(Image created by G. Volpe with the assistance of DALL·E 2)
What remains for humans to do in the age of AI? Date: 15 October 2026 Time: 13:05 Place: Uppsala, Sweden
Giovanni Volpe gave the invited talk “What remains for humans to do in the age of AI?” at the AI for Science Workshop, held in Uppsala, Sweden, on 15–16 October 2026.
The workshop was organized as a satellite meeting of the Swedish e-Science Academy 2026, bringing together researchers to discuss the growing role of artificial intelligence in scientific research and its potential to transform how scientific discoveries are made.
In his talk, Giovanni explored the question of what roles remain uniquely or essentially human as increasingly capable AI systems become integrated into scientific research. He discussed how artificial intelligence is reshaping scientific practices and how researchers can make productive use of these technologies while preserving the essential role of human creativity, critical thinking, and scientific intuition.
GAUDI leverages a hierarchical graph-convolutional variational autoencoder architecture, where an encoder progressively compresses the graph into a low-dimensional latent space, and a decoder reconstructs the graph from the latent embedding. (Image by M. Granfors and J. Pineda.)Unsupervised Graph Deep Learning Uncovers Protein Organization in Super-Resolution Microscopy
Mirja Granfors, Jesús Pineda, Blanca Zufiria Gerbolés, Joana B. Pereira, Carlo Manzo, and Giovanni Volpe Date: 7 October 2026 Time: 11:45 – 13:00 and 17:20 – 19:00 Place: Biblioteket, Arbetets museum, Norrköping
Modern optical techniques such as super-resolution microscopy generate rich, spatially resolved data that can be naturally represented as graphs. However, the structural variability of these graphs poses significant challenges for their analysis and classification, particularly when similar underlying processes can produce substantially different graph realizations.
To address this challenge, we introduce Graph Autoencoder Uncovering Descriptive Information (GAUDI), an unsupervised graph deep learning framework designed to capture both local details and global structure. GAUDI employs a hierarchical graph-convolutional variational autoencoder architecture with pooling and upsampling layers connected through skip connections, preserving essential connectivity information throughout the encoding–decoding process.
Despite substantial variability between graph realizations generated from identical or highly similar underlying parameters, GAUDI consistently maps them to nearby regions of a structured, continuous latent space. This enables the identification of invariant process-level features while separating them from stochastic variability.
We demonstrate this capability using super-resolution microscopy, where GAUDI characterizes protein assemblies and organizes structurally variable graphs according to their underlying state. The framework is further demonstrated on small-world network models, collective motion in the Vicsek model, and age-related changes in brain connectivity. These results highlight GAUDI as a broadly applicable approach for extracting meaningful information from complex graph-structured data across microscopy and other fields.
DeepTrack2 Logo. (Image by J. Pineda.)DeepTrack2: Physics-Based Microscopy Simulations for Deep Learning
Mirja Granfors, Alex Lech, Jiacheng Huang, Eduard Duta, Guillem Guigo i Corominas, Harshith Bachimanchi, Benjamin Midtvedt, Jesús Pineda, Carlo Manzo, and Giovanni Volpe Date: 7 October 2026 Time: 11:45 – 13:00 and 17:20 – 19:00 Place: Biblioteket, Arbetets museum, Norrköping
Deep learning has become an increasingly powerful tool for microscopy image analysis, but training robust models often requires large and diverse datasets that are difficult and time-consuming to acquire experimentally. Physics-based simulations provide a way to generate synthetic microscopy data while retaining control over the optical and physical properties of the imaging process.
DeepTrack2 is a flexible and scalable Python library designed to generate physics-based synthetic microscopy datasets for training deep learning models. It supports a wide range of imaging modalities, including brightfield, fluorescence, darkfield, and holography, and allows users to customize optical systems, incorporate optical aberrations and noise, simulate diverse objects and imaging conditions, and apply image augmentations.
Its modular architecture also enables simulations to be executed using a PyTorch backend, providing GPU acceleration and allowing backpropagation through the optical system. This makes it possible to optimize the optical setup and the deep learning model jointly. The DeepTrack2 repository further provides tutorials covering the complete workflow, from physics-based simulation to deep-learning-based analysis using its companion library, deeplay.
Together, these capabilities provide a flexible framework for developing and benchmarking deep learning methods for microscopy while reducing the need for large experimentally acquired training datasets.
