SLAS Europe 2026 Scientific Podium Sessions
Recorded On: 05/20/2026
Scientific Podium Presentations from three educational tracks:
The SLAS Scientific Program Committee selects conference speakers based on the innovation, relevance and applicability of research as well as those that best address the interests and priorities of today’s life sciences discovery and technology community. All presentations are published with the presenters' permission.
- Advances in Laboratory Automation
- Advances in Drug Discovery
- Screening Applications and Diagnostics
Conference Chairs:
James Pilling, MS
Associate Principal Scientist
AstraZeneca (England)
Vivian Lu Tan, PhD
Managing Director
Vienna BioCenter Core Facilities (Austria)
Key:
Complete
Next
Failed
Available
Locked
Keynote
Opening Keynote: Nanomedicine-based immunotherapy – Concepts, automation & clinical translation
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Immunotherapy is transforming the treatment of cancer, autoimmune diseases, and inflammatory conditions. While antibody- and cell-based approaches dominate current clinical practice, nanomedicine offers a powerful complementary modality with the potential to reprogram the immune system and achieve durable disease modification.
This keynote presentation highlights the translational journey of an innovative class of nanomedicines inspired by the body's endogenous lipoproteins. These dynamic assemblies of apolipoproteins and lipids can be recapitulated in the laboratory using cutting-edge recombinant technologies and microfluidics, creating versatile platforms for advanced drug delivery and controlled release.
In our research, we demonstrate how apolipoproteins provide exceptionally robust scaffolds for engineering nanomedicines that exhibit superior safety profiles and inherent targeting to key immune cell populations. Building on more than two decades of bioengineering progress, the presentation will showcase the development of nanomedicine-based immunotherapies that deliver long-term therapeutic benefits in immune-mediated diseases.
The presentation will additionally lay out entrepreneurial pathways and laboratory automation strategies to facilitate clinical translation,
Willem Mulder, PhD - Professor of Precision Medicine
Eindhoven University of Technology & Radboud University Medical Center
Advances in Laboratory Automation
Session: Next-Gen Discovery: Emerging Cellular & Biophysical Technologies
Increasingly complex biological targets and innovative drug modalities require ultra-sensitive technologies able to detect weak and transient interactions, capture dynamic processes in real-time, and interrogate target structure to power translation. This session will explore how innovative cellular and biophysical methods accelerate the development of diverse therapeutics including proximity-based modalities.
Induced Proximity Therapeutics for Cancer
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Targeted protein degradation and other induced proximity technologies are revolutionizing drug discovery. As PROTAC technology advanced through the development of high-throughput chemical synthesis, direct testing in cell-based assays, and understanding drug optimization, it also became clear that PROTACs target space is limited by availability of a target binder.
However, the success of Cereblon E3 ligase and the ability of its ligands to act as molecular glues, has led to rapid development of multiple molecular glue degraders and opened up a new way of drug screening based on protein-protein interactions. Molecular glues, that bind at the interface of a protein-protein complex tap into entirely different drug target space, previously thought of as undruggable.
At the ICR, we created a proprietary Cereblon molecular glue library, and we are exploring its target degradation potential through high-throughput proteomics profiling, target-focused- and phenotypic screening. We are also expanding targeted protein degradation technology by exploring novel E3 ligases. In my talk I will discuss how these efforts contribute to discovery of novel clinically relevant targets and cancer therapeutics.
Agnieszka Konopacka, PhD
How to Discover Molecular Glues de Novo
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Molecular glues are an emerging class of agents in drug discovery, enabling new mechanisms to modulate protein function through stabilization of protein-protein interactions. Despite their promise, discovering synthetic molecular glues de novo remains highly challenging, as productive hits must simultaneously engage a target and induce cooperative ternary complex formation. Several clinically and biologically impactful molecular glues, including CRBN acting thalidomide, lenalidomide, and the plant hormone auxin acting through TIR1, are small and structurally simple, illustrating that modest chemical matter can elicit profound biological effects when identified using appropriate discovery strategies
A central barrier in early molecular glue discovery is the complex behavior of multi-component assemblies. Initial hits must satisfy dual requirements: sufficient interaction with one binding partner and the ability to promote cooperative stabilization of a ternary complex. This "double filter" presents a high threshold for conventional screening approaches, which are typically optimized to rank compounds by affinity rather than by cooperativity. As a result, weak or transient binders with strong cooperative potential are frequently missed, highlighting the need for screening and profiling assays that directly report on ternary complex formation.
Here, we present a biophysics-driven framework for the de novo discovery of molecular glues across challenging targets, with a particular focus on cooperativity as a principle. The approach to optimize screening and follow-up methodologies that quantitatively relate binding affinity to ternary complex stabilization, allowing do derive cooperativity-affinity relationship early on. We further describe assay formats suitable for characterizing covalent molecular glues, including SPR-based Capture-Chase methodologies that provide mechanistic insight into irreversible ternary complex formation. Together, this framework outlines a generalizable strategy for identifying, validating, and prioritizing molecular glue candidates and aims to lower the barrier to discovering this emerging class of next-generation therapeutics.
Pim De Vink, Senior Scientist
Accelerating Molecular Glue Discovery Through Ultra-High-Throughput Profiling of a Continuously Expanding E3-Ligase-Biased Library
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Targeted protein degradation via molecular glues is a promising means to drug previously "undruggable" proteins. These small molecules stabilize novel protein-protein interactions that can ultimately result in protein destruction through downstream cellular processes. These complexes often consist of a degrader molecule, a target protein, and an E3 ubiquitin ligase.
To expedite the identification of molecular glues for novel targets, Amgen is developing biased chemical libraries of E3-ligase ligands. Our Small molecule Effector Library (SEL) is a collection of E3-binding scaffolds decorated with diverse fragments. This library can be quickly produced, resulting in a rapidly growing screening deck.
This presentation will describe Amgen's Automated Screening Platform team's efforts to adapt traditional ultra-high throughput screening workflows to the new SEL screening paradigm. This has required reevaluations at every stage of our processes from assay validation to data aggregation. We have tested the library against several assay technologies with a primary focus on HiBiT degradation assays using cell lines expressing tagged targets of interest. To date, over 60 HiBiT cell lines have been screened, generating millions of data points. SEL screening has identified several lead chemical series for novel targets of interest and will continue to serve as an integral part of our targeted protein degradation platform.
Franck G. Madoux, PhD
Session: From Patient Samples to Precision Medicine: Building the Infrastructure for Translational Research
High-quality biological samples are the foundation of translational research and personalised healthcare. As studies expand to larger patient cohorts and increasingly data-rich analytical platforms, maintaining sample integrity, traceability, and scalability across the research workflow becomes critical. This session explores the ecosystem that supports modern biomedical discovery—from patient cohort sampling and biospecimen acquisition through to large-scale storage infrastructure and advanced analytical technologies. Speakers will highlight how coordinated approaches to sample management, research infrastructure, and emerging diagnostic platforms enable more reliable and scalable discovery, bridging the gap between patient-derived samples and precision diagnostics to support more predictive, data-driven research.
NIHR BioResource : Advancing translational research
Open to view video. | Closed captions available
Open to view video. | Closed captions available
The NIHR BioResource is a unique recallable resource of well characterised patients and healthy volunteers willing to participate in biomedical research. The BioResource includes large cohorts of patients with a Rare or Common condition (e.g. IBD, IMID, NAFLD/MASLD, as well as Mental Health and COVID-19). Through its close relationship with NHSBT, the NIHR BioResource has recruited blood donors at scale. Lately, the NIHR BioResource has increased its EDI work, launching programmes dedicated to young people (D-CYPHR) and to individuals from Black communities (IBHO). The NIHR BioResource has enrolled over 350,000 participants to date, holds >11TB of genomic data and >1.6M biological samples. Automation for sample receipting, processing, storage and retrieval is central to all BioResource operations. The NIHR BioResource has helped delivering major programmes with direct impact in UK clinical care delivery including, most recently, the Blood Group Genotyping Programme that is now embedded in the NHS.
