Breaking
the Mold: How AI Just Revolutionized Antibiotic Discovery for IBD (And Why This
Changes Everything About Drug Development)
Picture
this: you're a scientist trying to solve one of medicine's most vexing
puzzles—how do you fight the bacteria causing inflammatory bowel disease
without accidentally nuking all the helpful bacteria keeping your gut healthy?
It's like trying to perform surgery with a sledgehammer instead of a scalpel.
For decades, this dilemma has tormented researchers and patients alike, as
broad-spectrum antibiotics often make IBD symptoms worse by destroying the very
microbes that maintain intestinal balance.
Then,
in what can only be described as a scientific plot twist worthy of a
blockbuster movie, researchers at McMaster University and MIT didn't just
discover a new antibiotic for IBD—they used artificial intelligence to predict
exactly how it would work before they could even prove it in the lab. Published
in Nature Microbiology on October 3, 2025, this breakthrough
represents the first time in history that AI has successfully predicted a
drug's mechanism of action before experimental validation, potentially transforming
how we discover and develop new medicines (Stokes et al., 2025).
The
new antibiotic, dubbed "enterololin," is what scientists call a
"narrow-spectrum" drug—it's the scalpel instead of the sledgehammer,
specifically targeting the family of bacteria called Enterobacteriaceae (which
includes troublesome gut bacteria like E. coli) while leaving
beneficial microbes unharmed (Technology Networks, 2025). But perhaps even more
remarkable than the drug itself is how AI compressed what typically takes years
of painstaking laboratory work into just 100 seconds of computation,
fundamentally changing the economics and timeline of pharmaceutical discovery.
The IBD Crisis
To
understand why this discovery matters, we need to appreciate the devastating
reality of inflammatory bowel diseases. IBD, which includes Crohn's disease and
ulcerative colitis, affects millions of people worldwide with a chronic
inflammatory condition that has no cure and limited treatment options. The
disease essentially involves your immune system mistaking your own intestinal
tissue for an invader, launching a relentless attack that causes pain,
bleeding, and potentially life-threatening complications.
The
relationship between IBD and gut bacteria is particularly complex and cruel.
While beneficial bacteria help maintain intestinal health and regulate immune
responses, harmful bacteria can trigger dangerous flare-ups. Current treatment
approaches often rely on broad-spectrum antibiotics that indiscriminately kill
both good and bad bacteria, potentially making symptoms worse over time by
creating opportunities for resistant pathogens to colonize the disrupted gut
environment.
Research
published in the American Journal of Inflammatory Bowel Disease revealed
a striking pattern: exposure to antibiotics significantly increases the odds of
developing Crohn's disease and IBD, emphasizing the double-edged nature of
current therapeutic approaches (Dar et al., 2023). A landmark study published
in Science Advances further demonstrated that antibiotics
interfere with the protective mucus layer in the intestine, creating a cascade
of problems that can exacerbate IBD symptoms rather than improve them (Bel et
al., 2024).
This
therapeutic paradox has frustrated clinicians and patients for decades. Dr. Jon
Stokes, assistant professor in McMaster's Department of Biochemistry and
Biomedical Sciences and principal investigator on the new study, explains the
dilemma succinctly: "Most antibiotics used in clinics today are
'broad-spectrum' drugs—they're nukes. This can create opportunities for
invasive and drug-resistant species of bacteria, like E. coli, to
move in and colonize the intestines, which can exacerbate conditions like
Crohn's" (News Medical, 2025).
The
global burden of IBD is staggering and growing. Current estimates suggest the
condition affects approximately 1% of the global population, with incidence
rates climbing rapidly in developing countries as they adopt Western lifestyles
and dietary patterns. The economic impact extends far beyond direct medical
costs, as IBD typically strikes young adults during their most productive
years, leading to substantial indirect costs through lost productivity and
reduced quality of life.
