TL;DR
This paper introduces the "iTome," an automated ultramicrotome designed for high-throughput, high-resolution serial sectioning for Transmission Electron Microscopy (TEM). By employing optical interferometry for precise distance control and a novel pickup mechanism, the iTome significantly advances the ability to reconstruct large, complex biological volumes at the nanoscale.
What Problem Does This Paper Address?
Characterizing the intricate 3D structure of biological tissues at the nanoscale, crucial for understanding function, has historically relied on manual serial sectioning using ultramicrotomes. This process is incredibly time-consuming, labor-intensive, and prone to human error, limiting the achievable volume and resolution. As the demand for reconstructing larger and more complex biological structures grows, the manual approach becomes a significant bottleneck, pushing the limits of human endurance and throughput. This paper addresses the need for an automated, reliable, and efficient system to produce the vast number of high-quality ultrathin sections required for volumetric reconstruction.
Methods
The iTome is an automated ultramicrotome that integrates several key innovations:
- Optical Interferometry for Precision Control: Instead of relying solely on mechanical positioning, the iTome uses fiber-fed optical interferometers to precisely measure and servo-regulate the distance between the cutting knife and the sample block face. This offers superior accuracy and robustness against environmental perturbations compared to purely mechanical systems.
- Multi-Block Sectioning: The system can hold and section multiple sample blocks sequentially, allowing for efficient processing of different samples or the use of "pusher" blocks.
- Novel Section Pickup Mechanism: Sections are picked up while still anchored to the knife edge. A "pusher" section of empty resin is cut after the desired tissue sections for a grid. This pusher section displaces the tissue sections away from the knife edge, ensuring they are precisely positioned within the open slot of the TEM grid, rather than on the metal rim. This method is also adapted for large tissue sections where the pusher resin is integrated into the block.
- Automated Grid Handling: A sophisticated robotic system manages TEM grids, including picking them from a cassette, presenting them for section pickup, and returning them for storage. This system utilizes machine vision for precise alignment and handling.
- Integrated Machine Vision: Cameras are used throughout the process for monitoring section quality, guiding grid handling, and establishing precise spatial relationships between the grid and the knife edge.
Datasets
The paper does not discuss datasets in the traditional machine learning sense. Instead, it presents results from sectioning biological samples:
- Drosophila melanogaster first-instar larva: A complete larva was sectioned into 4,866 sections.
- Mouse cortex: A volume measuring 400 µm x 2000 µm x 30 µm was sectioned.
- Adult Drosophila brain and ventral nerve column (VNC): A tissue volume measuring 400 µm x 800 µm x 344 µm was sectioned.
These represent new benchmarks for the volume and complexity of biological material sectioned for TEM.
Key Insights and Ideas
- Precision through Optics: The reliance on optical interferometry for distance measurement is a significant departure from traditional microtomes, offering a path to higher accuracy and reliability.
- The "Pusher" Section: The ingenious use of an empty resin section to precisely position tissue sections within the grid slot is a clever solution to a common problem in automated section pickup.
- TEM's Throughput Advantage: The paper highlights that TEM imaging, with its parallel nature, offers a significant throughput advantage over scanned imaging methods (like SEM) for reconstructing large volumes, making automated TEM sectioning highly valuable.
- Reproducibility is Key: The demonstrated high success rates (99.74% and 99.95%) and visually consistent section placement across multiple grids underscore the system's reliability.
Key Results
The iTome demonstrated impressive performance in sectioning various biological samples:
- Drosophila larva: 4,866 sections, 34 nm thick each, with a 99.74% pickup success rate.
- Mouse cortex: 750 sections, 40 nm thick each.
- Drosophila brain and VNC: 9,300 sections, 37 nm thick each, with a 99.95% pickup success rate.
The system achieved highly reproducible section placement within grid slots, with a low incidence of folds or wrinkles. The authors also noted a reduction in errors through iterative improvements and feedback from visual inspection of the sections.
Limitations and Future Work
- Error Events: While success rates are high, the paper details specific error events, including human error (typographical), mechanical failures (pump breakdown, servo actuator failure), and issues with section anchoring or cutting. These highlight areas for further system robustness.
- Environmental Sensitivity: Although optical interferometers reduce sensitivity to environmental perturbations, the system is still shielded by an acrylic enclosure to mitigate ambient temperature and humidity fluctuations.
- Future Work: The paper implicitly suggests future work in further refining the automation to minimize remaining error sources and potentially increasing sectioning speed. The ability to section large volumes for TEM opens doors for more comprehensive 3D reconstructions of biological systems.
Summary updated