Selecting particles from digital micrographs is an essential step in single-particle electron cryomicroscopy (cryo-EM). As manual selection of complete datasets—typically comprising thousands of particles—is a tedious and time-consuming process, numerous automatic particle pickers have been developed. However, non-ideal datasets pose a challenge to particle picking. Here we present the particle picking software crYOLO which is based on the deep-learning object detection system You Only Look Once (YOLO). After training the network with 200–2500 particles per dataset it automatically recognizes particles with high recall and precision while reaching a speed of up to five micrographs per second. Further, we present a general crYOLO network able to pick from previously unseen datasets, allowing for completely automated on-the-fly cryo-EM data preprocessing during data acquisition. crYOLO is available as a standalone program under http://sphire.mpg.de/ and is distributed as part of the image processing workflow in SPHIRE.

Selecting particles from digital micrographs is an essential step in single particle electron cryomicroscopy (cryo-EM). Since manual selection of complete datasets typically comprising many thousands of particles is a tedious and time-consuming process, many automatic particle pickers have been developed in the past few decades.However, non-ideal datasets pose a challenge to particle picking. Here, we present a novel automated particle picking software called crYOLO, which is based on the deep learning object detection system "You Only Look Once" (YOLO). After training the network with 500 -2,500 particles per dataset, it automatically recognizes particles with high recall and precision reaching a speed of up to five micrographs per second.Importantly, we demonstrate a powerful general network trained on more than 40 datasets to select previously unseen datasets, thus paving the way for completely automated "on-the-fly" cryo-EM data pre-processing during data acquisition. CrYOLO is available as a standalone program under http://sphire.mpg.de/ and will be part of the image processing workflow in SPHIRE.

The self-organizational properties of DNA have been used to realize synthetic hosts for protein encapsulation. However, current strategies of DNA–protein conjugation still limit true emulation of natural host–guest systems, whose formation relies on non-covalent bonds between geometrically matching interfaces. Here we report one of the largest DNA–protein complexes of semisynthetic origin held in place exclusively by spatially defined supramolecular interactions. Our approach is based on the decoration of the inner surface of a DNA origami hollow structure with multiple ligands converging to their corresponding binding sites on the protein surface with programmable symmetry and range-of-action. Our results demonstrate specific host–guest recognition in a 1:1 stoichiometry and selectivity for the guest whose size guarantees sufficient molecular diffusion preserving short intermolecular distances. DNA nanocontainers can be thus rationally designed to trap single guest molecules in their native form, mimicking natural strategies of molecular recognition and anticipating a new method of protein caging.

The elastic features of protein filaments are encoded in their component units and in the way they are connected, thus defining a biunivocal relationship between the monomer and the result of its self-assembly. Using DNA origami approaches, we constructed a reconfigurable module, composed of two quasi-independent domains and four possible interfaces, capable of facial and lateral growing through specific recognition patterns. Whereas the flexibility of the intra-domains region can be regulated by switchable DNA motifs, the inter-domain interfaces feature mutually and self-complementary shapes, whose pairwise association leads to filaments of programmable periodicity and variable persistence length. Thus, we show here that the assembly pathway leading to oligomeric chains can be finely tuned and fully controlled, enabling the emulation of protein-like filaments using a single construction principle. Our approach results in artificial materials with a large variety of ultrastructures and bending strengths comparable, or even superior, to their natural counterparts.