Abstract

Antimicrobial resistance (AMR) poses a significant threat to global public health, with multidrug-resistant pathogens becoming increasingly common. Combination therapies, which involve the simultaneous use of two or more drugs to treat infections, have emerged as a potential strategy to overcome this challenge. This overview explores the  application of AI/ML tools in combating multidrug resistance and focuses particularly on the how these in silico methods can be utilised to optimise combination therapies in the fight against AMR. 

Introduction: Antimicrobial Resistance

Antimicrobial resistance (AMR) is a pressing global issue, with the World Health Organization (WHO) declaring it one of the top ten global public health threats facing humanity and the World Bank estimating that AMR could result in $1 trillion additional healthcare costs by 2050.^1,2 Antibiotic resistance was estimated to be responsible for 1.27 million deaths in 2019 and contributed to nearly five million more.¹ The misuse of antimicrobials in humans, plants and animals are the main drivers for the emergence of drug-resistant pathogens and consequently puts many of the gains of modern medicine in jeopardy. Microbes are able to meet the evolutionary challenge of antimicrobial chemotherapy by transferring pre-existing resistance determinants from their respective gene pool via mobile genetic elements. Alternative resistance mechanisms within the microbial arsenal also include limiting the uptake of the antimicrobial, inactivating the drug, actively pumping the drug out the cells through efflux pumps or modifying the drug target (Figure 1).^3–5 Enzyme-catalysed deactivation of antibiotics is one of the most important drivers that confers antibiotic resistance and the most successful examples are β-lactamase inhibitors.⁶

‍

‍

Figure 1: Molecular Mechanisms of antimicrobial resistance in bacterial pathogens.⁷

Alternatives to antibiotics such as anti-virulence agents, antibodies and vaccines have been explored as therapies but these best serve as adjunctive or preventive therapies and conventional antibiotics are still indispensable.⁸ Only 12 antimicrobial drugs entered the market from 2017-2021 and very few are expected to gain market authorisation in the near future.⁹ The slow drug discovery progress in antimicrobials can be attributed to the lengthy pathway to approval, high R&D costs and low success rates. Antibiotic development is no longer considered an economically viable investment in the pharmaceutical industry since it takes 10-15 years to progress an antibiotic candidate from pre-clinical to clinical stages and the generation of resistant bacteria takes an average of 2 years thus rendering the development of antibacterial drugs far behind the speed of bacterial resistance.¹⁰ These scientific and commercial challenges has led to the majority of the pharmaceutical industry systematically dismantling antibiotic discovery programs. The inadequate levels of R&D in antimicrobials as well as the failure to recognise a patient’s risk factor for infection has resulted in an urgent need for additional measures to ensure equitable access to new and existing treatments in order to ensure appropriate treatment and favourable outcomes.

‍

Uncovering the Potential of Combination Therapies

The emergence of multidrug-resistant pathogens, such as Methicillin-resistant Staphylococcus aureus (MRSA) and extensively drug-resistant tuberculosis (XDR-TB), has made traditional antibiotic treatments less effective. In response to this challenge, researchers have turned to innovative solutions like combination therapies (the use of two or more drugs with different mechanisms of action to target multiple pathways in the pathogen) as an alternative strategy to produce more viable and potent drugs from the current arsenal.
Combination therapies have been used in medicine for decades, particularly in the treatment of HIV/AIDS and cancer. One notable example is the combination of nucleoside reverse transcriptase inhibitors (NRTIs) with protease inhibitors (PIs) in the treatment of HIV.¹¹ This approach has been highly successful in suppressing viral replication and reducing the risk of drug resistance. 

