Researchers at the Allen Institute for Cell Science, a Seattle research group founded by Microsoft co-founder Paul Allen, have created the first predictive 3D model of a live human cell. Using the model, scientists can digitally visualize and even manipulate cell behavior on a computer screen.
Called the Allen Integrated Cell, the model is the result of deep learning training with tens of thousands of high-quality cell images. It’s able to identify subcellular structures and project a 3D, multilayered image of a cell that shows how all its components interact simultaneously — something that has never been visualized in this way before.
“To me, it’s the closest thing I’ve ever seen in science to magic,” said Rick Horwitz, Executive Director of the Allen Institute for Cell Science. “Now you can see the inner workings of cells in action, in a movie and in three dimensions.”
To build the model, the researchers used NVIDIA TITAN Xp GPUs, an NVIDIA DGX Station AI supercomputer and the cuDNN-accelerated PyTorch deep learning framework, which studied images with protein tags that identified cellular structures such as the nucleus.
To see the 3D model and hear from the researchers themselves, watch the video below:
Beyond Slow, Costly Traditional Tools
Until now, scientists have used fluorescence microscopy — a microscope that uses fluorescence to study objects — to see structures within cells, which is an expensive, slow and damaging process. The microscopes can cost tens of thousands of dollars, all the proteins had to be specifically labeled, and the fluorescence could harm DNA within cells.
In the new model, researchers used a convolutional neural network to recreate the relationships between 3D transmitted light and fluorescence cell images in regards to major cellular structures. This teaches the model to build its own 3D structure from light images using a sample “buffet” of different proteins and organelles (the miniature organs in a cell body).
“This is a new way to see inside living human cells,” said Horwitz. “It’s like seeing the whole cell for the first time. In the future, this will impact drug discovery, disease research and how we frame basic studies involving human cells.”
The Allen Integrated Cell model can even predict the dynamics of mitotic (or cell dividing) events, such as the reorganization of the nuclear envelope and cell membrane. Such images were previously nearly impossible to predict, especially in 3D.
Capabilities like these give the model the potential to drastically change future studies regarding cells, whether in cancer research or entire industries such as pharmaceuticals and biotechnology.