Agentic artificial intelligence (AI) systems are emerging as teams of intelligent computational experts capable of rivaling human performance in labor-intensive tasks, including literature review, hypothesis formulation, data analysis and model interpretation. These systems are poised to accelerate labor-intensive biomedical research by making autonomous decisions based on contextual information and expert feedback.
Agentic AI systems have been developed for various applications, including drug discovery, data analysis and biomarker identification; however, several distinct challenges remain for making these systems broadly deployable in biomedical research. A new review published in Nature biotechnologies discusses three key algorithms and seven foundational building-block characteristics that contribute to the development of agentic AI systems. It highlights their biomedical applications, design considerations and the challenges and opportunities associated with deploying agentic AI systems to advance collaborative scientific research.
Reliance on cell lines and animals for toxicity testing has long been the cornerstone of preclinical drug discovery for intestinal diseases. However, the physiology, genetics, and disease etiology of animals differ significantly from humans, particularly in complex disorders such as inflammatory bowel disease (IBD), in which attrition rate is over 85%.
Regulatory shifts by the FDA with the Modernization Act 2.0 and subsequent New Approach Methodologies (NAMs) guidance are accelerating the transition toward human-relevant systems. Human intestinal organoids (HIOs) have emerged as transformative tools that faithfully replicate the architecture, function, and cellular diversity of the human gut. Advances in the HIOs allow study of drug absorption, metabolism, and toxicity in a dish, providing a bridge between in vitro assays and clinical outcomes to offer new opportunities for improving the prediction of toxicokinetics and pharmacokinetics.
Can we conclude on mutagenicity classification based in in vitro data only?
Currently, the genotoxicity testing strategy involves a step-wise approach, starting with in vitro tests that cover three key endpoints: gene mutations, clastogenicity and aneugenicity. If in vitro tests yield positive results, follow-up confirmatory in vivo tests are required. In a new study, the JRC analysed the EURL-ECVAM genotoxicity database which includes over 700 chemicals, exploring six combinations of in vitro tests that cover all the three key endpoints. Findings indicate that it may, in some cases, be possible to classify a chemical as a somatic cell mutagen using in vitro assays alone. Novel in vitro assays that address specific biological mechanisms providing supporting information in a weight of evidence may help increasing confidence in the results.
Putting Science into Standards – 3D Bioprinting: Towards Standards in Biomedicine
3D bioprinting stands at the forefront of biomedical innovation, offering transformative potential for creating functional tissues, thereby advancing regenerative medicine and in vitro testing methods. However, challenges such as reproducibility, scalability, and bioink standardisation remain. At the PSIS (Putting Science into Standards) workshop, with 56 experts from 17 countries, participants focused on standardisation needs for bioinks, biomaterials, printing equipment, and quality and safety assessments, paving the way for impactful advancements. This collaborative approach underscores the need for updated regulatory frameworks to safely integrate 3D bioprinting into healthcare, enhancing personalised medicine and fostering groundbreaking biomedical applications.
Jean-Laurent Casanova, MD, PhD, the Levy Family Professor at Rockefeller University, pediatrician at the Necker Hospital for Sick Children and Imagine Institute in Paris, and investigator of the Howard Hughes Medical Institute, known for his discovery of the human genetic and immunological determinants of infectious diseases, is the recipient of the 2026 Mechthild Esser Nemmers Prize in Medical Science at Northwestern University.
In nearly three decades, between the US and France, Casanova and his lab have discovered the rare and common genetic causes of more than 20 infections. They also discovered what they call “phenocopies” of these genetic disorders; some patients suffer from severe infection because of auto-antibodies that neutralize a component of host defense that is mutated in other patients with the same infection.
The UK Medical Research Council is inviting proposals to host a Pre-Clinical Hub ‘to bring together cutting-edge human disease modelling capabilities and essential data to support human in vitro model development for translational research’.
It is part of a broader initiative backed by £30m of funding, which has the stated aims to create more accurate predictive pre-clinical models of human disease ; reduce the use of animals by increasing the range of robust alternative methods ; be a national resource for in vitro model development and increase access to state-of-the-art technology, human tissue and patient samples and data across the UK ; and position the UK at the forefront of human in vitro model development for medical research. A webinar will be hosted on March 25th and the deadline to apply is May 21st, 2026 (4:00 pm).