Nathalie Kingston, PhD
High-throughput single-cell functional screening diagnostic enables rapid therapy selection and clinical benefit in late-stage cancer: first personalized medicine EXALT-1 trial results and randomized EXALT-2 trial feasibility
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Precision medicine (PM) offers a promising approach to guide treatment decisions for patients with relapsed or refractory blood cancers. Two key PM strategies have emerged: genomic precision medicine (gPM), which identifies actionable genetic alterations, and functional precision medicine (fPM), which measures drug responses directly in patient-derived cells. While gPM has advanced significantly, it yields effective therapy matches in fewer than 10% of cancer cases. In the first fPM trial, EXALT-1 (NCT04470947), we demonstrated that ex vivo single-cell drug response profiling enabled personalized treatment selection, resulting in meaningful clinical benefit for 54% of patients with advanced lymphoma or leukemia. Notably, 40% of responders experienced exceptional outcomes, with treatment durations exceeding threefold the historical expectation for their disease.
Building on this foundation, we developed a single-cell high-throughput flow cytometry-based drug screening assay that efficiently profiles malignant and microenvironmental cell populations, even from limited clinical samples requiring small sample volumes. This innovation addresses a major barrier to applying functional precision medicine across diverse cancer types. In the ongoing EXALT-2 trial (NCT04470947), a multicenter, randomized controlled study, we demonstrated the clinical feasibility of the assay, including its rapid turnaround times (3 days from biopsy)-essential for timely clinical decision-making. In the feasibility study cohort, functional profiling consistently delivered actionable results within timeframes compatible with routine oncology workflows. These results underscore the feasibility and impact of high-throughput, single-cell fPM testing in a multicenter randomized setting. Functional testing not only complements genomic profiling but also offers a faster and broader approach to personalized therapy. Final clinical outcome data from EXALT-2 will further define the role of functional precision medicine in transforming treatment paradigms for hematologic malignancies and beyond
Alexander Pichler
Large scale automated sample storage - design, optimization and integration with laboratory operations
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Designing large-scale automated storage requires balancing capacity, reliability, and seamless laboratory integration. This presentation explores key architectural decisions, workflow optimization, and data connectivity needed to support high-throughput environments. Drawing on real-world implementation experience, it highlights strategies to enhance sample integrity, operational efficiency, and scalability while enabling future-ready laboratory infrastructure.
Tony Cox, PhD
Session: Automating Drug Discovery in 3D-Models
As 3D cell culture, organoids, and complex co-culture systems become increasingly central to drug discovery and disease modeling, laboratories face new challenges in scalability, reproducibility, and integration with high-throughput workflows. This session highlights how advances in laboratory automation, robotics, liquid handling, and imaging are transforming the use of physiologically relevant models in research and development. Presentations will showcase innovative strategies for handling fragile 3D structures, adapting assay design for miniaturized and multiplexed readouts, and integrating multi-omics and biophysical measurements with automated platforms. Attendees will gain insight into how automation is bridging the gap between model complexity and experimental throughput, enabling more predictive, data-rich discovery pipelines.
Session Chair: Robin Pronk, BrainZell
In-organoid HCS: perspectives on cancer therapy development and synergistic drug discovery by harnessing cohorts of PDOs
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Identifying therapies with potential clinical relevance is a critical task for intervening on solid malignancies such as pancreatic ductal adenocarcinoma. High-throughput (HTS) holds promise as large-scale approach to uncover newer cancer therapies, especially when this interrogation accounts in highly physiological-competent human models such as collections of patient-derived tumor organoids (PDO)1. PDOs provide accurate replicas of the patient's tumor tissue from where they are established, with histopathological molecular and functional characteristics recapitulated at high fidelity, including the response to chemotherapy2-5. Within this framework, PDO collections enable to reconstruct the genotypic, epigenetic and phenotypic landscapes observed in tumors6, facilitating the representation of the diversity of tumor pathways activity on a patient-by-patient basis. Drug pharmacotyping in PDOs2,7 has emerged as an applicable approach for predictive personalized medicine scrutiny6. The correspondence between 'in-organoid' drug response and patient's clinical outcomes is expected to support the nomination of patients who are more alike to benefit from the treatment assessed. Current approaches based on single-agent PDO-prediction remain very variable, and introduction of multi-drug combination testing is required to tackle drug resistance and enhance the translational relevance of PDOs.
Drug response profiling in PDO cohorts provides a robust drug discovery pipeline8,9 by means of tumor diversity representation. High-content screening (HCS), which encompasses the automated acquisition and analysis of microscopy images10, is an eligible technology for drug discovery used to interrogate the effects of genetic and chemical perturbations at the phenotypic level in unbiassed manner. PDO-HCS has been explored for applications such chemical and antibody11 screening, drug discovery and target validation8,12, or synergy scoring including the profiling of de novo therapy-induced cytotoxicities8,9. However, achieving sufficient numbers of cells to generate the required number of organoid plates for HTS, and in extension, for more complex assays like drug synergism interrogation poses a significant challenge. Here, I will provide a perspective on the design of a PDO-based HCS pipeline that build confidence and support generation of highly significant hits while ensuring scalability, versatility and robustness.
1. Clevers, H. Cell 165, 1586-1597 (2016).
2. van de Wetering, M. et al. Cell 161, 933-45 (2015).
3. Boj, S. F. et al. Cell 160, 324-338 (2015).
4. Sato, T. et al. Gastroenterology 141, 1762-1772 (2011).
5. Kondo, J. et al. Proc. Natl. Acad. Sci. 108, 6235-6240 (2011).
6. Verstegen, M. M. A. et al. Nat. Med. 31, 409-421 (2025).
7. Tiriac, H. et al. Cancer Discov. 8, 1112-1129 (2018).
8. Cutrona, M. B., et al. iScience 27, 110289 (2024).
9. Mertens, S. et al. Cell Rep. 42, 112324 (2023).
10. Giuliano, K. A. et al. SLAS Discov. 2, 249-259 (1997).
11. Herpers, B. et al. Nat. Cancer 3, 418-436 (2022).
12. Betge, J. et al. Nat. Commun. 13, 3135 (2022).
Meritxell B. Cutrona
High-Throughput Screening of Clinically Used Neurological Therapeutics in Human Brain Organoids for Effect Profiling
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Accurate assessment of compound effects on brain function and viability is paramount to reduce risks in clinical trials and improve the success rate in treating disease. The abysmal success rate for neurological compounds is 6% and underscores the need for more effective methods. Advancements in technology, such as hiPSC derived brain organoids, provide a promising tool that mimics critical features of the developing human brain (1). This allows for testing compounds in a more physiologically relevant model system. In this study, we present a high-throughput screening (HTS) approach to brain organoid testing. We conducted screenings on 15,000 brain organoids using 58 neurological therapeutics used in everyday clinical practice and known reference toxic substances. This was accomplished with minimal resource utilisation, using automated liquid handling and microscopy, accomplishing a high degree of similarity between the organoids.
