École Polytechnique Fédérale de Lausanne

6 related articles

researchsquare.com Oct 14, 2026
Sensor networks stop guessing only locally: nodes first find same-state neighbors, then recover the whole network
In heterogeneous sensing networks, agents previously could infer only their own state or rely on a central processor to aggregate measurements. A decentralized method called cluster-aware social learning now lets each agent use local observations and neighbor exchanges to identify connected groups sharing the same hidden state, maintain one belief per group, and map those beliefs back to individual agents, recovering the network-wide state configuration while exchanging only cluster-level information. Synthetic Gaussian experiments and real gas-sensor data both show beliefs concentrating on each cluster's true state once the cluster structure is frozen.
In heterogeneous sensing networks, agents previously could infer only their own state or rely on a central processor to aggregate measurements. A decentralized method called cluster-aware social learning now lets each agent use local observations and neighbor exchanges to identify connected groups sharing the same hidden state, maintain one belief per group, and map those beliefs back to individual agents, recovering the network-wide state configuration while exchanging only cluster-level information. Synthetic Gaussian experiments and real gas-sensor data both show beliefs concentrating on each cluster's true state once the cluster structure is frozen.
University of Salerno École Polytechnique Fédérale de Lausanne Malek Khammassi Vincenzo Matta Ali H. Sayed
Decentralized Social Learning Network-Wide State Inference Cluster-Aware Social Learning
bioRxiv Oct 13, 2026
Proteins no longer fold one way: after terminal remodeling, two or three mutations shift the conformational balance
Protein design has long targeted a single static structure, while natural metamorphic proteins occupy multiple folded states under the same conditions. Using RFdiffusion to partially remodel the termini of 93 native domains yielded 2,291 native-alternative conformation pairs; two sequence-design routes then made one sequence compatible with both backbones, and AlphaFold3 with BioEmu selected state-stabilizing mutations. The designs showed broadened and missing NMR crosspeaks, indicating microsecond-to-millisecond exchange in the metamorphic region; two or three mutations shifted the equilibrium toward either state, yielding crystal and NMR structures consistent with the design models.
Protein design has long targeted a single static structure, while natural metamorphic proteins occupy multiple folded states under the same conditions. Using RFdiffusion to partially remodel the termini of 93 native domains yielded 2,291 native-alternative conformation pairs; two sequence-design routes then made one sequence compatible with both backbones, and AlphaFold3 with BioEmu selected state-stabilizing mutations. The designs showed broadened and missing NMR crosspeaks, indicating microsecond-to-millisecond exchange in the metamorphic region; two or three mutations shifted the equilibrium toward either state, yielding crystal and NMR structures consistent with the design models.
SIB Swiss Institute of Bioinformatics University of Geneva HES-SO University of Applied Sciences and Arts Western Switzerland University College London École Polytechnique Fédérale de Lausanne Interuniversity Consortium for Magnetic Resonance Christian Schellhaas Roberta Pierattelli Alisa Khramushin Bruno E. Correia Maddalena Elia Alberto Borsatto Luciano A. Abriata Francesco Luigi Gervasio
RAS Techniques and Instruments Sep 25, 2026
BYOL features plus Protege active learning rank 100 MGCLS candidates, 99 showing diffuse radio characteristics and 55 confirmed as cluster-related emission
The work feeds self-supervised BYOL features extracted from source cutouts into the Astronomaly: Protege active-learning framework and evaluates the pipeline on high-resolution (about 7 arcsec), convolved (15 arcsec), and concatenated-feature MeerKAT Galaxy Cluster Legacy Survey (MGCLS) datasets, using tracers from a human-labelled catalogue as both guidance and benchmark; high-resolution features identify diffuse sources earlier than convolved ones, concatenated features perform best overall, and of the top 100 sources ranked by Protege 99 exhibit some form of diffuse radio emission with 55 confirmed as cluster-related, recovering 55 of 121 tracers from 62,587 sources with only 300 human labels.
The work feeds self-supervised BYOL features extracted from source cutouts into the Astronomaly: Protege active-learning framework and evaluates the pipeline on high-resolution (about 7 arcsec), convolved (15 arcsec), and concatenated-feature MeerKAT Galaxy Cluster Legacy Survey (MGCLS) datasets, using tracers from a human-labelled catalogue as both guidance and benchmark; high-resolution features identify diffuse sources earlier than convolved ones, concatenated features perform best overall, and of the top 100 sources ranked by Protege 99 exhibit some form of diffuse radio emission with 55 confirmed as cluster-related, recovering 55 of 121 tracers from 62,587 sources with only 300 human labels.
