Lightnews — Scholar-powered news

César de la Fuente @delafuentelab.bsky.social · 21h

Thank you!

Reposted by César de la Fuente

UofT-TCAIREM @uoft-tcairem.bsky.social · 2d

🎥 @delafuentelab.bsky.social

Uncovering new #antibiotics with #AI | Penn Today

penntoday.upenn.edu/news/uncover...

#MedSky #AImedicine #AI #HealthAI #AIinHealthcare #MedTech

Uncovering new antibiotics with AI | Penn Today

César de la Fuente leverages machine learning to accelerate the discovery of lifesaving drugs and help reduce antibiotic resistance, a rising global health problem.

penntoday.upenn.edu

1 1

César de la Fuente @delafuentelab.bsky.social · 6d

Thank you so much for highlighting our latest work!

Nature Biotechnology @natbiotech.nature.com · 8d

Research Highlight: Mining archaea for antibiotics www.nature.com/articles/s41...

Mining archaea for antibiotics - Nature Biotechnology

Nature Biotechnology - Mining archaea for antibiotics

www.nature.com

2 4

Reposted by César de la Fuente

Nature Biotechnology @natbiotech.nature.com · 8d

Research Highlight: Mining archaea for antibiotics www.nature.com/articles/s41...

Mining archaea for antibiotics - Nature Biotechnology

Nature Biotechnology - Mining archaea for antibiotics

www.nature.com

1 2

César de la Fuente @delafuentelab.bsky.social · 6d

AI has already accelerated our ability to discover new antibiotic candidates so I think this is the beginning, though a lot more work needs to be done to take these candidates through clinical trials.

1

César de la Fuente @delafuentelab.bsky.social · 8d

En España mueren más de 4.000 personas al año por infecciones resistentes a los antibióticos. Urge innovar, sensibilizar y actuar. Gracias COPE por difundir nuestras últimas investigaciones. 🎙️🔬
www.cope.es/programas/he...

Las bacterias, cada vez más resistentes: "Mueren más personas al año por resistencia a los antibióticos que en accidentes de tráfico"

Jorge Alcalde, desgrana en 'Herrera en COPE', los hallazgos más sorprendentes de la ciencia que conecta el pasado con el presente

www.cope.es

1 1

César de la Fuente @delafuentelab.bsky.social · 10d

(2/2)We can now mine genetic data from extant and ancient biology for new compounds, design new-to-nature molecules from scratch using generative AI, and develop models to outsmart emerging resistant bacteria. Thank you @asm.org for covering our @upenn.edu lab's work! @pennmedcso.bsky.social

Harnessing AI to Revolutionize Antibiotic Discovery | ASM.org

The world desperately needs new antibiotics. Scientists are leveraging AI to accelerate drug discovery, transforming how—and where—we look for new compounds to defeat resistant bacteria.

asm.org

4

César de la Fuente @delafuentelab.bsky.social · 10d

(1/2) AI is redefining antibiotic discovery. What once took years with traditional methods can now be done in hours.

2 1

Reposted by César de la Fuente

Quincemil @quincemil.bsky.social · 13d

El equipo del gallego César de la Fuente presenta una plataforma de IA que diseña antibióticos nuevos #Galicia

El equipo del gallego César de la Fuente presenta una plataforma de IA que diseña antibióticos nuevos

ApexAmphion combina inteligencia artificial y aprendizaje por refuerzo para generar moléculas capaces de combatir bacterias resistentes Te puede interesar: ITG acoge en A Coruña una jornada internacional y gratuita sobre el futuro del cerebro humano

dlvr.it

1 1

Reposted by César de la Fuente

Rafael Laso Pérez @gecko1990.bsky.social · 15d

Excited to highlight our News & Views on the article by Torres et al. from @delafuentelab.bsky.social team 🚀 Archaea could help to discover new antibiotics!

Read it here:
www.nature.com/articles/s41...

Deep-mining the archaeal proteome for antibiotics - Nature Microbiology

A deep-learning algorithm unravels a collection of archaeasins, peptides from the archaeal proteome with potential antimicrobial activity and implications for the development of next-generation antibiotics

www.nature.com

1 3 4

César de la Fuente @delafuentelab.bsky.social · 13d

Amazing work!

1

Reposted by César de la Fuente

Nature Microbiology @natmicrobiol.nature.com · 15d

Deep-mining the archaeal proteome for antibiotics

News & Views by Rafael Laso-Pérez

@gecko1990.bsky.social @delafuentelab.bsky.social

Read it here: rdcu.be/eHNwa

Deep-mining the archaeal proteome for antibiotics

Nature Microbiology - A deep-learning algorithm unravels a collection of archaeasins, peptides from the archaeal proteome with potential antimicrobial activity and implications for the development...

rdcu.be

1 7 18

Reposted by César de la Fuente

Wilson Lab @wilsonlab.bsky.social · 15d

Deep-mining the archaeal proteome for antibiotics www.nature.com/articles/s41...