Our results revealed comprehensive dose response profiles of all tested compounds across multiple effect modalities. The validity of our assay was confirmed by identification of the highly toxic effects of Rotenone, a potent inhibitor of mitochondrial complex I, and known to induce a Parkinsonian phenotype in rats. In addition to viability-related effects, we simultaneously measured spontaneous and evoked and potassium chloride-induced calcium activity, confirming dose-dependent neural activity effects of anticonvulsants and antidepressants. For example, Carbamazepine showed clear dose-dependent modulation of neural activity, where our system demonstrated higher sensitivity than animal studies.
Conclusion: The data presented in this study exemplifies the potential of human brain organoids as a powerful tool for profiling compound effects in a more physiologically relevant model system, including both adverse viability-related effects and functional neural activity changes. This approach holds promise for improving early discovery and preclinical assessment of new drug candidates, potentially reducing the risk of failures in downstream clinical trials. The use of human brain organoids for high-throughput screening offers a valuable bridge between in vitro and in vivo assessments, with implications for enhancing the safety and success rates of drug development.
Robin Pronk, PhD
Scalable mechanically active Organs-on-Chip for Drug Safety and Efficacy Screening
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Organs-on-chip (OoC) are emerging as powerful preclinical in vitro systems for drug screening and discovery. By replicating key physiological and pathological cues, OoC provide more human-relevant models, improving the prediction of clinical drug responses. Despite their potential, standardization and compatibility with automated systems for these OOC platforms have only recently begun to advance.Here, we present our strategy to develop, characterize, and qualify OoC for specific contexts of use, supporting their growing acceptance. Our platforms incorporate the uBeat® technology, which applies precise mechanical stimulation to 3D microtissues, enhancing physiological function or inducing disease-relevant features. Physiological and pathological models were leveraged specifically for drug screening purposes for both safety and efficacy applications, exemplified by the cardiac uHeart and uScar models, as well as barrier-like models of intestine, such as uGut.
uHeart is a human 3D cardiac model generated by subjecting induced pluripotent stem cells (iPSC) -derived cardiomyocytes to physiological mechanical stimulation (10% uniaxial strain, 1 Hz). It was preliminarily qualified for QT prolongation and pro-arrhythmia detection following ICH-S7B guidelines. Drug-induced alterations in electrophysiological signals, evaluated using 11 reference compounds from the Comprehensive in vitro Proarrhythmia Assay (CiPA) list, demonstrated that uHeart achieved 83.3% sensitivity and 100% specificity in predicting QT-prolongation.
The uScar model consists of human atrial cardiac fibroblasts (haCFs) cultured in 3D within the uBeat®-based OoC. The applied mechanical stimulation was sufficient to induce fibrotic traits, characterized by high fibroblast-to-myofibroblast transition and by an increased expression of ECM proteins such as collagen and fibronectin. Standard-of-care drugs like Pirfenidone and Tranilast were tested for efficacy qualification and confirmed to effectively prevent the onset of fibrotic characteristics.
uGut is a human gastrointestinal model encompassing functional epithelial and endothelial barriers of the intestine, by recapitulating the organ peristaltic-like movement (10% uniaxial strain, 0.2 Hz) promoting epithelial polarization, mucus production, and villi-like structures, which can be leveraged in absorption studies.
All the models described demonstrate the versatility of our platforms and the strategy we are implementing to impact therapeutics progression. Towards this aim, Biomimx is currently translating all the models discussed across its applications into a scalable, high-throughput, pharma-oriented setup compatible with automated laboratory systems. This approach will increase screening capacity, robust data generation, and reduced operator-dependent variability, thereby supporting standardization.
Stefano Piazza, MSc
Session: Digitalization of Molecular Discovery
The integration of automation, agentic AI, and advanced data analytics is redefining molecular and drug discovery. By establishing iterative feedback loops between computational models and automated laboratory systems, these technologies significantly accelerate the design–make–test–analyze (DMTA) cycle. This session will examine recent advances enabling closed-loop experimentation — from computational frameworks that interpret the complex languages of chemistry and biology, to predictive modeling and exploration of ultralarge chemical spaces, to the autonomous execution and interpretation of laboratory experiments. Together, these innovations are laying the foundation for fully digitalized, self-driving discovery platforms.
Session Chair: Yao Fehlis, PhD - KUNGFU.AI
Document-Aware Agentic AI for Closed-Loop DMTA in Drug Discovery
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Closed-loop integration of computational models, laboratory automation, and analytics is accelerating the Design-Make-Test-Analyze (DMTA) cycle in drug discovery. While agentic AI has shown promise for orchestration and experiment execution, unstructured documents-protocols, SOPs, and deviation records-remain a major bottleneck to reliable, automated workflows.
In this talk, I present a document-aware agentic AI approach for closed-loop DMTA, combining prior experience in lab orchestration and multi-agent systems with applied document understanding developed at KUNGFU.AI. The framework embeds document intelligence directly into the DMTA loop, enabling agents to interpret experimental intent, enforce execution constraints, and generate traceable analytical and reporting artifacts.
I demonstrate a LangGraph-based proof of concept in which agents coordinate experimental design, simulated automated execution, and analysis, using document-derived acceptance criteria and execution telemetry to drive iterative feedback. This work provides a practical blueprint for integrating agentic AI, document intelligence, and analytics to reduce cycle time and support scalable, self-driving discovery platforms
Yao Fehlis, PhD
Vibe Coding the Autonomous Discovery Lab: Building an End‑to‑End AI‑Driven Research Platform from Scratch
Open to view video. | Closed captions available
Open to view video. | Closed captions available
The digital transformation of drug discovery laboratories is often slowed by rigid software ecosystems, monolithic LIMS implementations, and automation architectures that struggle to adapt to rapidly evolving science. At the AITHYRA Institute, we are pursuing a different paradigm: "vibe coding" an autonomous discovery laboratory from the ground up, where software, hardware, and experimental science are co‑designed in real time through rapid, intuition‑driven, human‑in‑the‑loop development augmented by AI.
This presentation describes the practical construction of a real, production‑scale discovery lab spanning chemical screening, a gene and enzyme foundry, and proteomics screening. Rather than deploying pre‑packaged automation solutions, we developed custom, lightweight software layers that directly orchestrate laboratory hardware-including acoustic dispensing, liquid handlers, robotic arms, and modular platforms such as Opentrons-while integrating experiment design, scheduling, execution, and downstream data analysis into a unified system. AI agents are embedded throughout the design-build-test-learn (DBTL) loop to assist with experimental planning, parameter optimization, quality control, and real‑time interpretation of results, while keeping scientists actively in the loop.
This approach enabled rapid lab activation and early scientific output. Within the first two months of lab commissioning, we executed a fully automated 2,000‑compound chemical library screen, from experimental design through Echo integration, robotic cell culture execution, and data analysis. Within four months, we deployed a closed‑loop gene and enzyme foundry workflow, automating design, synthesis, screening, and analysis with minimal manual intervention. These milestones were achieved without a traditional LIMS, relying instead on flexible software abstractions that evolved alongside the science.
The talk will focus on architectural decisions, integration patterns, and cultural practices that made this possible, as well as lessons learned and failure modes encountered while "vibe coding" a real laboratory. We will discuss how tight feedback loops between scientists, automation engineers, and AI systems can dramatically reduce setup time, increase experimental velocity, and improve scalability without sacrificing scientific rigor. Attendees will leave with a practical blueprint for building autonomous, AI‑enabled research environments that align automation with how modern discovery teams actually work.