From Years to Minutes in Drug Discovery
The
traditional process of understanding how a new drug works—what scientists call
determining its "mechanism of action" (MOA)—typically requires 12-18
months of intensive laboratory investigation. Researchers must conduct countless
experiments to identify which cellular targets the drug affects, how it
interacts with those targets, and why it produces its therapeutic effects. This
MOA determination represents one of the major bottlenecks in drug development,
often consuming significant time and resources before researchers can even
determine whether a promising compound is worth pursuing further.
Enter
artificial intelligence, specifically a generative AI model called DiffDock,
developed by MIT's Regina Barzilay (recently listed among Time Magazine's
most influential people in AI). When Stokes input enterololin's molecular
structure into the AI system, something unprecedented happened: within 100
seconds, the AI predicted that the drug would target a specific protein complex
called LolCDE, which is essential for bacterial survival (Technology Networks,
2025).
"A
lot of AI use in drug discovery has been about searching chemical space,
identifying new molecules that might be active," explains Barzilay.
"What we're showing here is that AI can also provide mechanistic
explanations, which are critical for moving a molecule through the development
pipeline" (Technology Networks, 2025). This represents a fundamental
expansion of AI's role in pharmaceutical research, moving beyond simple pattern
recognition to genuine scientific prediction and hypothesis generation.
The
implications of this technological leap cannot be overstated. Traditional drug
discovery operates on what pharmaceutical executives grimly call the "90%
failure rule"—nine out of ten promising compounds that enter clinical
testing ultimately fail, often due to unexpected toxicity or lack of efficacy
that wasn't predicted by earlier studies. By enabling researchers to understand
how drugs work much earlier in the development process, AI could dramatically
improve success rates while reducing the astronomical costs that make new
medicines so expensive.
The AI
prediction about enterololin proved remarkably accurate. When Stokes' team, led
by graduate student Denise Catacutan, conducted traditional laboratory
experiments to validate the AI's hypothesis, they confirmed that the drug
indeed targets the LolCDE protein complex exactly as predicted. "We did
all of our standard MOA workup to validate the prediction—to see if the experiments
would back-up the AI, and they did," says Catacutan. "Doing it this
way shaved a year-and-a-half off of our normal timeline" (Technology
Networks, 2025).
The Science Behind the Breakthrough
Enterololin
represents a paradigm shift in antibiotic design, embodying the principle of
"narrow-spectrum" activity that specifically targets disease-causing
bacteria while preserving beneficial microbes. This selectivity addresses one
of the fundamental problems with current antibiotic therapy: the collateral
damage inflicted on helpful bacteria that maintain gut health and protect
against opportunistic infections.
The
drug's target, the LolCDE protein complex, is part of a bacterial transport
system called lipoprotein trafficking. This cellular machinery is essential for
moving proteins from inside bacterial cells to their outer envelope, a process
critical for bacterial survival and virulence. What makes enterololin
particularly clever is that while this transport system exists in many types of
bacteria, the drug appears to exploit subtle differences in how
Enterobacteriaceae species perform this function, explaining its selective
toxicity.
Research
published in Nature has previously demonstrated the clinical
value of narrow-spectrum antibiotics through the discovery of abaucin, an
AI-discovered antibiotic that specifically targets Acinetobacter
baumannii while sparing other bacterial species (Collins et al.,
2023). The abaucin studies showed that narrow-spectrum activity minimizes the
risk of bacteria rapidly spreading resistance against the drug and reduces the
likelihood of causing dysbiosis—the harmful disruption of the gut microbiome
that can lead to secondary infections.
Enterololin's
mechanism appears to leverage what researchers call "bacterial pathway
specificity." While all Gram-negative bacteria express the proteins
involved in lipoprotein trafficking, subtle differences in how different
bacterial species perform this process create opportunities for selective
targeting. "We think it's because A. baumannii does
lipoprotein trafficking a little bit differently than other Gram-negative
species," explained Stokes in previous research on similar narrow-spectrum
antibiotics. "We believe that's why we're getting this narrow spectrum
activity" (MIT News, 2023).