Left: trapping of a healthy (biconcave) red blood cell. Right: trapping of an echinocyte-like cell with membrane protrusions. In the representation of the trapping (focused) beam, only rays originating from points close enough to the propagation axis are drawn. More peripheral rays are omitted. (Image from the manuscript.)How morphology affects optical trapping of red blood cells in health and disease
Emir Erdem, Gökberk Kabacaoğlu, Giovanni Volpe, Agnese Callegari, and Luca Biancofiore
Biomedical Optics Express 17, 5355-5372 (2026)
doi: 10.1364/BOE.609102
Optical tweezers are widely used to manipulate microscopic particles and biological cells through optical forces and torques, yet how cell shape controls trapping response has not been fully elucidated. Here, we develop a computational framework to investigate single-beam optical trapping of red blood cells with physiologically and pathologically relevant deformed morphologies. Representative cell shapes are generated parametrically and analyzed using static displacement and rotation tests, together with Brownian dynamics simulations. We show that all considered morphologies remain laterally and orientationally confined in a single-beam trap, while axial localization is weaker by pure optical means and is experimentally realized by the balance between the optical force and the buoyancy-corrected cell weight. Cell geometry strongly affects translational and rotational stiffness, preferred orientation, and Brownian confinement. In particular, shape-induced anisotropy produces distinct dynamical signatures, indicating that optical trapping measurements may be sensitive to morphological alterations beyond simple changes in cell size or volume. These results establish a quantitative connection between red-blood-cell morphology, trap stiffness, and dynamics, providing a computational framework for interpreting morphology-dependent optical trapping experiments.
Bridging scientific research and innovation to strengthen Europe’s competitiveness and societal impact. (Image generated by ChatGPT from A. Ciarlo.)Enhancing Europe’s competitiveness by empowering researchers as innovators
Amanda K. A. Silva Brun, Giorgio Presidente, Mangala Srinivas, Valentina Cauda, Alberto Fernandez-Tejada, Christine L. Mummery, Rami I. Aqeilan, Thomas Thum, Mehmet C. Onbasli, Robert Colebunders, Gianni Ciofani, Davide Iannuzzi, Marios Avraamides, João T. Barata, Romana Schirhagl, Ana Cecília A. Roque, Luana Dessbesell, Chloé Dupuis, Marco Paggi, Kerstin Kinkelin, Gerard J. J. Boink, Antonella Bongiovanni, Kevin Braeckmans, Kerem Pekkan, Lorenzo Moroni, Twan Lammers, Xiao Xiang Zhu, Adriele Prina-Mello, Carlo S. Casari, Max C. Lemme, Thomas Wolbers, Achillefs N. Kapanidis, Erdem Yörk, Jan S. Kirschke, Philipp-Leo Mengel, Giulio Cerullo, Volker Hessel, Víctor A. de la Peña O’Shea, Aitziber L. Cortajarena, Frank Kirchhoff, Harald H. H. W. Schmidt, Imre Berger, Anne Houdusse, Edgar R. Gomes, Frederic Amant, Tambet Teesalu, Joan Ballester, Juan José Vilatela, Raymond M. Schiffelers, Fabrizio d’Adda di Fagagna, Roman Jerala, Cristina Canal, Sophie Demolombe, Gaetano Gargiulo, Giovanni Volpe, Andrea C. Ferrari, Daniel Gros, Clemens Fuest, and Dario Polli
Science 393, 1302-1304 (2026)
doi: 10.1126/science.ady2388
Europe has a strong scientific research base, yet many promising discoveries still face barriers on the path from academia to technologies, products, and companies. In this Policy Forum, a broad group of European researchers, innovators, and economists examines how these barriers could be reduced to strengthen Europe’s capacity for breakthrough innovation and socioeconomic impact. Focusing particularly on Horizon Europe, the authors propose a more continuous pathway from fundamental research to commercialization, including closer integration between European Research Council Proof of Concept grants and European Innovation Council programs, stronger and more flexible support for early-stage high-risk innovation, and simpler funding procedures. They also highlight the importance of recognizing entrepreneurship within academic careers, improving intellectual-property management, and providing researchers with clearer access to mentoring, industrial partnerships, and spin-off creation. Together, these measures could help translate more of Europe’s scientific excellence into new technologies, businesses, and solutions for society, strengthening Europe’s long-term competitiveness.
Conceptual models of protein folding: funnel versus foldon. (Figure from the Authors of the manuscript.).
How do proteins fold? Date: 21 September 2026 Time: 13:15 Place: LINXS at The Loop, Lund
Giovanni Volpe will give the invited talk “How do proteins fold?” at the TR+ Conference 2026: From Static to Dynamics, organized by LINXS and MAX IV in Lund from 21 to 23 September 2026. The conference brings together researchers working on time-resolved methods to investigate the structural and functional dynamics of biomolecules.
How proteins fold remains a central unsolved problem in biology. While the idea of a folding code embedded in the amino acid sequence was introduced more than six decades ago, this code remains undefined. Although we now have powerful predictive tools to predict the final native structure of proteins, we still lack a predictive framework for how sequences dictate folding pathways.
Two main conceptual models dominate as explanations of folding mechanism: the funnel model, in which folding proceeds through many alternative routes on a rugged, hyperdimensional energy landscape; and the foldon model, which proposes a hierarchical sequence of discrete intermediates.
Recent advances on two fronts are now enabling folding studies in unprecedented ways. Powerful experimental approaches—in particular, single-molecule force spectroscopy and hydrogen–deuterium exchange assays—allow time-resolved tracking of the folding process at high resolution. At the same time, computational breakthroughs culminating in algorithms such as AlphaFold have revolutionized static structure prediction, opening opportunities to extend machine learning toward dynamics.
Together, these developments mark a turning point: for the first time, we are positioned to resolve how proteins fold, why they misfold, and how this knowledge can be harnessed for biology and medicine.
