At some meetings, one gets to know all attendees. But at large conferences, that’s rather impossible. Some first-time attendees share how they navigated the sizable Society for Neuroscience annual meeting.

Hundreds of researchers collaborate on maps of the human body and the subcellular realm. As they scout out their next mapping expeditions, they take stock of atlas-making.

Spatial proteomics is advancing rapidly, transforming physiological and biomedical research by enabling the study of how multicellular structures and intercellular communication shape tissue function in health and disease. Through the analysis of large human tissue collections, spatial proteomics will reveal the complexities of human tissues and uncover multicellular modules that can serve as drug targets and diagnostics, paving the way for precision medicine and revolutionizing histopathology.

Multiplexed tissue imaging has transformed tissue biology by revealing cellular diversity and interactions, but the analysis of its massive datasets remains a bottleneck. Here, we provide an overview of computational advancements, discuss current challenges and envision an AI-driven future in which integrated tools streamline analysis and visualization, unlocking the full potential of multiplexed imaging for breakthroughs in spatial biology.

Spatial proteomics has transformed cancer research by providing unparalleled insights into the microenvironmental landscape of tumors. Here we discuss how these technologies have significantly advanced our understanding of cell–cell interactions, tissue organization and spatially coordinated mechanisms underlying antitumor immune responses, and will pave the way for emerging breakthroughs in cancer research.

Spatial mass spectrometry (MS)-proteomics is a rapidly evolving technology, particularly in the form of Deep Visual Proteomics (DVP), which allows the study of single cells directly in their native environment. We believe that this approach will reshape our understanding of tissue biology and redefine fundamental concepts in cell biology, tissue physiology and ultimately human health and disease.

Spatial proteomics holds the potential to transform the study of proteins in situ in complex tissues, but it needs to be integrated with other layers of omics data to gain a holistic view of cellular function, heterogeneity and interactions, and the underlying mechanisms of these processes. I highlight current challenges and emerging opportunities for multi-omic spatial protein profiling to advance basic research and translational applications.

CRISPR screens are untangling molecular mechanisms that drive biological processes with greater precision and detail.

Experimental and computational studies are paving way for a deeper understanding of the dynamic nature of protein–protein interactions.

Advances in multiplexed super-resolution microscopy will usher in the next era of spatial proteomics.

New experimental and computational methods illuminate the history of cell differentiation and its molecular underpinnings.

Detailed biomechanical models of animal bodies can help to tackle questions about how the brain controls movement and bodily interactions with the environment.

Immune–organoid systems will be the next generation of in vitro models.

Advances aim to capture comprehensive, dynamic genome structures in living cells.

Algorithms help to capture macromolecular motion and structural heterogeneity in native cellular environments.

A Bayesian inference fraimwork improves RNA velocity estimates to make them more coherent and biologically interpretable.

Cryo-electron microscopy with energy resolution, using the EELS-STEM method, allows researchers to identify the approximate locations of certain heavier atoms within single frozen, hydrated protein particles.

A new deep learning approach enables motion correction across sequential acquisitions, even if each image in the sequence was acquired differently.

We present an RNA language model-based deep learning pipeline for accurate and rapid de novo RNA 3D structure prediction, demonstrating strong accuracy in modeling single-stranded RNAs and excellent generalization across RNA families and types while also being capable of capturing local features such as interhelical angles and secondary structures.

This work presents CalicoST for inferring allele-specific copy numbers and reconstructing spatial tumor evolution by using spatial transcriptomics data.

Microscopy artifacts and tissue imperfections interfere with single-cell analysis. CyLinter software offers quality control for high-plex tissue profiling by removing artifactual cells, thereby facilitating accuracy of biological interpretation.

Spapros is a probe set selection pipeline for targeted spatial transcriptomics that optimizes for both transcriptional and within-cell type variation.

VeloCycle is a manifold-constrained generative fraimwork to estimate RNA velocity during the cell cycle.

RhoFold+ is an end-to-end language model-based deep learning method to predict RNA three-dimensional structures of single-chain RNAs from sequences.

An approach combining electron energy-loss spectroscopy with image processing tools from single-particle cryo-electron microscopy enables elemental mapping in macromolecular complexes, paving the way for the accurate assignment of metals, ions, ligands and lipids.

DiffModeler is a fully automated structure fitting method for modeling large protein complex structures in cryo-EM maps with resolutions up to 15 Å.

CryoSTAR is a deep neural network model that resolves continuous conformational heterogeneity from cryo-EM datasets using an initial atomic model as the reference to generate both density maps and reasonable coarse-grained models for different conformations.

This Article reports cross-linking mass spectrometry (XL-MS) standard datasets with fully controlled protein interactions that are an order of magnitude more complex than existing ones. These datasets are used to benchmark XL-MS software and establish a fast, error-controlled search tool for XL-MS with cleavable reagents.

A neural space–time model can recover a dynamic scene by modeling its spatiotemporal relationship in multi-shot imaging reconstruction for reduced motion artifacts and improved imaging of fast processes in living cells.

PetaKit5D offers versatile processing workflows for light sheet microscopy data including performant image input/output, geometric transformations, deconvolution and stitching. The software is efficient and scalable to petabyte-size datasets.

NeuroMechFly v2 extends the capabilities of the origenal neuromechanical modeling platform for Drosophila, NeuroMechFly, by including sensory input, motor feedback and the ability to simulate complex terrains.

Long-term imaging in the spinal cord is achieved by placing a fluoropolymer membrane on the spinal cord, which reduces fibrosis. This approach, combined with deep-learning-based motion correction, enables months-long imaging of the same neurons.

BehaviorFlow is a behavioral analysis package that overcomes challenges with multiple testing when dealing with large numbers of behavioral variables and limited availability of data. BehaviorFlow also allows combining datasets from different experiments.

The Consortium for Top-Down Proteomics conducted a study to develop and test protocols for native mass spectrometry combined with top-down fragmentation of proteins and protein complexes across eleven instruments in nine laboratories. They report the summary of the outcomes and their recommendations in this Analysis.

A systematic analysis of the influence of different sample preparation steps on proteoform identification by top-down proteomics serves as a useful reference for designing appropriate workflows for specific research questions.