The
clinical implications of this selectivity are profound. Traditional
broad-spectrum antibiotics often create what clinicians call "ecological
disruption" in the gut microbiome, clearing the way for opportunistic
pathogens like Clostridium difficile to establish dangerous
infections. Narrow-spectrum drugs like enterololin could potentially provide
therapeutic benefits without these devastating side effects, offering a more
sustainable approach to treating bacterial infections in IBD patients.
Beyond IBD
While
enterololin was specifically designed to target IBD-associated bacteria, the
AI-driven discovery methodology represents a broadly applicable platform that
could accelerate antibiotic development across multiple disease areas. The
global antibiotic resistance crisis—described by the World Health Organization
as one of the top threats to human health—makes this technological advancement
particularly timely and crucial.
Current
statistics paint a sobering picture of antimicrobial resistance: the US Centers
for Disease Control reported that rates of dangerous bacterial infections
surged by 69% between 2019 and 2023, with so-called "nightmare
bacteria" proving increasingly difficult to treat with existing
antibiotics (Nature, 2025). Globally, 1.1 million deaths annually are linked to
bacterial resistance to antimicrobial drugs, a number projected to increase
dramatically without new therapeutic options.
The
economic burden of antibiotic resistance extends far beyond healthcare costs. A
comprehensive analysis published in PMC estimated that
developing a new antibiotic typically costs 5-10 million euros just for the
early discovery phases, not including the massive investments required for
clinical testing and regulatory approval (David et al., 2021). The high failure
rates and lengthy development timelines have caused many pharmaceutical
companies to abandon antibiotic research entirely, creating a dangerous
innovation deficit precisely when new drugs are most urgently needed.
AI-driven
discovery platforms like the one used to develop enterololin could
fundamentally change this economic equation. By reducing the time and cost
required to identify promising compounds and understand their mechanisms of
action, artificial intelligence makes antibiotic development more financially
attractive while accelerating the pace of innovation. César de la Fuente, a
machine biologist at the University of Pennsylvania who has pioneered AI
approaches to antibiotic discovery, emphasizes this transformation: "The
whole process of discovering a candidate, creating it in the laboratory and
testing it in cells can be done 'within a week or two'" compared to the
traditional timeline of years (Nature, 2025).
The
methodology's broad applicability is already being demonstrated. Stokes'
spin-out company, Stoked Bio, has licensed enterololin from McMaster University
and is currently testing modified versions of the antibiotic against other
drug-resistant bacteria, including Klebsiella, with promising early
results (News Medical, 2025). This suggests that the AI-driven approach could
generate multiple therapeutic candidates from a single discovery platform,
dramatically improving the economics of antibiotic development.
More Than Just Speed
The
successful prediction of enterololin's mechanism of action represents more than
just a time-saving innovation—it demonstrates artificial intelligence's
potential to provide genuine scientific insights that might be impossible for
human researchers to achieve independently. This capability addresses what many
experts consider the central challenge in modern drug discovery: the gap
between identifying active compounds and understanding how they work.
Traditional
drug discovery operates through what pharmaceutical scientists euphemistically
call "phenotypic screening"—essentially, testing thousands of
compounds against disease models and seeing what works, often without any
understanding of why effective compounds produce their beneficial effects. This
approach, while historically successful, creates significant risks when
promising compounds advance to clinical testing without a clear understanding
of their mechanisms.
Regina
Barzilay's DiffDock AI system represents a fundamentally different approach,
using generative artificial intelligence to predict how small molecules will
interact with specific protein targets. The system learns from vast databases
of known protein-drug interactions, developing sophisticated pattern
recognition capabilities that can identify likely binding sites and interaction
mechanisms for novel compounds.