Deep-mining the archaeal proteome for antibiotics - Nature Microbiology

A deep-learning algorithm unravels a collection of archaeasins, peptides from the archaeal proteome with potential antimicrobial activity and implications for the development of next-generation antibiotics

www.nature.com

1 3

César de la Fuente @delafuentelab.bsky.social · 13d

(9/9)Incredible Team work: Hanqun Cao, Marcelo D. T. Torres, Jingjie Zhang, Zijun Gao, Fang Wu, Chunbin Gu, Jure Leskovec, Yejin Choi, Guangyong Chen, Pheng-Ann Heng

#ReinforcementLearning #AI #DrugDiscovery #Antibiotics #AMR #ProteinLM #SyntheticBiology #Biotech #ApexAmphion

A deep reinforcement learning platform for antibiotic discovery

Antimicrobial resistance (AMR) is projected to cause up to 10 million deaths annually by 2050, underscoring the urgent need for new antibiotics. Here we present ApexAmphion, a deep-learning framework ...

arxiv.org

2

César de la Fuente @delafuentelab.bsky.social · 13d

(8/9)We’re compressing years of trial-and-error into hours-scale loops, expanding the space of antimicrobial discovery, and executing a programmable RL pipeline that moves from games to potential life-saving therapeutics.

1

César de la Fuente @delafuentelab.bsky.social · 13d

(7/9)In other words, our ApexAmphion model has now mastered antibiotic discovery and can design promising candidates within hours, often in a single morning or afternoon.

1

César de la Fuente @delafuentelab.bsky.social · 13d

(6/9)• Multi-objective rewards: predicted MIC + key physicochemical properties
• Rapid & steerable: tune potency and developability on demand
• Amphorium: >2 million machine-annotated candidates for follow-up

1

César de la Fuente @delafuentelab.bsky.social · 13d

(5/9)Iteration by iteration, the system learns to design better molecules.
𝐇𝐢𝐠𝐡𝐥𝐢𝐠𝐡𝐭𝐬 (𝐠𝐫𝐨𝐮𝐧𝐝-𝐭𝐫𝐮𝐭𝐡 𝐞𝐱𝐩𝐞𝐫𝐢𝐦𝐞𝐧𝐭𝐬):
• 100/100 designed molecules active in vitro (100% hit rate)
• 99/100 active against ≥2 clinically relevant pathogens (incl. MDR)

1

César de la Fuente @delafuentelab.bsky.social · 13d

(4/9)Instead of chess moves, the agent proposes peptide sequences (candidate antibiotics). Instead of a scoreboard, it receives multi-objective rewards tied to predicted potency (minimum inhibitory concentration, MIC) and developability—properties that make a drug more likely to succeed.

1

César de la Fuente @delafuentelab.bsky.social · 13d

(3/9)In games, RL agents explore vast decision spaces and optimize toward rewards (winning). ApexAmphion brings the same logic to biology.

1

César de la Fuente @delafuentelab.bsky.social · 13d

(2/9)I’m thrilled to introduce ApexAmphion—inspired by the systems that mastered Go and StarCraft. We pair a 6.4B-parameter protein language model with PPO in a closed loop to generate → score → optimize antibiotic candidates at digital speed.

A deep reinforcement learning platform for antibiotic discovery

Antimicrobial resistance (AMR) is projected to cause up to 10 million deaths annually by 2050, underscoring the urgent need for new antibiotics. Here we present ApexAmphion, a deep-learning framework ...

arxiv.org

1

César de la Fuente @delafuentelab.bsky.social · 13d

(1/9)For over a century, antibiotics were found by slow, painstaking, serendipitous screening.
Today we share a different path: the first demonstration of deep reinforcement learning (RL) for antibiotic discovery.

1 1 6

César de la Fuente @delafuentelab.bsky.social · 15d

Over 200 leaders — including 10 Nobel laureates & top AI experts — call for AI red lines to prevent dangerous uses. I’ve signed the Global Call for AI Red Lines for responsible use of AI: red-lines.ai

200+ prominent figures endorse Global Call for AI Red Lines

AI could soon far surpass human capabilities and escalate risks such as engineered pandemics, widespread disinformation, large-scale manipulation of individuals including children...

red-lines.ai

1 3

Reposted by César de la Fuente

Philadelphia Inquirer @inquirer.com · Sep 3

Scientists are using tools and technologies like artificial intelligence to look for new medicines in plants and other natural products, advancing a centuries-old approach to drug discovery.

How Philadelphia scientists turn toxic fungi, snake venom, and trees into medicine

Scientists are using new tools and technologies like artificial intelligence to look for new medicines in plants and other natural products, advancing a centuries-old approach to drug discovery.

www.inquirer.com

9 81 380

César de la Fuente @delafuentelab.bsky.social · 16d

Grateful to @inquirer.com and Kela Yub for the feature on our lab's work. Using AI, we are unlocking the hidden therapeutic potential of nature, including in venoms, turning centuries-old drug discovery into approaches that reveal antibiotic candidates in minutes.
Congrats to Xue Sherry Gao as well

How Philadelphia scientists turn toxic fungi, snake venom, and trees into medicine

Scientists are using new tools and technologies like artificial intelligence to look for new medicines in plants and other natural products, advancing a centuries-old approach to drug discovery.

www.inquirer.com

5