Imagine being able to see inside a transparent human stem cell, like the “Visible Man and Woman” models in biology class. That’s what the Seattle-based Allen Institute for Cell Science lets you do with its brand-new data imaging platform, the Allen Cell Explorer.
The cells you see on the screen aren’t made-up animations: They’re based on an analysis of high-quality photomicrographs documenting more than 6,000 induced pluripotent stem cells, or IPS cells, derived from human skin cells.
The IPS cells underwent gene editing to attach fluorescent markers to 11 different types of structures that make up the cells’ machinery – and that’s not all. The institute then applied deep-learning computational methods to predict the complete structure of each cell, based on their glowing patterns.
“This is the first time researchers have used deep learning to try and understand the elusive question of how actual cells are organized,” Rick Horwitz, the institute’s executive director, said in a news release.
“The cartoons we rely on in textbooks, which are based on an artist’s interpretation of data from a relatively small number of cells, will eventually be replaced by data-driven models of this kind from very large numbers of cells,” he said.
Developing the database isn’t just an academic exercise: “We’re actually having brainstorming sessions to see what else can we do with this data,” Molly Maleckar, director of modeling at the institute, told GeekWire.
One example has to do with drug screening: Suppose a researcher has developed a drug to address a mitochondrial disorder, one of the trickiest types of illnesses to diagnose and treat. If the drug affects the cell’s mitochondria in a certain way, the Allen Cell Explorer could theoretically anticipate what happens to the rest of the cell’s machinery.
“The idea that we can build a cell, and think of the cell as a whole, and see how that cell changes as it grows, divides, differentiates or gets sick – to me, it’s really a new frontier in thinking about cells and the cellular basis of disease,” said Susanne Rafelski, the institute’s director of assay development.
Those sorts of applications could be waiting down the line. But for now, just understanding the interactions involving all the parts of the cell – for example, the mitochondria, microtubules and cell-to-cell junctions – is enough of a challenge.
“There’s this assumption outside the world of cell biology that people must know where everything is. … That’s actually not true,” Maleckar said.
Rafelski compares the Allen Cell Explorer’s visualizations to city maps that highlight the highways and the toll booths, the power plants and control centers inside stem cells.
One of the online portal’s tools, the Integrated Cell Model, predicts the entire cell’s structure on the basis of the maps (and shows how close the predicted map comes to the actual image data.)
There’s also a 3D Cell Viewer that offers thousands of 3-D stem cell visualizations, and a Cell Catalog that provides detailed information about the institute’s gene-edited stem cell lines.
Researchers can order the ingredients to replicate the cell lines, plus the instructions on how to use them, through the non-profit AddGene repository.
Like all of the Allen Institute’s databases, the Allen Cell Explorer is freely available to all – thanks to $100 million in contributions from Microsoft co-founder Paul Allen. And it’s far from finished.
Researchers aim to add visualizations of stem cells as they transform themselves into differentiated cells, such as heart muscle cells. Eventually, stem cell lines with mutations linked to cancer and other diseases could be added to the mix.
“It’s astounding that we get to do this, that we get to do science and then share science as beautifully as possible,” Rafelski said.
“I would absolutely agree,” Maleckar added. “Science altruism: It really works.”