternative resistance including efflux pump and highlighting that next-generation antibacterial agents may still clinically pathways that limit long-term to with inhibition of and some D activity against many et al., use of broad-spectrum antibiotics remains essential for but their overuse antimicrobial resistance (AMR) and disrupts the microbiome. targeted antimicrobial are the key to AMR while clinical develop these the of functional genomics, and structural is to identify specific pathogen targets for artificial intelligence helps researchers compounds and predict resistance mechanisms before they precision antimicrobial therapy on rapid and These targeted diagnostics identify both the pathogen and its resistance within a clinically actionable to this we require clinical trials, antimicrobial and that the development of therapies. Precision antimicrobial therapy antimicrobial by facilitating treatment, whereas provide the clinical framework for appropriate antimicrobial Together, these approaches more targeted therapy, unnecessary antimicrobial and improve treatment et al., 2024).Precision antimicrobial approaches have in intestinal infections, infections and therapy by in a model of the development of the as a targeted therapeutic strategy et al., The commensal Staphylococcus a novel antibiotic that selectively Staphylococcus highlighting the potential of precision antimicrobial et al.,
Source: Frontiers in Microbiology
Institutions: University of Nebraska–Lincoln, KIRND Institute of Research and Development Pvt Ltd, Centre for Cellular and Molecular Biology, Council for Scientific and Industrial Research, Council for Scientific and Industrial Research, Council of Scientific and Industrial Research, Houston Methodist, Cornell University, Weill Cornell Medicine
Language: English
Subfield: Applied Microbiology and Biotechnology
Field: Immunology and Microbiology
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7168759588