The
implications extend well beyond antibiotic discovery. Research published
in Drug Patent Watch describes how AI platforms are
revolutionizing drug discovery across therapeutic areas by enabling researchers
to screen billions of compounds computationally before investing in expensive
laboratory synthesis and testing (Drug Patent Watch, 2025). This
"virtual-first" approach could dramatically reduce the failure rates
that make pharmaceutical development so costly and risky.
Perhaps
most importantly, AI-driven mechanism prediction could enable researchers to
design drugs with specific mechanisms of action from the outset, rather than
discovering mechanisms after the fact. This capability could lead to more
rational drug design approaches that optimize for specific therapeutic outcomes
while minimizing off-target effects and toxicity.
The
integration of AI into pharmaceutical research is happening at multiple levels
simultaneously. Advanced machine learning algorithms are being applied to
target identification, compound design, toxicity prediction, and clinical trial
optimization. A comprehensive review published in PMC found
that AI applications in drug discovery have shown particular promise in
reducing the time and cost associated with early-stage drug development while
improving success rates in later-stage clinical testing (Serrano et al., 2024).
Clinical Translation
The
path from laboratory discovery to patient treatment remains challenging, even
with AI acceleration. Enterololin must still navigate the complex regulatory
pathway required for all new pharmaceuticals, including extensive safety
testing, clinical trials, and regulatory approval. However, the AI-driven
understanding of the drug's mechanism of action provides significant advantages
for this translation process.
Regulatory
agencies like the FDA increasingly emphasize the importance of understanding
drug mechanisms when evaluating new therapeutic candidates. Clear mechanistic
understanding enables more informed safety assessment, better clinical trial
design, and more rational approaches to dosing and patient selection. The fact
that enterololin's mechanism was predicted and validated before clinical
development begins represents a significant advantage over traditional
discovery approaches.
Stokes
anticipates that enterololin could be ready for human trials within three
years—a timeline that his research team is eager to meet (News Medical, 2025).
This aggressive schedule reflects both the urgent medical need for new IBD
treatments and the advantages provided by AI-driven development approaches.
Jeff Skinner, CEO of Stoked Bio, emphasizes the translation potential:
"The identification of enterololin underscores the remarkable science
emerging at McMaster. We are proud to partner with the university on
translating this breakthrough into real therapies for patients" (News
Medical, 2025).
The
clinical development strategy for enterololin will likely focus on patient
populations with IBD who have failed to respond to conventional therapies or
who experience significant microbiome disruption with broad-spectrum
antibiotics. This targeted approach could demonstrate the drug's value while
building evidence for broader application in IBD management.
Early
clinical trials will need to carefully monitor both therapeutic efficacy and
microbiome preservation, documenting enterololin's narrow-spectrum activity in
human patients. Success in these studies could establish a new paradigm for
antibiotic therapy in IBD, emphasizing selective bacterial targeting rather
than broad microbial suppression.
The
broader implications for clinical practice could be substantial. If enterololin
proves effective in human trials, it could provide clinicians with a new tool
for managing IBD flares without the microbiome disruption associated with
conventional antibiotics. This capability could improve long-term patient outcomes
while reducing the risk of antibiotic resistance development.
Changing the Drug Development Equation
The
economic implications of AI-driven drug discovery extend far beyond reducing
development timelines and costs. By improving success rates and enabling more
rational drug design, artificial intelligence could fundamentally alter the
risk-reward equation that determines pharmaceutical investment decisions and
ultimately affects drug pricing and accessibility.
Current
estimates suggest that developing a new drug costs between $1-3 billion and
takes 10-15 years, with the vast majority of this investment lost when
compounds fail in clinical testing. These economics have created a
pharmaceutical industry focused on blockbuster drugs with broad market appeal
rather than precision medicines for smaller patient populations or rare
diseases.
AI-driven
approaches could enable a more diverse and sustainable pharmaceutical
ecosystem. By reducing early-stage development costs and improving success
rates, artificial intelligence makes it economically viable to develop drugs
for smaller markets and more specific patient populations. This could be
particularly important for conditions like IBD, where patient heterogeneity
means that different therapeutic approaches may be optimal for different
disease subtypes.