‍

Drug-drug interactions can be classified into three types: synergetic, no interaction and antagonistic, the focus of this overview will be on synergistic combinations. The three classes of synergistic drug-drug combinations include, I) the combination of antibiotic and another antibiotic that act in a tandem or parallel manner by targeting distinct essential molecular processes¹², II) an antibiotic and a non-antibiotic agent that enhance the efficacy of antibiotics commonly known as adjuvants or potentiators¹³, III) two non-antibiotic compounds that target non-essential but synthetically legal functions.¹⁴ The unprecedented discovery of  clavulanic acid as β-lactamase inhibitors led to the first antibiotic-adjuvant combination Augmentin (amoxicillin-clavulanic acid pair)¹³ and since then there has been a growing interest into the use of adjuvants in combination therapies as a complementary strategy to fight antimicrobial resistance. This is due to the notable advantages of combination therapies in antimicrobial development over single-agent therapies. These include the synergistic and efficacious effects some combinations can generate, the reduced resistance by the targeting of multiple pathways and the ability to invoke a broad spectrum approach to targeting a wide range of pathogens making them effective against polymicrobial infections. The rational design of adjuvants requires an understanding of the correlation between the chemistry of adjuvant molecules and the biology of resistance mechanisms. The four known modes of action used to divide adjuvants include, enzyme inhibitors, membrane saboteurs, signalling inhibitors and immune enhancers (Figure 2). Resistance inhibitors target bacterial enzymes and efflux pumps and the inhibition of enzyme-mediated drug resistance has been proven to be a clinically successful regimen.^7,15 Membrane saboteurs potentiate the activity of some specific antibiotics by enhanced membrane permeability¹⁶ and signalling inhibitors target the transduction systems that bacteria employ to respond to different environmental changes.¹⁷ Given the vital role of the host defence mechanism in confronting invasive bacteria, targeting immune mechanism offers an alternative set of targets for antibiotic adjuvants.¹⁸ 

‍

‍

Figure 2: Antimicrobial resistance mechanisms and the modes of action of antibiotic adjuvants.⁷

Recently, scientists at Ineos Oxford Institute (IOI) found a new combination therapy to combat AMR by targeting two key bacterial enzymes involved in resistance.¹⁹ Meropenem, a critical antibiotic used to treat multi-drug resistant infections is becoming increasingly ineffective due to AMR and this study explored the use of adjuvants to inhibit bacterial enzymes such as metallo-β-lactamases (MBL) or serine-β-lactamases (SBL). A triple-drug combination of meropenem, an MBL (InC58) and SBL inhibitor (AVI) was found to be more effective at inhibiting the growth of bacteria compared to either of the dual-drug combinations with an MIC₅₀ that was 64 times lower than the dual-drug combinations. Whilst further development is needed to show in vivo activity, such novel treatments could significantly extend the antibacterial activity of carbapenems.

‍

In the fight against antimicrobial resistance the use of adjuvants to tailor the properties of existing antibiotics has also been explored (Figure 3). De novo drug discovery pipelines are time consuming and expensive and can cost more than $2 billion from discovery to market. With only 10% of de novo drugs put through to clinical trials, repurposing can deliver efficacious therapeutics to patients much faster by leveraging pre-existing human trial data.²⁰ This can increase the ROI for pharmaceutical companies due to significant reductions in cost and the subsequent impact on R&D budgets. The screening of suitable antibiotic adjuvants from previously approved drugs could cut the associated development times especially the stage of clinical evaluation of safety, thus successfully ushering new antimicrobials through the pipeline and into the market faster. The repurposing of FDA-approved drugs is not a novel idea and many drugs have been repurposed to treat various diseases for example, Benevolent AI’s use of AI algorithms to detect Baricitinib, a drug created for rheumatoid arthritis was proven to be an effective antiviral medication for the treatment of hospitalised adults with coronavirus.²¹ Many studies have shown the use of this approach to successfully elucidate  new drug-drug interactions. Farha et al. conducted a high-throughput chemical screen for antagonists of targocil and ticlopidine, inhibitors of wall teichoic acid (WTA) in Methicillin-resistant Staphylococcus aureus (MRSA) and identified clomiphene, a widely used fertility drug as an inhibitor of undecaprenyl diphosphate synthase.²² Additionally,  Loperamide, an antidiarrheal drug, was identified as an adjuvant of the semi-synthetic tetracycline antibiotic minocycline.²³ 

‍

‍

Figure 3: Representative examples of drug repurposing as β-lactam adjuvants against MRSA.⁷

A growing body of research is also focussing on developing antimicrobial peptides as antibiotic adjuvants which has been shown to increase membrane permeabilization via a demonstrable capability to bind and disrupt both negatively charged and zwitterionic membranes.^24,25 Researchers at the University of Pennsylvania’s Perelman School of Medicine leveraged AI to mine the genomes and proteomes of extinct organisms to identify a number of clinical antimicrobial peptidic candidates displaying in vitro and in vivo activities. This molecular de-extinction of ancient antimicrobial peptides enabled by paleoproteome mining offers another alternative framework for antibiotic discovery.²⁶

