We have not been seeing cells as they really are, says Rick Horwitz, and it is time for that to change. Traditional laboratories focus on one or two aspects of cells, and not on the way the hundreds of different pieces of molecular machinery within the cell affect one another. To improve that perspective, Horwitz is leaving his cell biology lab at the University of Virginia to become the inaugural director of the Allen Institute for Cell Science in Seattle. The enterprise is being started today with a gift of $100 million from philanthropist Paul Allen, and will join the Allen Institute for Brain Science.
Horwitz says the institute will kick off a large project, the Allen Cell Observatory, creating animated models of cells combined with an enormous database of cell parts that allows users to zoom into the genes and proteins that make these parts work in health and disease.  “It will be rather like a Google Maps for cells,” he says. Horwitz spoke with Scientific American to describe the project, the new institute, and his hopes for the ways they can transform cell biology.
[An edited transcript of the interview follows.]
What do we not know about cells? Scientists have been studying them for a long time.
Cells are so very complicated, and we do not have any predictive models of their behavior and how that gets disturbed in disease. The cell comprises hundreds and hundreds of molecular machines, organelles like mitochondria and ribosomes. We do not know how genetic alterations in these machines change how they affect each other in time and space, and how that produces diseases like cancer. For example, a lot of cancer seems to be related to activities in the wrong place in the cell, at the wrong time.
What has been holding scientists back?
Traditionally, academic labs focus on small subsets of molecules, or small sets of cell activities. But to understand and predict cell behavior, you really need the overview. How many mitochondria are in a cell, for instance? How many ribosomes? When do they change in the cell cycle? How are these changes related to one another? We need to study them together because the cell is really an integrated system.
How will the Allen Cell Observatory allow this?
We want to create animated, visual models of the cell. We need to see how changes affect cells spatially and temporally.
But why animation?
The literature in cell biology has been presented as long stretches of text. It’s like reading the phone book. What we want is something closer to Google Maps. We want to zoom in and see protein-scale structures. We want to zoom out and see connections between organelles. That will be very powerful.
We can use this to build predictive models.  At the moment, I haven’t heard anyone make a bold prediction about cell behavior that turned out to be right. But once you see how a cell behaves normally, you can perturb one gene, or a set of genes, and see how that affects the life of that cell. That calls for sophisticated IT, a big database, new biosensors, and new cell assays to measure activity.
This is not happening elsewhere in cell biology?
Not really. A lot of research now is trying to be very translational, focused on helping people with different diseases. We are building tools and infrastructure that will be open-access and available to everyone. We are not trying to develop a drug. So we are not restricted by funds that are only supposed to go to developing treatments. I think what we are doing will lead to treatments, but we need to build the basic tools to understand cells first.
What cells will you use?
That’s a very big deal. We’ve decided to use human induced pluripotent stem cells. They can differentiate into any number of cell types, and they can go anywhere in the body. We think we can get pretty close to a real-life situation. But actually picking the right line of cells is going to be tricky.
Why so tricky?
If you pick the wrong cell line and it has an aberration, you’ll end up with a lot of very wrong stuff. We think that we will study three to five different lines and see if we keep getting consistent results. Then we’ll figure out what to do next.
When will you get started, and how long will it take before you get results?
We will move into our new building in Seattle in the fall of 2015, and between now and then we’ll be looking to hire about 75 people to work at the institute. Once we’re moved in and working, I think we’ll be able to report significant progress in three years.