Wali Mustafa Malik, MA
Capability-Driven Automation for Self-Driving Laboratories: The LiSAH Framework
Open to view video. | Closed captions available
Open to view video. | Closed captions available
The vision of AI driven scientific discovery has renewed interest in Self Driving Laboratories, where intelligent systems design and execute experiments autonomously. While large language models can increasingly support experimental planning, most current SDL implementations remain tightly coupled to specific laboratory hardware. This limits interoperability and makes it difficult to adapt experiments when laboratory setups change.<br><br><br>This talk introduces LiSAH, a framework for building capability driven Self Driving Laboratories. Instead of linking experimental procedures to specific instruments, LiSAH separates what an experiment requires from how it is technically executed. Experimental workflows specify required capabilities, while laboratory systems provide these capabilities through standardized services and executable skills.
The framework combines semantic representations of samples, experimental sequences, and laboratory systems with digital twins of laboratory devices based on the Asset Administration Shell. These standardized interfaces allow heterogeneous instruments, including legacy devices, to be integrated into a shared automation environment.
By combining capability based automation, digital twins, and multi agent orchestration, LiSAH enables laboratory workflows that are more interoperable, adaptable, and scalable, providing a practical foundation for future Self Driving Laboratories.
Henrik Schu, M. Eng.
Advances in Drug Discovery
Session: AI-Driven Drug Discovery: From Prediction to Precision
Recent advances in computational modeling deliver predictive power that surpasses experimental measurement capabilities across biological scales, from protein structure and binding to single-cell perturbation effects, pathology image annotations, and genetic variant effects. Paired with high-throughput, multi-omic platforms, computational approaches that leverage large-scale datasets with diverse measurement modalities further bridge predictions and mechanistic insights across scales, enabling tissue- and patient-specific predictions of molecular interactions and therapeutic interventions. This session highlights computational advances with the potential to scale therapeutic discovery for precision medicine.
Session Chair: Xinyi Zhang, PhD - Aithyra
Prediction of protein subcellular localization in single cells
Open to view video. | Closed captions available
Open to view video. | Closed captions available
The subcellular localization of a protein is important for its function, and its mislocalization is linked to numerous diseases. Existing datasets capture limited pairs of proteins and cell lines, and existing protein localization prediction models either miss cell-type specificity or cannot generalize to unseen proteins. Here we present a method for Prediction of Unseen Proteins' Subcellular localization (PUPS). PUPS combines a protein language model and an image inpainting model to utilize both protein sequence and cellular images. We demonstrate that the protein sequence input enables generalization to unseen proteins, and the cellular image input captures single-cell variability, enabling cell-type-specific predictions. Experimental validation shows that PUPS can predict protein localization in newly performed experiments outside the Human Protein Atlas used for training. Collectively, PUPS provides a framework for predicting differential protein localization across cell lines and single cells within a cell line, including changes in protein localization driven by mutations.
Unbiased Phenotypes at Scale: AI/ML Transforms High-Content Imaging for Drug Discovery
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Classical phenotypic screens often rely on distal readouts (reporters, viability) that slow target deconvolution and extend DMTA timelines, particularly for intractable targets with unknown or evolving mechanisms. While image-based High-Content Screening (HCS) provides multiplexed, sub-cellular specificity, conventional assay development remains slow, subjective, and variable, depending on hand-picked endpoints and manual thresholds that can oversimplify complex biology and miss emergent phenotypes.
PhenoSpace, an AstraZeneca AI/ML platform, can automate single cell phenotyping to deliver robust, scalable HCI for drug discovery. PhenoSpace combines object detection, machine learning (ML)-powered phenotypic classification, and contrastive deep learning to identify the phenotype of each cell without manual feature selection or workflow tuning. By enabling unbiased detection of subtle and novel phenotypes across diverse biological systems, the platform expands biological signal capture while reducing assay-development cycles and increasing reproducibility. Here, we present a novel approach to utilising a high dimensional ML-powered phenotypic readout to support target-proximal Hit ID and target deconvolution at scale for intractable targets. We showcase how this AI-first approach is especially impactful for intractable targets-highlighting the robustness and efficacy of the PhenoSpace-based workflow in enabling faster hypothesis generation and more efficient progression from phenotypic signal to actionable chemistry.
Aditi Pradhan
Delivering Safer Drugs Through Innovation: AstraZeneca's Implementation of AI, Automation and High-Content Imaging
Open to view video. | Closed captions available
Open to view video. | Closed captions available
FDA data suggests that hepatotoxicity accounts for approximately 10-15% of all drug withdrawals from the market. Mitigating these safety risks is essential for evaluating the hepatotoxicity profile of potential molecules and elucidating their structure-activity relationships. Early assessment at the discovery stage holds the promise of rescuing programs without significant losses of time and resources.
The current approach requires testing a high volume of molecules in a high-throughput manner to identify compounds with the desired safety profile. To narrow down the number of compounds coming through the pipeline, our team has integrated an innovative approach that merges AI tools, statistical models, and advanced lab automation coupled with high-content imaging, enabling the identification of diverse parameters that serve as sensitive predictors of hepatotoxicity. This methodology includes leveraging legacy data together with the active generation of new datasets, enabling active learning to deliver safer medicines to our patients.
Our results demonstrate the power of this approach to accelerate decision-making, reduce safety liabilities, and inform molecular design, setting a new benchmark for early safety de-risking in drug discovery. This methodology showcases best practices in translational toxicology and demonstrates the role of automation and innovative approaches in elevating preclinical screening methodologies.
Anastasiia Gryniukova, PhD MBA
Session: Multiomics and Spatial Biology
Single cell and spatial omics technologies are revolutionising our ability to profile disease within the tissue context. This session aims to highlight how technological and computational advances in spatial biology are allowing us to scale experiments with impacts for understanding mechanisms of disease, allowing target and biomarker identification for patient stratification, and opening the doors to precision medicine across disease indications.
Session Chair: Nina Corsini, PhD - IMBA
Single cell analysis of clonal composition helps to decipher the logic of embryonic development
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Understanding how complex tissues emerge from multipotent progenitors remains a central challenge in developmental biology. While single-cell transcriptomics has provided detailed snapshots of cell states, it often lacks direct information about lineage relationships and clonal dynamics. In this talk, I will present a clonal atlas of mouse embryonic development generated using high-diversity lentiviral barcoding combined with single-cell RNA sequencing. This approach enables reconstruction of lineage relationships at scale, linking transcriptional states to their clonal origins. To interpret this complexity, we developed Clone2vec, a machine learning framework that embeds clones based on their cellular composition and transcriptional trajectories. This allows systematic comparison of clonal behaviors, identification of multipotent versus fate-biased progenitors, and quantification of clonal diversity across developmental contexts. By integrating lineage tracing with transcriptomic and computational analysis, we uncover how clonal structure shapes developmental outcomes and how progenitor cells distribute their potential across multiple lineages. These results provide a new perspective on the logic of embryonic development, revealing it as a probabilistic and spatially regulated process emerging from heterogeneous clonal strategies.
Igor Adameyko, PhD
Exploring human brain development and disease using human brain organoids
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Disorders of the human brain constitute a tremendous burden on human health worldwide. More than one in three people is affected by such neurological conditions, which makes them a leading cause of illness and disability.
Human neurological disease and brain development has been largely explored using animal models, in particular rodents. However, the human brain has undergone a tremendous evolutionary expansion enabled by an amplified and diversified repertoire of progenitor cell types. Human brains have a higher proportion of interneurons versus excitatory neurons compared to mice. In addition, human interneurons generate specific interneuron-interneuron networks (Loomba et al., 2022). Interneuron progenitors are also involved in the characteristic protracted development of the human brain, as streams of newly born interneurons migrate into the human cortex in the perinatal and postnatal period (Paredes et al., 2016). The increased size of the human brain also requires long range axonogenesis, a process often affected in neurodevelopmental disease (Martins-Costa et al., 2024).