‍

In this post-antibiotic era, the discovery and development of novel anti-infective regimens is paramount and the three different combinations as described above afford a promising pipeline to such. Combination I (antibiotic + antibiotic) can achieve broad spectrum coverage against pathogens, combination II (antibiotic + adjuvant) offers a promising opportunity to prolong the life of clinically validated antibiotics and combination III (adjuvant + adjuvant) whilst a relatively unexplored frontier, may accelerate the development of narrow-spectrum drug combinations. However, the clinical implementation of combination therapies faces several challenges, including regulatory hurdles, managing the inherent increase in toxicity and the need for more robust clinical trials to establish efficacy and safety.

‍

The Role of AI in AMR and the Optimisation of Combination Therapies

Artificial intelligence (AI) has emerged as a powerful tool for drug discovery and optimization and more broadly across medical research and clinical practice. Whilst to date there has not been a FDA-approved AI-created drug, an AI-designed anti-fibrotic small molecule inhibitor (INS018_055) developed by Insilico Medicine and currently in phase II clinical trials, could soon well be the first.²⁷ The data analysis capabilities achieved with AI/ML solves the problem of limited rational decision making due to insufficient information or time constraints and coupled with sufficient compute power can significantly advance the predictive tools we apply to antimicrobial drug development. The integration of machine learning (ML) to AMR is becoming increasingly pervasive where tools are helping to predict early antibiotic resistance, develop novel therapies, diagnose AMR and repurpose existing drugs (Figure 4). 

‍

‍Figure 4: Antimicrobial drugs whose development has utilised computational and AI technologies.²⁸ 

‍

Currently AMR is principally diagnosed using two techniques in clinical microbiology, one is classical culture-based antimicrobial susceptibility (AST) and the other is a whole-genome sequencing alternative (WGS-AST). These tests are notoriously slow and have benefitted from computational methods for bacterial susceptibility profiling.²⁹ ML methods now serve as a bridge between specimen collection and molecular susceptibility analysis facilitating the time-sensitive empirical antibiotic choices. In the context of AMR prediction, besides the detection of antimicrobial resistance phenotypes, different ML models have been employed in the pursuit for more targeted treatment options. Goodman et al. used recursive partitioning to build a decision tree for the prediction of extended-spectrum β-lactamase (ESBL) production in Escherichia coli and Klebsiella pneumoniae based on patient epidemiological and microbiological data.³⁰ Additionally, researchers supported by the Oxford Martin Programmes used a combination of fluorescence microscopy and AI to detect antimicrobial resistance.³¹  The method relied on deep learning models that could analyse bacterial cell images and detect structural changes that may occur in cells when targeted with antibiotics. 

‍

The most extensive application of AI/ML in AMR is in the guided approach to drug discovery. It is predicted that there are close to 10³⁰ possible drug-like molecules so a few hundred million is merely scratching the surface when trying to explore the entire chemical space. In this drug design paradigm, AI can learn from chemical datasets to predict new molecular structures with unique properties. AI-facilitated small molecule discovery takes the shape of identifying biosynthetic gene clusters (BGCs), MOA-driven compound library screening, protein structure-function guided rational design and drug repurposing. Researchers at Stanford medicine and McMaster university devised a new AI model SyntheMol which was trained to construct drugs using a library of more than 130,000 molecular building blocks and a set of validated chemical reactions and in less than nine hours SytheMol generated 25,000 possible antibiotics and the recipes to make them. This ultimately led to six novel drugs aimed at killing resistant strains of Acinetobacter baumannii, a leading pathogen responsible for antibacterial resistance-related deaths.³²

