Image 1 of 1
AMBER postdoctoral fellowship project (final call)
PRISM: Platform for Robotic Imaging and Synchrotron Microscopy
ID30B, a jointly operated ESRF-EMBL beamline, is a high-throughput, fully automated macromolecular crystallography (MX) beamline that will soon be equipped with new instrumentation to extend its capabilities toward X-ray microscopy applications. The upgraded S-TOMCAT and I-TOMCAT beamlines at the Swiss Light Source provide an ideal environment for high-resolution X-ray tomographic microscopy. These beamlines offer stable, high-flux imaging with submicron resolution, well suited to biological samples.
PRISM (Platform for Robotic Imaging and Synchrotron Microscopy) builds on this foundation by placing precision robotisation at the centre of new, fully automated imaging workflows. The core of the project is the adaptation of the existing Flex robotic sample-changing system (Papp et al., 2017)—originally developed for MX applications—for the handling of X-ray microscopy samples. This approach replaces manual intervention with a reliable, continuous, and reproducible pipeline.
Development of the new sample changer system will take place at EMBL Grenoble, followed by commissioning at the ESRF ID30B beamline. The successful candidate will join the Papp Team (https://www.embl.org/groups/papp/) and collaborate closely with the CAD, mechatronics and software engineers of the group. The commissioning and deployment of the system at ID30B will be supported by the McCarthy Team (https://www.embl.org/groups/mccarthy/) and ESRF colleagues. A finalised setup will then be installed at the PSI I-TOMCAT beamline, headed by Marco Stampanoni, in collaboration with ARINAX. Biological specimens can thus be automatically loaded, positioned, and exchanged, ensuring consistent imaging conditions and enabling long, unattended acquisition sequences essential for large-scale studies. This robust automation significantly increases throughput while reducing errors and variability.
Artificial intelligence within PRISM complements this robotic backbone and will evolve progressively in sophistication. Initially, AI will support tasks such as monitoring data quality and flagging potential anomalies during acquisition. As the project advances, more exploratory applications will be developed, including intelligent feature detection to trigger adaptive tuning of experimental parameters, as well as increasingly autonomous optimisation of scanning strategies. These forward-looking capabilities build upon the strong, reproducible framework established by robotisation, enhancing PRISM’s performance without compromising reliability.
By combining high-performance beamlines with robust automation and progressively evolving intelligent tools, PRISM will transform X-ray tomography into an accessible, high-throughput, and future-ready platform for the life sciences, enabling studies at a scale and consistency previously unattainable.
Location: EMBL, Grenoble, France
Organisation: EMBL, Grenoble
AMBER postdoctoral fellowship project (final call)
Development of AI-assisted sample optimisation workflows for cryo-EM
Recent developments in cryo-EM workflows have highlighted how machine learning approaches can be successfully applied at several stages of the pipeline. While these tools are relatively easy to implement at the image processing level (e.g., particle picking or reconstruction), their application to sample preparation and vitrification remains challenging.
This is largely due to the lack of automated and reproducible workflows, which are a necessary prerequisite for generating consistent experiments and enabling comprehensive data logging.
The EasyGrid instrument developed at EMBL is currently one of the most advanced and versatile automated sample preparation systems in the field. It enables fully automated preparation of cryo-EM samples, covering all steps from plasma treatment of sample supports, through sample dispensing and spreading, to jet vitrification and storage in cryo-EM boxes. The workflow is complemented by a quality control instrument (EasyGrid Control) capable of generating nanometer-scale thickness maps of the prepared specimens.
Together, the sample preparation and quality control instruments can be integrated into a closed-loop workflow, enabling optimization of cryo-EM sample preparation parameters without requiring immediate access to a cryo-EM microscope.
In parallel, a database is being developed to automatically store all metadata associated with each prepared sample, including chemical composition, preparation and vitrification parameters, pre-screening atlases, storage conditions, and cryo-EM screening data. Mining this comprehensive dataset will enable the training of specialized AI models, ultimately allowing automated and data-driven optimization of cryo-EM sample preparation.
The EasyGrid database will be integrated with EBI's structural biology resources, including EMPIAR, EMDB, PDB. This will cover alignment of the EasyGrid sample preparation metadata with EMPIAR's deposition schema (building on REMBI and related community standards), and cross-resource linking so that sample records can be related to related sequence, model, and structure entries. The AMBER postdoc will act as the primary point of contact between Grenoble and EBI, supported by the EMPIAR and EMDB teams.
The fellow will join the Papp Team & Djinovic Group at EMBL Grenoble. The selected candidate will work in close collaboration with the conceptors and super users of the EasyGrid system. EMBL EBI collaboration will be established with Hartley (EMPIAR) and Morris (EMDB) teams.
Location: EMBL, Grenoble, France
Organisation: EMBL, Grenoble
Links
AMBER call in EURAXESS main call (starting point for application)
Kristina Djinovic Carugo's profile in EMBL's Research portal