Check out more fundings in the NAMs space
The Tufts Center for the Study of Drug Development (Tufts CSDD) is conducting a study to assess the use of modeling and simulation in drug discovery and development. The results of this study will provide drug development professionals with valuable insight on the advantages and challenges of modeling and simulation.
Your experience and expertise will help benchmark the current use and how it is changing the landscape. The survey takes approximately 10 minutes to complete.

Liquid AI and Insilico Medicine have released LFM2‑2.6B-MMAI, a general small AI foundation model trained into a system intended to support multiple stages of drug discovery while allowing companies to keep their data inside their own computing environments without transmitting it to external cloud services.
The 2.6‑billion-parameter scientific foundation model is designed to run on private pharmaceutical infrastructure and was trained with Insilico Medicine’s recently launched Science MMAI Gym training framework, a framework that helps turn general Large Language Models (LLMs) into specialized biomedical and chemical tools. The developers state that the model is designed to support tasks across the drug discovery workflow.
Foetal bovine serum (FBS) has been widely used as a nutrient-rich supplement in mammalian cell culture for over 6 decades; however, its usage has increasingly raised various concerns and challenges related to quality variations, unethical collection practices, supply-demand imbalance and regulatory challenges.
Starting from a comprehensive analysis of components of FBS and their functions in cell growth, a new review compares the main types of FBS alternatives, i.e., human and animal-derived, plant-based alternatives and serum free media. Future perspectives discussed include the development of application-specific FBS alternatives, improvements in the quality and specialized formulation of FBS, optimization of existing alternatives and the establishment of databases and incentive mechanisms to facilitate the transition away from FBS.
Read the publication in Frontiers
Organoids are among the most advanced in vitro models for studying human biology, offering strong potential across research and medicine. However, their usefulness is still limited by incomplete maturation, variability between samples, limited long-range interactions, and especially the lack of functional blood vessel systems. Integrating organoids into microfluidic platforms can help address some of these problems by creating controlled environments with precise physical and biochemical conditions.
A major focus in improving these systems is vascularization, which enables the formation of stable, perfusable networks of organoids. A new review discusses the biological rationale, current strategies, and technical considerations for vascularizing organoids-on-chip, highlighting their potential to improve physiological relevance, functional performance, personalization and translational applicability.
Read the publication in Lap on a Chip
The FDA Modernization Act 2.0 has shifted from animal testing to NAMs for preclinical drug development. Three-dimensional (3D) bioprinting has emerged as a premier NAM, capable of fabricating human-relevant complex models with high fidelity.
A review surveys the primary bioprinting modalities and analyzes the operating mechanisms, advantages, and limitations of each technique, from high-resolution inkjet patterning to the rapid, layerless fabrication enabled by volumetric methods. Although challenges regarding regulatory validation, vascularization, and anatomical complexity persist, advanced bioprinting strategies incorporating biochemical cues, cellular diversity, and multi-tissue interactions are paving the way for more predictive and humane therapeutic pipelines.This work outlines the technological landscape and future directions for biofabrication in a post-animal testing era.
Read the publication in npj biomedical innovations
Twenty years have passed since the first demonstration of mouse induced pluripotent stem cells (iPSCs). What began as an unexpected observation in Kyoto quickly transformed stem cell biology and regenerative medicine worldwide.
Over the past two decades, scientists have gained profound insights into the molecular mechanisms underlying cellular reprogramming and pluripotency. The technology has continued to evolve — becoming safer, more efficient, and more versatile. Today, iPSCs serve as a foundation for wide-ranging applications, from disease modeling and drug discovery to regenerative therapies and rejuvenation research. In a new perspective paper, the 2012 Nobel Prize winner Dr. Yamanaka reflects on the scientific journey of iPSCs, highlights key milestones in our understanding of reprogramming, and discusses the expanding clinical and societal impact of iPSCs.
Read the perspective in Cell Stem Cell
