Technion breakthrough unlocks secrets of cells in 3D for first time
To solve this problem, the researchers turned to the field of deep learning, developing an artificial neural network capable of formulating its own solution
By DONNA RACHEL EDMUNDSSuper-resolved image of mitochondria rendered as a 2D histogram. Scale bar 5 μm.(photo credit: COURTESY OF NATURE METHODS AND AUTHORS)
A new microscope developed by Technion – Israel Institute of Technology promises to revolutionize the field of biology by offering scientists a 3D glimpse of cells in action at super-resolution.Microscopes as a rule produce two-dimensional images of cells allowing biologists a view of what is going on inside, but such images innately miss information as the world is three-dimensional. To date, scientists have worked around the problem by scanning a sample in layers to produce a computerized composite image of all the layers as a 3D object.But the process is limited as the sample must remain still for long periods of time to allow each layer to be scanned, meaning it can't be used to watch a cell in action in 3D.In addition, standard optical microscopy, in which light is passed through the sample and a lens to magnify the image, is limited by the diffraction limit of the lens.Enter DeepSTORM3D - a super-resolution 3D mapping system developed by Technion's researchers. Not only is it able to map images with a resolution ten times that achievable through standard optical microscopy, but it is able to map 3D images in moving systems."To get depth information from a 2D image we use wavefront shaping – an optical method that encodes the depth of each molecule in the image obtained on the camera,' explained Asst. Prof. Yoav Shechtman, who led the development of DeepSTORM3D. "The problem with this method is that if several molecules are close by, their images overlap on the camera and this drastically impairs spatial and temporal resolution, to the point that some samples cannot produce useful images at all."To solve this problem, the researchers turned to the field of deep learning, developing an artificial neural network capable of formulating its own solution. By teaching the system using a large number of virtual samples to the network, the neural network was able to train itself on how to produce super resolution 3D images from real-world microscopy data."The new technology has advanced us towards realizing one of the holy grails of biological research – mapping biological processes in living cells in super-resolution," Shechtman said. "It is important that the life sciences benefit from our instrumentation, and we maintain close relationships with biologists who explain their needs to us."Using the system, the researchers were able to demonstrate the system's ability to 3D map mitochondria, the cell's energy maker, and understand volumetric imaging of fluorescently labeled telomeres in live cells, which are, among other things, responsible for cell division.
Not only that, but the researchers found that the neural network was not only able to map the data, but it was also able to improve the instrumentation used."This is perhaps the most exciting direction to emerge from the current development," Shechtman said. "The neural network has provided us with the optimal physical design of the optical system. In other words, the computer not only analyzes the data but has shown us how to build the microscope. This concept can also be applied in non-microscopy-related fields, and we are working on it."