‍

Whilst the large combinatorial space to explore for combination therapies presents a daunting challenge, a number of data-driven approaches like machine learning algorithms have been employed to identify novel synergistic drug interactions from millions of potential combinations. The potential combinations between novel and existing antibiotics and adjuvants presents a unique opportunity to leverage machine learning algorithms to efficiently sift through the expansive chemical space that exists. AI/ML can be used to predict synergistic combinations, optimise dosing regimens whilst minimising toxicity, personalised treatments and predict drug repurposing candidates. There are many data-driven approaches currently employed in the development of combination therapies, an approach based on drug information such as Combination Synergy Estimation (CoSynE) uses structural data of individual compounds to predict drug interaction outcomes.³³ Monitoring pathogen response can also provide insights into the chemogenomic and transcriptomic data that drive synergistic interactions, one approach is INDIGO (inferring drug interactions using chemogenomics and orthology) which correctly identifies synergistic drug combinations used in the clinic.³⁴ ML tools also extends into the screening of drug combinations where advances in sequencing has led to the development of new screening techniques for the discovery of novel antibiotic adjuvants from huge libraries of non-antibiotic agents. A combined experimental-computational approach based on high-throughput metabolomics was developed to predict drug-drug interactions.^35,36 The study applied high-throughput metabolomics to monitor the metabolic response of E. coli to a library of 1,279 chemical compounds that have little to no direct antimicrobial activity. The drug metabolome profiles were used to make de novo predications of modes of action of drugs and this led to several novel drug combinations.³⁷ As mentioned above, the repositioning of existing drugs in combination therapies is a growing area of research and drug repurposing AI is a key innovation area (Figure 5). Drug repurposing used to be serendipitous but now computational methods are utilised as a more methodical approach ranging from molecular docking and binding-site detection to pathway mapping, genetic associations and machine learning. 

‍

Figure 5: Key players in drug repurposing AI.³⁸ 

‍

Challenges with AI adoption in AMR and combination therapies 

The integration of AI/ML in AMR whilst evidently opens up a multitude of opportunities to revolutionise the fight against multi-drug resistant pathogens also presents some significant challenges. Despite their high predictive performance, most current AI systems, particularly those that rely on deep neural networks, still suffer from several limitations namely, a large proportion of existing AI models remain “black-boxes”  with low interpretability, meaning that humans cannot comprehend the rules behind decisions made. This lack of transparency and accountability can lead to severe consequences, especially when it comes to FDA approval and the requirement to understand the mechanism of action of these drugs. This necessitates the need for “explainable models” to ensure one can follow the logical reasoning behind the predictions. Data availability and reproducibility is also important. Many studies already share their source code publicly but improvements can be made in data sharing, open accessibility and thorough code documentation.³⁹  AI has significantly shortened the in vitro stages in the drug-discovery pipelines of novel antimicrobials. However, little aid has been gained from AI regarding how those drugs behave in the human body. Current drug combination experiments are performed in growth conditions that are not representative of infection sites. In vitro conditions often only account for a small set of fixed metabolic conditions, which are drastically different from an in vivo environment that is characterized by complex and dynamic metabolic conditions. This discordance contributes to the challenge with translating synergistic drug combinations to the clinic. 

‍

Conclusions and Future Directions 

Antimicrobial resistance is increasingly undermining existing anti-infective agents and hence constitutes a global challenge in public health. An urgent need to identify novel therapeutic programs, high failure rates and costs in discovery of new antibiotics has prompted a growing interest in new approaches like combination therapies. Combination therapies offer a promising approach to combating antimicrobial resistance and overcoming multidrug-resistant infections. By targeting multiple pathways in the pathogen, these treatments can reduce the risk of resistance emerging and improve patient outcomes. With the advent of artificial intelligence, researchers have new tools at their disposal for optimizing combination therapies. The efficiency with which AI confers will hopefully incentivise pharmaceutical companies to speed up the slow pipeline of antibiotics. Perhaps AI will be the tool that can outpace the evolutionary advantage conferred to an organism which has utilised resistance to ensure it survival. However, addressing the challenges associated with regulatory approval, data availability, and interdisciplinary collaboration will be crucial for realizing the full potential of combination therapies in the fight against antimicrobial resistance. Perhaps finding individual solutions may not be sufficient but rather applying them in unison. Human intelligence, novel technologies, suitable policies, and attitudinal shifts may well be the combination that is needed to mitigate the current state where antimicrobial resistance continues to be a human threat. 

‍

References 

(1) Antimicrobial resistance. World Health Organisation. https://www.who.int/news-room/fact-sheets/detail/antimicrobial-resistance.

(2) Drug-Resistant Infections: A Threat to Our Economic Future. World Bank. https://www.worldbank.org/en/topic/health/publication/drug-resistant-infections-a-threat-to-our-economic-future.