To study human brain development, we employ cerebral organoids. Cerebral organoids are 3D cell culture models of the human brain grown from human ES or iPSCs that have been instrumental in uncovering mechanism of brain development and disease. Using a combination of single cell and spatial omics techniques and advanced brain organoid models allows us to uncover mechanisms of neuronal development and provides unprecedented insights into neurodevelopmental disease that provide inroads into treatment.
Nina Corsini, PhD
High-Throughput Functional Screening for Novel Multispecific Target Discovery in Inflammatory and Fibrotic Diseases
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Modulating two or more disease-associated pathways simultaneously via multispecific therapy represents a key strategic approach at Sanofi. Traditionally, paired target selection has relied on combining established monospecific compounds or selecting from a limited pool of literature-supported targets. To overcome these limitations, an innovative approach was developed for uncovering novel bispecific therapeutic hypotheses for complex diseases through unbiased combinatorial pairing of targets that modulate disease-modifying pathways. This comprehensive strategy integrates three essential pillars: (1) identification of novel targets, by combining patient data with deep phenotypic characterization in translational functional assays, (2) generation of unbiased combinatorial pairings of bispecific antibodies/NANOBODY® molecules, and (3) screening with the same translational functional assays.
The Rheumatoid Arthritis Fibroblast-Like Synoviocytes (RA-FLS)/CD4 T cell co-culture assay investigates the critical fibroblast-T cell interaction as a disease driver. Omics analysis of this co-culture assay demonstrated correlation with multiple inflammatory diseases, underscoring its translational potential across indications, particularly in fibrotic subtypes. This functional co-culture assay reflects upregulation of fibrotic markers by fibroblasts, upon interaction with activated T cells. The assay's robustness was validated through quantification of the fibrotic key markers, with clinically validated antibodies serving as benchmarks. The assay was successfully miniaturized and automated to a 384-well format to enable higher throughput screening.
The selection of novel multispecific targets follows a data-driven strategy that integrates comprehensive disease information with omics data from the functional RA-FLS/CD4 co-culture assay. This systematic approach aims to identify and validate promising bispecific therapeutic candidates for complex inflammatory and fibrotic diseases, thereby enriching Sanofi's target portfolio.
In conclusion, this innovative paradigm for bispecific target discovery leverages three pillars of Sanofi's strategy, the immunoscience foundational knowledge, disease data ecosystem, and drug discovery expertise. Through this interdisciplinary and collaborative approach, Sanofi Research is developing a diversified pipeline portfolio to address unmet medical needs.
Mozhgan Dehghan Harati, PhD
Session: Drug Repurposing and Systems Pharmacology: Accelerating Therapeutic Innovation
Explore how drug repurposing and systems pharmacology converge to revolutionize therapeutic innovation. This session delves into leveraging existing drugs and computational models to uncover novel treatments, optimize efficacy, and reduce development timelines thus presenting opportunities for a systems-level approach to tackling unmet medical needs with speed and precision.
Session Chair: Emre Guney, PhD - STALICLA
Visualizing and Navigating Complex Biomedical Data: Challenges and Opportunities in Network Medicine
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Virtually all processes in health and disease rely on the careful orchestration of a large number of diverse individual components ranging from molecules to cells and entire organs. Networks provide a powerful framework for describing and understanding these complex systems in a holistic fashion. They offer a unique combination of a highly intuitive, qualitative description, and a plethora of analytical, quantitative tools. In my presentation, I will review how molecular networks can be understood as maps for elucidating the relation between molecular-level perturbations and their phenotypic manifestations. I will then sketch out a number of challenges in the areas of network biology and network medicine, as well as recent efforts of my group to address them. These efforts range from methodological work on the visualization and interpretation of large biomedical data combining artificial intelligence with virtual reality technology, to translational efforts towards concrete clinical applications in rare diseases and drug repurposing.
Joerg Menche, Prof Dr
Repositioning drugs at scale for mechanism-based therapies in complex disorders
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Drug discovery is reaching a pivotal milestone as AI significantly accelerates the pace of virtual screening and de novo molecular design. Open source automation software and Large Language Models (LLMs) that can be run locally democratize access to reasoning and custom tool usage, enabling academic labs and biotech companies alike to advance assets rapidly without compromising confidentiality of the data. Despite these technological leaps, an urgent gap remains in developing mechanism-based therapies for complex disorders. While polygenic diseases exhibit high clinical heterogeneity, most current AI models rely on broad, organ-centric phenotypes that fail to capture the causal mechanisms driving patient-level variability. In this presentation, I will first highlight the initiatives of REPO4EU, the Euro-Global Platform for Mechanism-based Drug Repurposing, in establishing standardized workflows that leverage network pharmacology to identify synergistic, disease-modifying therapies. I will then introduce STALICLA's DEPI (Databased Endophenotyping Precision Identification) platform that integrates diverse multi-omic and clinical datasets. I will demonstrate how DEPI enables rational drug repositioning through AI and systems medicine by identifying mechanism-specific treatments tailored to distinct patient subgroups in neurodevelopmental and neuropsychiatric disorders.
Emre Guney, PhD
Drug repurposing screen identifies translational candidates for Multiple Sulfatase Deficiency
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Multiple Sulfatase Deficiency (MSD, MIM #272200) is an ultra-rare lysosomal storage disorder. It is caused by mutations in the SUMF1 gene, which encodes the formylglycine-generating enzyme (FGE). FGE's function is essential for the post-translational activation of all 17 cellular human sulfatases, most of which are localized in the lysosome. Their dysfunction leads to the accumulation of glycosaminoglycans and sulfatides, resulting in lysosomal storage and, consequently, cell pathology, progressive organ failure, and death. No curative treatment options are currently available for MSD. In a small-scale drug screen, the FDA-approved retinoids tazarotene and bexarotene were found to synergistically ameliorate MSD pathophysiology (Schlotawa et al., EMBO Molecular Medicine, 2023). However, both compounds are primarily used for topical treatment of skin diseases and their chronic oral administration could therefore result in adverse events. To find alternative drug candidates, we developed a high-throughput assay using patient-derived fibroblasts to identify therapeutic agent candidates that restore arylsulfatase A (ARSA) activity, one of the most important sulfatases affected in MSD. We screened a highly annotated small-molecule repurposing library containing 5,632 compounds, including (pre-)clinical candidates, marketed, and withdrawn drugs. We identified 80 compounds that significantly increased ARSA activity and confirmed their potency in 56 of these compounds through concentration-response titration experiments in the primary assay. Based on this, six top candidates were selected and ranked according to potency, toxicity, predicted target classes, and blood-brain barrier permeability. Using human induced pluripotent stem cell (iPSC)-derived blood-brain barrier (BBB) models, we will further predict drug candidates' potential to permeate into the central nervous system. Further, multi-omics studies will be carried out to identify underlying targets and mechanisms of action. Our findings highlight repurposable compounds with the potential to restore sulfatase activity and paving the way for the fast-track development of targeted treatments for this devastating disorder.
Annika Wittich
Session: RNA as an Emerging Target or Modality
This session will showcase cutting-edge research positioning RNA as both an innovative drug target and a versatile therapeutic modality. Presenters will discuss strategies for identifying key RNA molecules, including non-coding RNAs, that drive disease processes and how novel technologies are advancing RNA-targeting drug discovery. The session will highlight recent breakthroughs in RNA therapeutics, from small molecules to oligonucleotides and CRISPR-based approaches
Session Chair: Alessandro Bonetti, PhD - AstraZeneca
Turning junk into gems: long non-coding RNAs unlock new therapeutic horizons
Open to view video. | Closed captions available
Open to view video. | Closed captions available
In my talk I will give an overview of the non-coding field and the rationale that supports non-coding elements as novel pharmacological targets. I will discuss our current efforts for developing a technological pipeline aimed at identification and validation of non-coding pharmacological targets and present an example on how RNA therapeutics targeting non-coding elements can ameliorate disease phenotype in vitro.