The
enterololin discovery demonstrates how AI can compress development timelines
while maintaining scientific rigor. The 18-month time savings achieved through
AI mechanism prediction represents substantial cost reduction, as
pharmaceutical development costs typically accumulate at millions of dollars
per month during active research phases. Multiplied across the entire
pharmaceutical industry, such efficiency gains could significantly reduce drug
development costs and ultimately drug prices.
International
implications are equally significant. The democratization of AI tools could
enable research institutions in developing countries to participate more
actively in drug discovery, potentially addressing global health challenges
that have been neglected by traditional pharmaceutical approaches. Open-source
AI platforms for drug discovery could level the playing field for academic
researchers and smaller biotechnology companies.
The
venture capital and biotechnology investment communities are already responding
to AI's potential in drug discovery. Investment in AI-driven pharmaceutical
companies has increased dramatically over the past several years, with
investors attracted by the potential for higher success rates and faster
returns compared to traditional drug development approaches.
The Next Frontier of AI-Driven Medicine
The
successful prediction of enterololin's mechanism of action represents just the
beginning of AI's potential impact on pharmaceutical research. Current
developments in generative artificial intelligence, machine learning, and
computational biology suggest even more sophisticated applications are on the
horizon.
One
particularly promising area involves the integration of AI drug discovery with
personalized medicine approaches. By analyzing individual patient genomic,
microbiome, and clinical data, AI systems could potentially design
patient-specific therapeutic compounds optimized for individual biological characteristics.
This "precision pharmaceutical" approach could maximize therapeutic
efficacy while minimizing adverse effects.
Research
groups are already exploring the use of AI to design entirely novel antibiotic
molecules from scratch, rather than optimizing existing compounds. A study
published in Cell described the use of deep learning to create
halicin, a compound with novel antibacterial properties that killed many
drug-resistant bacterial strains (Stokes et al., 2020). These
"new-to-nature" compounds could provide therapeutic options that
bacteria have never encountered, potentially circumventing existing resistance
mechanisms.
The
combination of AI with other emerging technologies could create even more
powerful discovery platforms. Integration with high-throughput screening
robotics, advanced microscopy, and real-time analytical chemistry could enable
fully automated drug discovery pipelines that operate continuously without
human intervention. Such systems could screen millions of compounds while simultaneously
predicting their mechanisms, optimizing their properties, and assessing their
safety profiles.
Regulatory
agencies are beginning to develop frameworks for evaluating AI-driven drug
discovery, potentially streamlining approval pathways for compounds developed
using validated AI platforms. The FDA has already approved numerous AI-based
diagnostic tools and is developing guidance for AI applications in drug
development, suggesting that regulatory acceptance of AI-discovered drugs may
be forthcoming.
The
global impact of democratized AI drug discovery tools could be transformative
for addressing neglected diseases that disproportionately affect developing
countries. By reducing the cost and complexity of pharmaceutical research, AI
could enable local research institutions to develop treatments for
region-specific health challenges without requiring massive pharmaceutical
industry investment.
The Reality Check
Despite
its revolutionary potential, AI-driven drug discovery faces significant
challenges that must be addressed before the technology can achieve its full
promise. The enterololin breakthrough, while remarkable, represents a single
success story in a field where statistical validation requires many examples
across diverse therapeutic areas.
One
fundamental limitation involves data quality and availability. AI systems are
only as good as the data used to train them, and pharmaceutical research has
historically generated fragmented, proprietary datasets that may not be suitable
for machine learning applications. Creating comprehensive, high-quality
training datasets will require unprecedented collaboration between
pharmaceutical companies, academic institutions, and regulatory agencies.
The
"black box" nature of many AI systems presents another challenge for
regulatory approval and clinical acceptance. While DiffDock successfully
predicted enterololin's mechanism, the AI couldn't explain its reasoning
process in terms that human scientists could readily understand and validate.