University of the Western Cape South African Radio Astronomy Observatory École Polytechnique Fédérale de Lausanne Verlon Etsebeth Emma Tolley Konstantinos Kolokythas Kenda Knowles Michelle Lochner
Galaxies: Formation, Evolution, Phenomena Radio Astronomy Observations and Technology Astronomy and Astrophysical Research
PLoS Biology Sep 21, 2026
AlloPool: a graph neural network framework that infers protein allostery from molecular dynamics simulations
The study presents AlloPool, a graph neural network framework that combines temporal attention with iterative edge pooling to learn minimal, time-evolving residue interaction networks from equilibrium and non-equilibrium molecular dynamics trajectories, reconstructing trajectories at sub-angstrom RMSD across Pin1, the GAIN mechanosensor domain, dopamine D2 and beta-1 adrenergic receptors, the SdrG adhesin, PDZ3, and the Engrailed homeodomain, and using those networks to map allosteric pathways, predict ligand pharmacology, mechanical loading states, and mutation effects.
The study presents AlloPool, a graph neural network framework that combines temporal attention with iterative edge pooling to learn minimal, time-evolving residue interaction networks from equilibrium and non-equilibrium molecular dynamics trajectories, reconstructing trajectories at sub-angstrom RMSD across Pin1, the GAIN mechanosensor domain, dopamine D2 and beta-1 adrenergic receptors, the SdrG adhesin, PDZ3, and the Engrailed homeodomain, and using those networks to map allosteric pathways, predict ligand pharmacology, mechanical loading states, and mutation effects.
Ludwig Cancer Research École Polytechnique Fédérale de Lausanne Aisima Chatzi Souleiman Lucas Guirardel Patrick Barth M. Marfoglia Miguel A. Pedraza-Joya
Graph Neural Network Protein Allostery Inference
bioRxiv Sep 18, 2026
LDDM: A Unified 3D Generative Model for Synthesizable Structure-Based Drug Design
The work introduces LDDM (Large Drug Discovery Model), a unified 3D generative framework supporting constrained and unconstrained docking, fragment linking and growing, and de novo design, together with a programmable design algorithm that produces synthetically accessible compounds satisfying fine-grained objectives; the authors experimentally validated designed or optimised ligands for five therapeutically relevant protein targets, achieving high success rates and identifying molecules with confirmed binding affinity while synthesizing only a small number of generated compounds, with the best designs structurally characterised by NMR spectroscopy and X-ray crystallography indicating high prediction accuracy.
The work introduces LDDM (Large Drug Discovery Model), a unified 3D generative framework supporting constrained and unconstrained docking, fragment linking and growing, and de novo design, together with a programmable design algorithm that produces synthetically accessible compounds satisfying fine-grained objectives; the authors experimentally validated designed or optimised ligands for five therapeutically relevant protein targets, achieving high success rates and identifying molecules with confirmed binding affinity while synthesizing only a small number of generated compounds, with the best designs structurally characterised by NMR spectroscopy and X-ray crystallography indicating high prediction accuracy.
AITHYRA GmbH- Research Institute for Biomedical Artificial Intelligence (Austria)University of California San Diego Institute of Molecular Physical Science École Polytechnique Fédérale de Lausanne Austrian Academy of Sciences Università degli Studi del Piemonte Orientale “Amedeo Avogadro”University of California, San Francisco Merck KGaA, Darmstadt (Germany)Laurin Kanis Fabio Testori Pierre A. Miranda Herrera David Robin Pavel Iosub Nicolas Thoma Kara Zielinski Philippe Schwaller Ana R. Lemos
Structure-Based Drug Design Generative Drug Design Fragment Linking and Growing
bioRxiv Sep 17, 2026
Microscope control with a natural language agent: an overview of MicroClaw
The authors present MicroClaw, an AI agent that advises on experiment- and system-specific parameters and approaches and collaboratively plans and executes diverse, complex, and reusable imaging workflows across microscope platforms, without requiring expert knowledge.
École Polytechnique Fédérale de Lausanne University of Applied Sciences Upper Austria University of Vienna University of California, San Francisco Vienna Biocenter Max Perutz Labs Anvita Manjunath Zach Marin Fabian Hauser Néstor M. Castillo Nico Stuurman Jonas Ries Amr Abouelezz Sarah Schweighofer
Smart Microscopy Automation AI Agent for Microscope Control

You’re all caught up

Sensor networks stop guessing only locally: nodes first find same-state neighbors, then recover the whole network

Proteins no longer fold one way: after terminal remodeling, two or three mutations shift the conformational balance

BYOL features plus Protege active learning rank 100 MGCLS candidates, 99 showing diffuse radio characteristics and 55 confirmed as cluster-related emission

AlloPool: a graph neural network framework that infers protein allostery from molecular dynamics simulations

LDDM: A Unified 3D Generative Model for Synthesizable Structure-Based Drug Design

Microscope control with a natural language agent: an overview of MicroClaw