Alessandro Bonetti, PhD
Small molecule-mediated modulation of RNA structure and function
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Targeting RNA with small molecules offers a transformative approach for modulating gene expression for disease targets that are currently deemed undruggable. Conventional screening strategies have traditionally prioritized thermodynamically stable, low-entropy RNA motifs that adopt well-defined conformations, yet such regions offer limited energetic opportunity for small molecules to induce functional effects by altering RNA structure. In contrast, many RNA molecules inherently sample dynamic ensembles of interconverting structures, providing a rich landscape where small molecules can in principle stabilize specific conformations or induce the formation of alternative conformational states, thereby modulating function. In this talk, I am gonna present some recent findings from our lab, showing how drugs can stabilize specific conformations within RNA structural ensembles, and how binding to RNA is not sufficient to elicit functional oucomes. I will show how transcriptome-wide chemical probing analysis can be leveraged to reveal the binding preferences of small molecules to RNA, further showing that only a subset of bound sites undergo structural rearrangements, ultimately affecting translation efficiency. Our findings provide mechanistic insights into drug-mediated modulation of RNA structure and function and establish a framework for systematically dissecting RNA-small molecule interactions beyond binding, illuminating new avenues for therapeutic intervention.
Danny Incarnato
Transcriptomics-Based Screening in Human Adipocytes for Innovative Metabolic Disease Therapeutics
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Purpose
Obesity and metabolic disorders remain major global health challenges, demanding innovative therapeutic strategies.
Adipocytes-central players in these diseases-are notoriously difficult to study due to their large size, high lipid content, and complex biology. Conventional functional assays are low throughput and capture only limited aspects of adipocyte physiology. To address this, we implemented an unbiased transcriptomics-based screening approach that delivers multidimensional data for informed decision making in drug discovery.
Experimental Approach
We established a robust, scalable assay using immortalized human white adipocytes (iHWA) differentiated in vitro and integrated it with ScreenSeq™, our proprietary 384‑well transcriptomics platform, as the primary readout. Handling viscous, lipid‑rich lysates required substantial protocol adaptations, leveraging ScreenSeq™'s proven versatility across ~150 cell models.
To illustrate how transcriptomics can resolve complex adipocyte phenotypes, we selected white‑to‑beige transdifferentiation as one therapeutic avenue and developed a corresponding beiging transcriptomic signature. This signature was derived by harmonizing published human and rodent datasets of in vitro as well as in vivo beiging models and applied through a dedicated scoring algorithm as a case study for disease‑relevant readouts.
Summary of Data
Pilot studies confirmed both assay robustness and our iHWA model's high transcriptomic similarity to in vivo human adipocytes. Using this platform, we screened 320 tool compounds to explore adipocyte biology and identify phenotype‑modifying agents.
Dimension reduction analysis showed tight clustering of compounds with shared targets or pathways, providing strong evidence that the assay captures coherent biological mechanisms rather than technical artifacts. Application of the beiging signature enabled systematic detection of compounds that partially activate thermogenic programs in iHWA. Among these were inhibitors of pathways previously shown to regulate brown adipocyte differentiation in vivo. These findings provide proof of concept for leveraging transcriptomics in complex cell models to uncover novel, disease‑relevant mechanisms.
Conclusion
ScreenSeq™ technology allows sizeable drug screens using transcriptomics as primary readout even with very difficult‑to‑handle cell types like adipocytes. Its data richness drives compound prioritization and mechanistic insights, providing new starting points for obesity and metabolic disease therapies. Its scalability-evidenced by ~4 million transcriptomes profiled to date-supports precise assessment of cell state and compound efficacy, mechanism of action, and safety at the earliest stages of discovery.
Future Directions
Building on this proof of concept, we have expanded screening to a larger tool compound set and 2,500 novel compounds, with data analysis underway. Additional transcriptomic signatures derived from obese patient material will support alternative hit identification. Further, we aim to broaden chemical space and perform deep validation of prioritized hits with the ultimate goal of accelerating therapeutic discovery for obesity and metabolic diseases.
Screening Applications and Diagnostics
Session: Perturbomics as a powerful tool to discover new targets, leads and biomarkers
Perturbation screens have become a cornerstone for uncovering novel biological insights with direct translational impact. By systematically disrupting genes, pathways, or cellular states, these approaches reveal previously hidden targets, validate mechanisms, and identify predictive biomarkers. Their versatility spans from early-stage discovery of druggable nodes to generating leads that accelerate therapeutic pipelines. This session will showcase cutting-edge applications of perturbation screening in both academic and industry settings, highlighting how these tools are reshaping the way we discover and develop new interventions.
Session Chair: Kilian Huber, PhD - University of Oxford
Chemical Biology Approaches for Drug Target Discovery and Validation
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Chemical biology has become an essential approach for connecting biological insight to therapeutic opportunity. This presentation highlights how chemical probes and emerging modalities such as targeted protein degradation (e.g., PROTACs) enable precise interrogation of protein function in native cellular contexts, supporting confident drug target discovery and validation while advancing fundamental biological understanding.
Using representative examples, the talk illustrates how complementary chemical perturbations-ranging from classical small-molecule inhibition to induced protein degradation-provide orthogonal views of protein function and mechanism. These approaches allow rapid testing of causality, help distinguish direct from indirect effects, and reveal protein roles that are not accessible through inhibition alone, including non-enzymatic and context-dependent functions. When combined with fit-for-purpose measurements of target engagement and pathway response, chemical biology tools strengthen mechanistic confidence and reduce uncertainty early in discovery.
Overall, integrating chemical probes with new therapeutic modalities creates a flexible and scalable framework for translating biological signals into validated targets, lead concepts, and biomarker hypotheses. Continued innovation in chemical perturbation strategies is expected to further accelerate drug discovery while uncovering new principles of cellular regulation.
Kilian Huber, PhD
CRISPR Screening for Target Discovery in Primary Human Macrophages and Dendritic Cells
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Macrophages are key players in maintaining tissue homeostasis and are directly involved in numerous diseases, including infections, inflammatory conditions, and cancer. Their central role in pathogenesis makes them promising targets for therapeutic intervention, however, their specific functions in disease mechanisms often remain poorly understood. CRISPR-based genetic screens have become a primary discovery engine in modern biology and are widely used to identify novel drug targets. Traditionally, these screens have been conducted in simplified cell models with basic readouts like cell viability, which often lack the resolution needed to dissect complex cellular processes. By combining genetic perturbations with single-cell transcriptomic profiling, high-content CRISPR screens enable rich, multidimensional insights into cellular phenotypes. Despite their potential, performing CRISPR screens in primary human macrophages has been technically challenging, limiting their accessibility for large-scale studies.