(3) Leus, I. V.; Weeks, J. W.; Bonifay, V.; Smith, L.; Richardson, S.; Zgurskaya, H. I. Substrate Specificities and Efflux Efficiencies of RND Efflux Pumps of Acinetobacter Baumannii. J. Bacteriol. 2018, 200, e00049-18.

(4) Zhang, W.-J.; Xu, X.-R.; Schwarz, S.; Wang, X.-M.; Dai, L.; Zheng, H.-J.; Liu, S. Characterization of the IncA/C Plasmid pSCEC2 from Escherichia Coli of Swine Origin That Harbours the Multiresistance Gene Cfr. J. Antimicrob. Chemother. 2014, 69, 385–389.

(5) Zhang, W.; Fisher, J. F.; Mobashery, S. The Bifunctional Enzymes of Antibiotic Resistance. Curr. Opin. Microbiol. 2009, 12, 505–511. 

(6) Bush, K.; Bradford, P. A. β-Lactams and β-Lactamase Inhibitors: An Overview. Cold Spring Harb. Perspect. Med. 2016, 6, a025247. 

(7) Liu, Y.; Tong, Z.; Shi, J.; Li, R.; Upton, M.; Wang, Z. Drug Repurposing for Next-Generation Combination Therapies against Multidrug-Resistant Bacteria. Theranostics 2021, 11, 4910–4928. 

(8) Czaplewski, L.; Bax, R.; Clokie, M.; Dawson, M.; Fairhead, H.; Fischetti, V. A.; Foster, S.; Gilmore, B. F.; Hancock, R. E. W.; Harper, D.; Henderson, I. R.; Hilpert, K.; Jones, B. V.; Kadioglu, A.; Knowles, D.; Ólafsdóttir, S.; Payne, D.; Projan, S.; Shaunak, S.; Silverman, J.; Thomas, C. M.; Trust, T. J.; Warn, P.; Rex, J. H. Alternatives to Antibiotics-a Pipeline Portfolio Review. Lancet Infect. Dis. 2016, 16, 239–251. 

(9) Lack of innovation set to undermine antibiotic performance and health gains. World Health Organisation. https://www.who.int/news/item/22-06-2022-22-06-2022-lack-of-innovation-set-to-undermine-antibiotic-performance-and-health-gains.

(10) 2021 antibacterial agents in clinical and preclinical development: an overview and analysis. World Health Organisation. https://www.who.int/publications-detail-redirect/9789240047655.

(11) Joly, V.; Descamps, D.; Yeni, P. NNRTI plus PI Combinations in the Perspective of Nucleoside-Sparing or Nucleoside-Failing Antiretroviral Regimens. AIDS Rev. 2002, 4, 128–139.

(12) Indrelie, J. A.; Wilson, W. R.; Matsumoto, J. Y.; Geraci, J. E.; Washington, J. A. Synergy of Imipenem or Penicillin G and Aminoglycosides against Enterococci Isolated from Patients with Infective Endocarditis. Antimicrob. Agents Chemother. 1984, 26, 909–912.

(13) White, A. R.; Kaye, C.; Poupard, J.; Pypstra, R.; Woodnutt, G.; Wynne, B. Augmentin (Amoxicillin/Clavulanate) in the Treatment of Community-Acquired Respiratory Tract Infection: A Review of the Continuing Development of an Innovative Antimicrobial Agent. J. Antimicrob. Chemother. 2004, 53, i3-20.

(14) Aziz, R. K.; Monk, J. M.; Lewis, R. M.; In Loh, S.; Mishra, A.; Abhay Nagle, A.; Satyanarayana, C.; Dhakshinamoorthy, S.; Luche, M.; Kitchen, D. B.; Andrews, K. A.; Fong, N. L.; Li, H. J.; Palsson, B. O.; Charusanti, P. Systems Biology-Guided Identification of Synthetic Lethal Gene Pairs and Its Potential Use to Discover Antibiotic Combinations. Sci. Rep. 2015, 5, 16025. 

(15) Liu, Y.; Li, R.; Xiao, X.; Wang, Z. Molecules That Inhibit Bacterial Resistance Enzymes. Molecules 2018, 24, 43.