Developing "explainable AI" systems that provide transparent
reasoning for their predictions will be crucial for gaining clinical and
regulatory acceptance.
Intellectual
property and commercialization issues could limit the widespread adoption of AI
drug discovery tools. If AI-discovered drugs prove highly valuable, determining
ownership rights between AI developers, pharmaceutical companies, and research
institutions could become contentious. Clear legal frameworks for AI-generated
intellectual property will be essential for sustainable development of the
field.
The
integration of AI tools into existing pharmaceutical research infrastructure
presents practical challenges. Many research organizations lack the
computational resources and technical expertise required to implement
sophisticated AI systems effectively. Building this capacity will require
substantial investment in both technology and human resources.
Validation
and standardization represent ongoing challenges for the field. While
individual AI systems may show promise in specific applications, establishing
industry-wide standards for AI drug discovery tools will be essential for
regulatory acceptance and clinical adoption. This standardization process
typically takes years and requires extensive collaboration between technology
developers and end users.
Democratizing Drug Discovery
The
potential for AI to democratize drug discovery has profound implications for
global health equity and access to essential medicines. Traditional
pharmaceutical development has been dominated by companies in wealthy countries
focused on diseases affecting profitable markets, leaving many global health
challenges unaddressed.
AI-driven
drug discovery could change this dynamic by dramatically reducing the resources
required for pharmaceutical research. Academic institutions and research
organizations in developing countries could potentially use open-source AI
tools to address local health challenges without requiring the massive
infrastructure investments traditionally associated with drug development.
The
enterololin discovery methodology, if widely disseminated, could enable
researchers worldwide to develop targeted antibiotics for region-specific
bacterial infections. This capability could be particularly valuable for
addressing antimicrobial resistance patterns that vary geographically based on
local antibiotic usage patterns and bacterial populations.
International
collaboration in AI drug discovery could accelerate progress while ensuring
that benefits are shared globally. The open science approach adopted by many AI
research groups, including the teams behind DiffDock and similar platforms, facilitates
knowledge sharing and collaborative improvement of these technologies.
The
World Health Organization has identified antimicrobial resistance as a global
priority requiring coordinated international response. AI-driven antibiotic
discovery platforms could provide the technological foundation for a
coordinated global effort to develop new antibiotics targeting the most
dangerous resistant pathogens.
Training
and capacity building will be crucial for realizing AI's potential for global
health applications. International programs to develop computational biology
and AI expertise in developing countries could create the human resources
needed to implement these technologies effectively worldwide.
Conclusion:
The
discovery of enterololin and the successful AI prediction of its mechanism of
action mark a watershed moment in the history of pharmaceutical research. For
the first time, artificial intelligence has demonstrated the ability to not
just identify promising therapeutic compounds, but to predict how they work
with remarkable accuracy and unprecedented speed.
This
breakthrough represents more than just a new antibiotic or a faster way to
understand drug mechanisms—it embodies the potential for artificial
intelligence to transform our entire approach to medicine. By compressing years
of laboratory work into minutes of computation, AI enables a more rational,
efficient, and ultimately successful approach to drug discovery that could
address some of humanity's most pressing health challenges.
The
implications extend far beyond inflammatory bowel disease. The AI-driven
methodology that produced enterololin could be applied to any disease where new
therapeutic approaches are needed, from antibiotic-resistant infections to cancer,
neurological disorders, and rare genetic diseases. By making drug discovery
faster, cheaper, and more successful, artificial intelligence could usher in an
era of pharmaceutical abundance rather than scarcity.
Yet
the technology's greatest promise may lie in its potential to democratize drug
discovery, enabling researchers worldwide to address local health challenges
with sophisticated computational tools that were previously available only to
major pharmaceutical companies. This democratization could help address global
health inequities while accelerating progress on neglected diseases that affect
the world's most vulnerable populations.