At Myllia, we have developed VISTA™, a pioneering single-cell CRISPR screening platform designed to investigate genetic regulators of biological processes such as activation states and identity in macrophages and dendritic cells. Using VISTA™, we conducted a pooled perturbation of approximately 100 genes and analyzed cellular responses under interferon gamma (IFNγ) stimulation in primary human monocyte-derived macrophages. Transcriptomic analysis revealed that depletion of both IFNγ receptor subunits IFNGR1 and IFNGR2 suppressed the expression of multiple interferon-stimulated genes. Conversely, depletion of SOCS1, a key negative regulator of JAK/STAT signaling, led to their upregulation. We also confirmed components of the FACT complex and m6A mRNA writers as negative regulators of IFNγ responses. The detection of both positive and negative regulators, along with transcription factors within the IFNγ pathway, validates the robustness and versatility of our assay. Notably, unbiased transcriptomic analysis uncovered MED12, a mediator complex component, as a novel regulator of a set of genes encoding cell surface proteins known to orchestrate immune responses. Moreover, MED12 also appeared to regulate genes involved in cytokine signaling associated with macrophage inflammatory states. These findings underscore the power of VISTA™ to unravel immunoregulatory pathways relevant to human disease in macrophages. We are currently broadening the scope of our screens and integrating complementary readouts to gain a more comprehensive understanding of macrophage biology, with the goal to enable translational insights into autoimmune disease and immunooncology.
Tilmann Bürckstümmer
Turning Primary Cells into Screening Ready Models: Affordable, High Throughput Functional Genomics Screening
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Functional genomics screens are essential for identifying novel drug targets, but their translational value depends on physiologically relevant in vitro disease models. Moving beyond immortalized cancer cell lines, primary cells better recapitulate disease biology but are difficult to perturb at scale because standard CRISPRn delivery by electroporation, while efficient, is costly, hard to scale, and requires large cell numbers which can be hard to acquire for primary cell types.
We evaluated lipid-based transfection as a scalable, a lower cost alternative for arrayed whole genome CRISPR Cas9 screens in primary human lung fibroblasts and aortic smooth muscle cells (AoSMCs). Using Cas9 mRNA and synthetic gRNAs, we screened eleven reagent and protocol variables to optimize editing. For AoSMCs we employed an I-optimal, high dimensional experimental design to assess multiple factors and levels in a single, compressed workflow, reducing experimental footprint from nine 384 well plates to one and enabling a STOP/GO decision in a single experiment. Optimized conditions produced >80% knockout efficiency in both primary cell types. These new transfection workflows are more amenable to high-throughput automation while also reducing transfection reagent costs by ~80% and reducing cells-per-well more than 16 times, making large-scale arrayed CRISPR screens in these primary human cells feasible. Through this approach, we have expanded arrayed whole‑genome CRISPR screening to physiologically relevant primary cell models, a crucial step toward identifying targets with higher translational potential.
Jenna M. Bradley
Session: Improving physiological relevance of models and predictive value of assays
Improving the physiological relevance of models is central to bridging the gap between experimental systems and human biology. More predictive assays not only reduce late-stage failures but also accelerate the translation of discoveries into effective therapies. Advances in complex cell systems, 3D cultures, organoids, and integrated assay platforms are driving this shift. This session will highlight innovations that enhance both the fidelity of models and the reliability of predictions across academic and industrial research.
Session Chair: Fernando Ramon Olayo, PhD - Servier
Cardioid-based drug discovery at scale
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Cardiovascular Disease (CVD) is the leading cause of death, accounting for 32% of global mortality. Yet, unlike other disease indications, breakthrough therapies for CVD have historically been difficult to identify. We propose that a significant factor in the persistent decline in CVD drug discovery is the absence of reliable in vitro models that can accurately predict human heart chamber pathologies suitable for high-throughput screening. The recent development of more complex 3D self-organizing cardiac chamber organoids (Cardioids) derived from induced pluripotent stem cells (iPSCs) promises to more faithfully recapitulate these pathologies. Here, I report on the progress HBB has made in developing its fully automated, integrated high-throughput drug discovery platform for generating, maintaining, and quality-controlling Cardioids. The Cardioids develop into chamber-like structures in a highly reproducible manner and allow for the co-development of the three main cell types of the human heart. By modulating protocols, we have significant control over cell type composition and function, enabling us to generate fit-for-purpose models to investigate mechanistically phenotype-specific questions. Using this approach, we have successfully established several Cardioid disease models and relevant readouts amenable to functional assessment in genetic and small-molecule HTS screening in 384-well format. We have employed it to model and screen disease (genetic cardiomyopathies) and to assess different aspects of cardiotoxicity by subjecting Cardioids to multiple genetic alterations and drugs. Overall, this platform and the Cardioids' versatility demonstrate practical applications and biological significance as a promising CVD drug-discovery tool, as well as relevance for toxicity screening, while highlighting the importance of high-throughput, quality-controlled, reproducible production and analysis.
Florian Fuchs
Bioreactor-Driven hiPSC-Derived 3D Models for Physiologically Relevant Preclinical Assays
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Human induced pluripotent stem cells (hiPSCs) have revolutionised biomedical research by enabling the generation of patient-specific, physiologically relevant cellular models. However, translating these advances into robust preclinical platforms requires overcoming challenges of reproducibility, scalability, and functional maturation. Bioreactor-based strategies offer a powerful solution by providing controlled microenvironments that support the development of complex three-dimensional (3D) tissue models.
In this talk, I will present recent advances in bioreactor technologies tailored for hiPSC-derived 3D constructs, including dynamic culture systems that enhance nutrient exchange, mechanical stimulation, and cellular organisation. I will highlight how these approaches improve physiological fidelity compared to conventional static cultures, while simultaneously enabling scalable production suitable for drug screening, toxicology, and disease modelling. Case studies will demonstrate applications in neural and hepatic tissues, underscoring the potential of bioreactor-driven platforms to bridge the gap between in vitro models and in vivo physiology.
Catarina Brito, PhD
CPSA: A novel assay technology for the assessment of drug-target engagement across the drug discovery process
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Methods to rapidly determine ligand target engagement at a throughput that is compatible with drug discovery and in a format that presents the target protein in its native form are valuable tools in drug discovery. To address this need Medicines Discovery Catapult has developed a proprietary, lysate-based chemical protein stability assay (CPSA) that measures compound-binding induced target stabilisation as a measure of target engagement. Our patented CPSA technology is based on the principle of chemical denaturant-induced protein unfolding, and the observation that drug-target binding stabilises this chemical unfolding. Unlike conventional thermal-shift or proteolysis-based methods, CPSA uses chemical denaturants to unfold proteins and monitors the remaining folded fraction with and without compound as a measure of target engagement. The use of cell lysate enables a simple high-throughput target engagement assay that presents the protein in its native cellular form and provides a route for targets that are intractable to large-scale protein purification. The technology features a streamlined, single-well "mix-and-read" format that requires no plate transfers, reducing variability and simplifying workflows. It is compatible with high-throughput screening in both 384- and 1536-well formats, making it applicable to early hit identification and lead compound profiling. As a label-free method requiring no target modification, it preserves the true pharmacology of ligands and allows the identification of compounds binding outside the active site, offering broader chemical-space exploration.
Comparative validation has shown high concordance with alternative cellular target engagement assays: we have generated dose response curves for a training set of p38α MAPK inhibitors and demonstrated assay robustness and reproducibility and pharmacology that aligns closely with alternative methods. We have further developed the method for use in single-point high throughput screening demonstrating excellent Z-prime values and run-to-run reproducibility.
Through partnering we have generated a package of data to show its effectiveness in HTS once miniaturized to 1536 well format and demonstrated the method is compatible with a range of target proteins. We have also independently validated the method in partner labs to strengthen our assertion of the robustness of the technology. As a result, this technology has now been licenced to multiple partners who have shared their data to confirm its compatibility with a variety of target classes.
In this presentation, we will outline the various formats and applications of CPSA, show benchmarking data and demonstrate use of the miniaturized assay for single-point screening and concentration-response analysis. Future developments of the assay will aim to validate the technology for use with different sample types, drug targets and protein detection endpoints including mass spectrometry proteomic profiling.