(16) Ejim, L.; Farha, M. A.; Falconer, S. B.; Wildenhain, J.; Coombes, B. K.; Tyers, M.; Brown, E. D.; Wright, G. D. Combinations of Antibiotics and Nonantibiotic Drugs Enhance Antimicrobial Efficacy. Nat. Chem. Biol. 2011, 7, 348–350. 

(17) Wuichet, K.; Cantwell, B. J.; Zhulin, I. B. Evolution and Phyletic Distribution of Two-Component Signal Transduction Systems. Curr. Opin. Microbiol. 2010, 13, 219–225.

(18) Perry, J. A.; Koteva, K.; Verschoor, C. P.; Wang, W.; Bowdish, D. M. E.; Wright, G. D. A Macrophage-Stimulating Compound from a Screen of Microbial Natural Products. J. Antibiot. (Tokyo) 2015, 68, 40–46.

(19) Ling, Z.; Farley, A.; Lankapalli, A.; Zhang, Y.; Premchand-Branker, S.; Cook, K.; Baran, A.; Gray-Hammerton, C.; Rubio, C.O.; Suna, E.; Mathias, J.; Brem, J.; Sands, K.; Nieto-Rosado, M.; Trush, M.; Rakhi, N.; Martins, W.; Zhou, Y.; Schofield, C.J.; Walsh, T. The Triple Combination of Meropenem, Avibactam, and a Metallo-β-Lactamase Inhibitor Optimizes Antibacterial Coverage Against Different β-Lactamase Producers. Engineering 2024.

(20) Selvaraj, C.; Chandra, I.; Singh, S. K. Artificial Intelligence and Machine Learning Approaches for Drug Design: Challenges and Opportunities for the Pharmaceutical Industries. Mol. Divers. 2022, 26, 1893–1913. 

(21) Dupuis, D.; Fritz, K.; Ike, E.; Arogundade, O.; Adewara, E. O.; Monday, E. O.; Ayinde, B. O. Current Use of Baricitinib in COVID-19 Treatment and Its Future: An Updated Literature Review. Cureus 2022, 14, e28680. 

(22) Farha, M. A.; Czarny, T. L.; Myers, C. L.; Worrall, L. J.; French, S.; Conrady, D. G.; Wang, Y.; Oldfield, E.; Strynadka, N. C. J.; Brown, E. D. Antagonism Screen for Inhibitors of Bacterial Cell Wall Biogenesis Uncovers an Inhibitor of Undecaprenyl Diphosphate Synthase. Proc. Natl. Acad. Sci. USA 2015, 112, 11048–11053. 

(23) Worthington, R. J.; Melander, C. Combination Approaches to Combat Multi-Drug Resistant Bacteria. Trends Biotechnol. 2013, 31, 177–184. 

(24) Boman, H. G. Antibacterial Peptides: Basic Facts and Emerging Concepts. J. Intern. Med. 2003, 254, 197–215. 

(25) Alalwani, S. M.; Sierigk, J.; Herr, C.; Pinkenburg, O.; Gallo, R.; Vogelmeier, C.; Bals, R. The Antimicrobial Peptide LL-37 Modulates the Inflammatory and Host Defense Response of Human Neutrophils. Eur. J. Immunol. 2010, 40, 1118–1126. 

(26) Maasch, J. R. M. A.; Torres, M. D. T.; Melo, M. C. R.; Fuente-Nunez, C. de la. Molecular De-Extinction of Ancient Antimicrobial Peptides Enabled by Machine Learning. Cell Host & Microbe 2023, 31, 1260-1274. 

(27) Ren, F.; Aliper, A.; Chen, J.; Zhao, H.; Rao, S.; Kuppe, C.; Ozerov, I. V.; Zhang, M.; Witte, K.; Kruse, C.; Aladinskiy, V.; Ivanenkov, Y.; Polykovskiy, D.; Fu, Y.; Babin, E.; Qiao, J.; Liang, X.; Mou, Z.; Wang, H.; Pun, F. W.; Ayuso, P. T.; Veviorskiy, A.; Song, D.; Liu, S.; Zhang, B.; Naumov, V.; Ding, X.; Kukharenko, A.; Izumchenko, E.; Zhavoronkov, A. A Small-Molecule TNIK Inhibitor Targets Fibrosis in Preclinical and Clinical Models. Nat. Biotechnol. 2024. 