Reference: Bustamante, C., Kaiser, C. M., Lindahl, E., Sosa, R., & Volpe, G. (2026). “How proteins fold.” Nature Reviews Molecular Cell Biology. Advance online publication.
The Soft Matter Lab participates to the SPIE Optics+Photonics conference in San Diego, CA, USA, 23-27 August 2026, with the presentations listed below.
Agnese Callegari Geometrical optics for optical manipulation (Invited Paper)
26 August 2026 • 2:45 PM – 3:15 PM PDT | Conv. Ctr. Room 5B [link on SPIE]
Giovanni Volpe is a coauthor of the following presentations:
Mite Mijalkov In-silico cognition signals aging and cognitive decline
23 August 2026 • 3:15 PM – 3:30 PM PDT | Conv. Ctr. Room 2 [link on SPIE]
Iok-ui Tiunn Unsupervised analysis of 3D human bone reveals age-sensitive niches for CXCL12+ stromal cells
24 August 2026 • 2:15 PM – 2:30 PM PDT | Conv. Ctr. Room 2 [link on SPIE]
Alexandra Badea Graph-Based Embeddings of Laminar Cortical Columns Reveal Interpretable Microstructural Patterns Associated with Alzheimer’s Disease Risk (Invited Paper)
26 August 2026 • 10:30 AM – 11:00 AM PDT | Conv. Ctr. Room 2 [link on SPIE]
Massimiliano Passaretti Clinical progression and genetic pathways in body-first and brain-first Parkinson’s disease
26 August 2026 • 2:15 PM – 2:30 PM PDT | Conv. Ctr. Room 2 [link on SPIE]
Mite Mijalkov The locus coeruleus gap: a novel marker of neurodegenerative resilience
27 August 2026 • 8:45 AM – 9:00 AM PDT | Conv. Ctr. Room 2 [link on SPIE]
DeepTrack 2 Logo. (Image from DeepTrack 2 Project)Artificial Intelligence for Microscopy: From Pixels to Physical Insight
Giovanni Volpe
(Invited Paper) Date: 24 August 2026 Time: 9:45 AM – 10:10 AM PDT Place: Conv. Ctr. Room 15B
Advances in artificial intelligence are transforming microscopy from a tool of observation into a tool of discovery. Modern machine learning algorithms—especially deep neural networks—enable us to reconstruct, segment, and interpret microscopic images with unprecedented speed and precision. In this talk, I will discuss how AI can enhance every step of the microscopy pipeline, from data acquisition and noise reduction to virtual staining and quantitative analysis. Drawing on examples from my group’s work—including the DeepTrack framework for physics-aware deep learning and AI-driven virtual microscopy—I will show how we can train models to reveal hidden structures, track dynamic processes, and even uncover underlying physical laws directly from images. Beyond technical performance, I will also address the broader implications of this transformation: how AI reshapes the role of the experimenter, the reproducibility of quantitative microscopy, and the future of scientific understanding in an era of data-driven discovery.
EquiNET uses a graph neural network to infer entropy production from non-equilibrium molecular trajectories and uses this information to estimate equilibrium free-energy differences. (Image by S. K. Manikandan.)Free-Energy Differences from Nonequilibrium Fluctuations in High Dissipation
Sreekanth K Manikandan and Giovanni Volpe
arXiv: 2608.23394
Equilibrium free-energy differences can be measured from nonequilibrium work fluctuations using fluctuation theorems, such as the Jarzynski equality and the Crooks fluctuation theorem. However, these approaches become statistically inefficient in high-dissipation regimes because their convergence requires trajectories with negative entropy production, events that occur exponentially rarely. Here, we demonstrate that repeated measurements of trajectory fluctuations are sufficient to determine entropy production without relying on such rare events, or to determine a lower bound on it when only partial measurements are available. This information then yields exact estimates of free-energy differences, or rigorous bounds. We validate this approach, named EquiNET, with numerical simulations, including one of biomolecular folding and unfolding, showing that it recovers accurate free-energy estimates even in regimes where conventional approaches fail.
An energy- and data-efficient foundation model for high-content phenotypic screening
Benjamin Midtvedt, Jesús D. Pineda Castro, Henrik Klein Moberg, Jim Denholm, Mattias Goksör, Christos Matsoukas, Bjarki Johannesson, Giovanni Volpe Date: 24 August 2026 Time: 5:30 PM – 7:30 PM PDT Place: Conv. Ctr. Exhibit Hall B1
High-content screening (HCS) enables large-scale characterization of cellular responses to chemical and genetic perturbations, yet current deep learning approaches remain limited by label scarcity, batch effects, and high computational cost. We present an energy- and data-efficient foundation model for high-content phenotypic screening that unifies these challenges within a single architecture. Trained on the full JUMP Cell Painting dataset, the model combines active learning with contrastive supervision to focus representation learning on biologically informative variation while reducing sensitivity to technical artifacts. By supporting multimodal integration of molecular metadata alongside fluorescence morphology, the model learns generalizable cellular representations that transfer across experimental conditions and downstream phenotypic tasks.
