Understanding the molecular structure, dynamics, and reactivity requires bridging processes that occur across widely separated timescales. Conventional molecular dynamics simulations provide an atomistic resolution, but their femtosecond time steps limit access to the slow conformational changes and relaxation processes that govern chemical function. Here, we introduce a deep generative modeling framework that accelerates sampling of molecular dynamics by four orders of magnitude while retaining physical realism. Applied to small organic molecules and peptides, the approach enables quantitative characterization of equilibrium ensembles and dynamical relaxation processes that were previously only accessible by costly brute-force simulation. The method generalizes across chemical composition and system size, extrapolating to peptides larger than those used for training, and captures chemically meaningful transitions on extended timescales. By expanding the accessible range of molecular motions without sacrificing the atomistic detail, this approach opens opportunities for probing conformational landscapes, thermodynamics, and kinetics in systems central to chemistry and biophysics.

Optical tweezers are widely used in single-molecule biophysics, cell biomechanics and soft matter physics, but require a human operator, limiting throughput and repeatability. Here we present a smart optical tweezers platform, named SmartTrap, capable of performing complex experiments autonomously by integrating real-time three-dimensional particle tracking, custom electronics and a microfluidics system. Through a series of experiments, we demonstrate it can operate continuously, acquiring high-precision data over extended periods of time. By bridging the gap between manual experimentation and autonomous operation, SmartTrap establishes a robust and open-source framework for the next generation of optical tweezers research, capable of performing large-scale studies in single-molecule biophysics, cell mechanics and colloidal science with minimal experimental overhead and operator bias.

Activation of G protein-coupled receptors (GPCRs) by extracellular ligands is crucial for cellular communication and modulates numerous physiological processes. Despite sharing highly similar orthosteric binding sites, catecholamine GPCRs exhibit exquisite selectivity for their native agonists, even among nearly identical chemical messengers. However, the molecular basis and evolution of receptor selectivity remain poorly understood. To elucidate the structural mechanisms of GPCR selectivity, we focus on the prototypical human β2-adrenergic and D1 dopaminergic receptors, which are important drug targets and respond to the catecholamines adrenaline/noradrenaline and dopamine, respectively. Guided by structural and sequence data, we identify a small set of residues responsible for ligand selectivity. By exchanging residues at four positions in the β-adrenergic receptors and seven in the D1-like dopaminergic receptors, we swap the pharmacological profiles of the two subfamilies. Unexpectedly, the switch in selectivity not only involves residues interacting with the ligand, but is also controlled by regions outside the orthosteric binding site. Cryo-electron microscopy structures and computational models of the mutant receptors identify distinct molecular mechanisms contributing to selectivity in a concerted manner. Our findings provide insights into GPCR evolution and highlight strategies for protein engineering and drug design.

Single-molecule techniques are exceptionally well suited for analyzing the complex dynamic behavior of macromolecules involved in fundamental biological processes. Nevertheless, time and cost usually restrict current single-molecule methods to examining a limited number of different samples. At the same time, a broad sequence or chemical space often needs to be investigated to gain a thorough understanding of complex biological phenomena. To address this urgent need, we have developed multiplexed single-molecule characterization at the library scale (MUSCLE), a method that combines single-molecule fluorescence microscopy with next-generation sequencing to enable highly multiplexed observations of complex dynamics on millions of individual molecules spanning thousands of distinct sequences or barcoded entities. In this protocol, we outline the implementation of MUSCLE and present examples from our recent research, such as the sequence-dependent dynamics of Cas9-induced target DNA unwinding and rewinding. This example demonstrates that MUSCLE can be applied to study protein–nucleic acid interactions, going beyond nucleic-acid-only model systems. We detail the sample and library design, high-throughput single-molecule data acquisition, next-generation sequencing, spatial registration of single-molecule fluorescence and sequencing data and downstream data analysis. The ligation-based surface immobilization approach of MUSCLE ensures high clustering efficiency (>40%), increasing throughput and simplifying registration. In addition, MUSCLE includes a 3D-printed flow cell adapter that enables liquid exchange during single-molecule fluorescence microscopy. The complete procedure typically spans 3–4 days and yields a dataset that comprehensively characterizes the dynamic behavior of a library of constructs.

Biological function is fundamentally determined by nucleic acid and protein sequence. Beyond encoding genetic information, nucleic acids also display complex physicochemical parameters that shape structure, dynamics, and interactions. Understanding how sequence variation sculpts the energetic landscapes underlying these properties requires methods that capture both molecular diversity and dynamic behavior. Single-molecule techniques are ideally suited to this task, but conventional formats remain time and cost intensive. Recent breakthroughs have enabled highly multiplexed approaches for observing molecular dynamics across millions of individual molecules representing thousands of sequences or barcoded entities. Though still in development, these methods have begun to bridge sequence, structure, dynamics, and function at scale, opening new opportunities in drug discovery, molecular diagnostics, and functional genomics.