(28) Liu, G.-Y.; Yu, D.; Fan, M.-M.; Zhang, X.; Jin, Z.-Y.; Tang, C.; Liu, X.-F. Antimicrobial Resistance Crisis: Could Artificial Intelligence Be the Solution? Mil. Med. Res. 2024, 11. 

(29) Waddington, C.; Carey, M. E.; Boinett, C. J.; Higginson, E.; Veeraraghavan, B.; Baker, S. Exploiting Genomics to Mitigate the Public Health Impact of Antimicrobial Resistance. Genome Med. 2022, 14, 1-15.

(30) Goodman, K. E.; Lessler, J.; Cosgrove, S. E.; Harris, A. D.; Lautenbach, E.; Han, J. H.; Milstone, A. M.; Massey, C. J.; Tamma, P. D.; for the Antibacterial Resistance Leadership Group. A Clinical Decision Tree to Predict Whether a Bacteremic Patient Is Infected With an Extended-Spectrum β-Lactamase–Producing Organism. Clin. Infect. Dis. 2016, 63, 896–903. 

(31) Zagajewski, A.; Turner, P.; Feehily, C.; El Sayyed, H.; Andersson, M.; Barrett, L.; Oakley, S.; Stracy, M.; Crook, D.; Nellåker, C.; Stoesser, N.; Kapanidis, A. N. Deep Learning and Single-Cell Phenotyping for Rapid Antimicrobial Susceptibility Detection in Escherichia Coli. Commun. Biol. 2023, 6, 1–12. 

(32) Stokes, J. M.; Yang, K.; Swanson, K.; Jin, W.; Cubillos-Ruiz, A.; Donghia, N. M.; MacNair, C. R.; French, S.; Carfrae, L. A.; Bloom-Ackermann, Z.; Tran, V. M.; Chiappino-Pepe, A.; Badran, A. H.; Andrews, I. W.; Chory, E. J.; Church, G. M.; Brown, E. D.; Jaakkola, T. S.; Barzilay, R.; Collins, J. J. A Deep Learning Approach to Antibiotic Discovery. Cell 2020, 180, 688-702. 

(33) Mason, J.D.; Stott, I.; Ashenden, S.; Weinstein, Z.B.; Karakoc, I.; Meral, S.; Kuru, N.; Bender, A.; Cokol, M. Prediction of Antibiotic Interactions Using Descriptors Derived from Molecular Structure. J. Med. Chem. 2017, 60, 3902-3912.

(34) Chandrasekaran, S.; Cokol-Cakmak, M.; Sahin, N.; Yilancioglu, K.; Kazan, H.; Collins, J. J.; Cokol, M. Chemogenomics and Orthology-Based Design of Antibiotic Combination Therapies. Mol. Syst. Biol. 2016, 12, 872-884.

(35) Brown, J. C. S.; Nelson, J.; VanderSluis, B.; Deshpande, R.; Butts, A.; Kagan, S.; Polacheck, I.; Krysan, D. J.; Myers, C. L.; Madhani, H. D. Unraveling the Biology of a Fungal Meningitis Pathogen Using Chemical Genetics. Cell 2014, 159, 1168–1187. 

(36) Shao, F.; Li, H.; Hsieh, K.; Zhang, P.; Li, S.; Wang, T.-H. Automated and Miniaturized Screening of Antibiotic Combinations via Robotic-Printed Combinatorial Droplet Platform. Acta Pharm. Sin. B. 2024, 14, 1801–1813. 

(37) Campos, A. I.; Zampieri, M. Metabolomics-Driven Exploration of the Chemical Drug Space to Predict Combination Antimicrobial Therapies. Mol. Cell 2019, 74, 1291-1303.

(38) Artificial intelligence: who are the leaders in drug repurposing AI for the pharmaceutical industry? GlobalData Patent Analytics. https://www.pharmaceutical-technology.com/data-insights/innovators-ai-drug-repurposing-ai-pharmaceutical/ 

(39) Cantrell, J. M.; Chung, C. H.; Chandrasekaran, S. Machine Learning to Design Antimicrobial Combination Therapies: Promises and Pitfalls. Drug Discov. Today 2022, 27, 1639–1651. 

‍