The
path from breakthrough to widespread clinical application will undoubtedly
present challenges. Regulatory frameworks must evolve to accommodate
AI-discovered drugs, healthcare systems must adapt to new therapeutic
approaches, and the global research community must develop new standards for
AI-driven pharmaceutical research. But the potential rewards—better treatments,
lower costs, and more equitable access to essential medicines—justify the
effort required to overcome these obstacles.
As Dr.
Stokes aptly summarizes his approach: "Drug resistance and our lack of new
drugs is a leaky faucet. You can leave it be for a while, but you're eventually
going to have a big problem. AI is my wrench—it's a tool for fixing the leak
before it becomes a flood" (Technology Networks, 2025). The enterololin
discovery suggests that we now have that wrench in hand, ready to fix problems
we've struggled with for decades.
The
revolution in AI-driven drug discovery is no longer coming—it's here, swimming
in vast oceans of biological data, guided by artificial intelligence that can
see patterns invisible to human eyes and predict outcomes we couldn't imagine.
The age of intelligent medicine has begun, and for millions of patients
suffering from inflammatory bowel disease and countless other conditions, that
intelligence offers something precious: hope backed by science, precision
guided by artificial intelligence, and the promise that tomorrow's medicine
will be fundamentally better than today's.
References
Bel, S., et al. (2024). Antibiotic-induced disruption of
intestinal mucus barrier function. Science Advances, 10(37),
eadj8090. https://doi.org/10.1126/sciadv.adj8090
Collins, J. J., et al. (2023). Deep learning-guided
discovery of an antibiotic targeting Acinetobacter baumannii. Nature
Chemical Biology, 19(11), 1342-1350. https://doi.org/10.1038/s41589-023-01349-8
Dar, S. H., et al. (2023). The association of antibiotic
exposure with new-onset inflammatory bowel disease: A systematic review and
meta-analysis. Journal of Gastroenterology and Hepatology, 38(7),
1089-1098. https://doi.org/10.1111/jgh.16142
David, L., et al. (2021). Artificial intelligence and
antibiotic discovery. Antibiotics, 10(11), 1376. https://doi.org/10.3390/antibiotics10111376
Drug Patent Watch. (2025, September 8). AI in action:
Accelerating the drug discovery pipeline. Drug Patent Watch Blog. https://www.drugpatentwatch.com/blog/ai-in-action-accelerating-the-drug-discovery-pipeline/
MIT News. (2023, May 24). Using AI, scientists find a drug
that could combat drug-resistant infections. MIT News Office. https://news.mit.edu/2023/using-ai-scientists-combat-drug-resistant-infections-0525
Nature. (2025, October 2). AI has designed thousands of
potential antibiotics. Will any reach patients? Nature News. https://doi.org/10.1038/d41586-025-03201-6
News Medical. (2025, October 3). AI serves as a tool to
fight drug resistance and accelerate new antibiotic development. News
Medical Life Sciences. https://www.news-medical.net/news/20251003/AI-serves-as-a-tool-to-fight-drug-resistance-and-accelerate-new-antibiotic-development.aspx
Serrano, D. R., et al. (2024). Artificial intelligence (AI)
applications in drug discovery and development. Pharmaceuticals, 17(10),
1328. https://doi.org/10.3390/ph17101328
Stokes, J. M., et al. (2020). A deep learning approach to
antibiotic discovery. Cell, 180(4), 688-702. https://doi.org/10.1016/j.cell.2020.01.021
Stokes, J. M., et al. (2025). Discovery and artificial
intelligence-guided mechanistic insight of enterololin: A narrow-spectrum
antibiotic targeting Enterobacteriaceae. Nature Microbiology, 10(1),
142-156. https://doi.org/10.1038/s41564-025-02142-0
Technology Networks. (2025, October 2). AI accurately
predicts how a new IBD antibiotic works. Technology Networks. https://www.technologynetworks.com/drug-discovery/news/ai-accurately-predicts-how-a-new-ibd-antibiotic-works-405431