John P. Vincent
Session: Capturing heterogeneity: Biomarker signatures from complex data
Biological heterogeneity is both a challenge and an opportunity in understanding disease and treatment response. Advances in high-dimensional technologies generate complex datasets that capture this variation at molecular, cellular, and patient levels. Extracting meaningful biomarker signatures from such data is key to precision medicine, enabling better stratification and prediction of clinical outcomes. This session will explore strategies and case studies where heterogeneity is harnessed to define actionable biomarkers in academic and industry contexts.
Session Chair: Andre Rendeiro, PhD - CeMM Research Center for Molecular Medicine of the Austrian Academy of Sciences
Population-scale single-cell genomics for precision blood diagnostics
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Blood tests are foundational to medicine but routine blood counts provide only coarse summaries of a highly diverse cellular system. Single-cell genomics offer unprecedented resolution in blood diagnostics with potential for mechanistic understanding into underlying causes of blood abnormalities. In this talk, I will discuss our emerging efforts to build population-scale single-cell atlases of human blood and how they can transform blood diagnostics. Drawing on analyses of more than 10 million single-cell profiles from ~6,000 individuals across multiple biobanks, integrated with matched complete blood count data across time, I will demonstrate how atlases define subtypes of blood based on high-resolution cell states. Single-cell genomics reveals molecular outliers in blood through cell composition and gene regulatory networks, capturing signatures of clonal haematopoiesis. Together, this work illustrates how single-cell genomics can lay the groundwork for next-generation blood diagnostics and general strategies for defining subtypes and extremes across tissues beyond blood.
Olli Dufva, MD, PhD
m3DinAI Drug Quest: A Label-Free High-Content Imaging Pipeline for Automated Morphometric Profiling in 3D Cancer Models
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Relevant in vitro models are essential for advancing cancer drug discovery, especially for aggressive tumors with no actionable targets like triple negative breast cancer (TNBC). We have developed m3DinAI Drug Quest, a Python-based pipeline that leverages label-free high content imaging (HCI) to characterize the effects of chemotherapeutic agents on tumor spheroids. m3DinAI was used to define the phenotypic profiling of TNBC spheroids with different molecular subtypes treated with eight chemotherapeutic agents spanning four mechanistic classes (antimetabolites, anthracyclines, topoisomerase inhibitors, and taxanes). The magnitude and temporal persistence of drug-induced effects were quantified using morphological, textural, and radiomic features extracted from brightfield HCI. m3DinAI relies on a new concept, the morphological disruption concentration (MDC), defined as the lowest dose that induces detectable spheroid phenotypic changes. The results demonstrated an increased drug tolerance in TNBC spheroids compared to 2D monolayers. Unsupervised machine learning (ML) using UMAP (Uniform Manifold Approximation and Projection) revealed treatment-specific morphological signatures, facilitating the discrimination of drug classes and mechanisms of action based solely on label-free imaging data.
m3DinAI represents a first-in-class, label-free drug discovery pipeline that combines HCI, ML, and time-resolved activity mapping to define the pharmacological profiling of TNBC spheroids in less than one week, opening new avenues for precision oncology.<br><br>Future experiments include the screening of microbial natural products of diverse origin to identify compounds with antitumoral activity in 3D TNBC spheroids and map their mechanisms of action using UMAP.
Rosario Fernandez, PhD
Histological aging signatures enable tissue-specific disease prediction from blood
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Aging, the leading risk factor for numerous diseases, manifests through diverse structural and architectural changes in human tissues, providing an opportunity to quantify and interpret tissue-specific aging. To address this, we present a comprehensive assessment of tissue changes occurring during human aging, utilizing a vast array of whole slide histopathological images from the Genotype-Tissue Expression Project (GTEx), primarily reflecting non-diseased tissue samples. Using deep learning, we analyzed 25,712 images from 40 distinct tissue types across 983 individuals, quantifying nuanced morphological changes that tissues undergo with age. We developed 'tissue clocks' - predictors of biological age based on tissue images - which serve as a quantitative measure of tissue structural integrity and physiological fitness. These clocks were associated with established aging markers, including telomere attrition, subclinical pathologies, and comorbidities. In a systematic assessment of biological age rates across organs, we uncovered several associations between demographic, lifestyle, and medical history factors and tissue-specific acceleration or deceleration of biological age, highlighting potential modifiable risk factors that influenced the aging process at the tissue level. Finally, by combining paired histological images and gene expression data, we developed a strategy to predict tissue-specific age gaps from blood samples. This approach was validated in independent cohorts covering eight diseases, ranging from acute conditions like stroke to chronic diseases such as cystic fibrosis and Alzheimer's disease. It successfully recovered significant associations with disease-relevant organs and revealed patterns of systemic and tissue-specific aging that may reflect broader physiological changes in health and disease. This work offers a new perspective on the aging process by positioning tissue structure as an integrator of cellular and molecular changes that reflect the physiological state of organs in health and disease. It underscores the value of histopathological imaging as a tool for understanding human aging and provides a foundation for the monitoring of tissue-specific aging processes in age-associated diseases.
Andre F. Rendeiro, PhD
Session: New tools for improving antimicrobial preparedness - Sponsored by ELRIG
The threat of emergence of new pathogens or multi-resistant forms of known ones remains one of the most important health challenges. Recent pandemics have demonstrated the power of aligning all the living forces of innovation to produce an adequate response in a short time. In contrast, both public and private investment in this arena are challenging. This session will focus on the advancements to enhance diagnostics and accelerate discovery of new therapeutic solutions for infectious diseases.
Session Chair: Tilmann Buerckstuemmer - Myllia Biotechnology GmbH
Targeting de novo Purine Biosynthesis: A Novel Therapeutic Strategy for Tuberculosis
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Tuberculosis (TB) remains the leading cause of death from infectious diseases worldwide. Our research has identified JNJ-6640, a first-in-class small-molecule inhibitor targeting PurF, the first and committed step in Mycobacterium tuberculosis (Mtb) de novo purine biosynthesis. JNJ-6640 exhibits potent and selective bactericidal activity, effectively disrupting Mtb DNA replication. Crucially, we found that nucleobase levels in both mouse and human lung tissues are insufficient to rescue the inhibitory effects of JNJ-6640 on PurF. Using a long-acting injectable formulation we achieved in vivo efficacy, highlighting the therapeutic potential of PurF, and other purine biosynthesis inhibitors, as key components in future TB treatment regimens.
Dirk Lamprecht, PhD
From detection to prevention: Preserving Health in the era of antimicrobial resistance
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Antimicrobial resistance (AMR) is a growing global health threat that cannot be addressed through antibiotics alone. This presentation explores how next-generation bacterial vaccines can shift the paradigm from treatment to prevention, using C. difficile as a model. It highlights the importance of scalable diagnostics, the role of vaccines in reducing antibiotic use and resistance pressure, and Elaris' dual-mode vaccine strategy targeting both toxin neutralization and bacterial colonization.
Christian Taucher, PhD
Multi-dimensional Ribolysin optimization: From improving pharmacokinetics to synergy with standard of care antibiotics
Open to view video. | Closed captions available
Open to view video. | Closed captions available
Antimicrobial resistance (AMR) is a growing global health threat that cannot be addressed through antibiotics alone. This presentation explores how next-generation bacterial vaccines can shift the paradigm from treatment to prevention, using C. difficile as a model. It highlights the importance of scalable diagnostics, the role of vaccines in reducing antibiotic use and resistance pressure, and Elaris' dual-mode vaccine strategy targeting both toxin neutralization and bacterial colonization.
Zehra Visram, Dr.