“What we had to do was try to find a way in which we could merge high-end computation on the one hand, with the common things that biologists do in the wet lab, and with some new state-of-the-art imaging tools that biologists are starting to incorporate in their research,” says Paul Matsudaira, director of the Whitehead-MIT BioImaging Center. “The challenge was that we had to do it in a nine-month span.”

Now two years old, the Center has become the cornerstone of MIT’s imaging and computing capacity.

“It unites the power of leading-edge microscopy and advanced computational systems to study the structure, dynamics, and function of molecules in challenging problems faced by biology, medicine, and bioengineering,” says Matsudaira.

MIT is a premier teaching and research institution, made up of departments and research centers, explains Matsudaira. Founded almost 25 years ago, the Whitehead Institute is independent of MIT, but is affiliated with it; all 15 Whitehead faculty are MIT faculty whose research is funded through the Whitehead Institute. Within the Whitehead Institute are a number of research initiatives, including the 17,490-sf BioImaging Center, which is administered by MIT. Each researcher is working on four or five projects at a time, developing new types of microscopes and new ways of storing, organizing, and querying the image databases.

The Whitehead Institute occupies two floors of a nine-story building constructed in the 1970s on the Polaroid headquarters campus. It was most recently occupied by Akami, an internet content distribution company.

“There were about five floors that MIT was going to build out; two for the Institute for Soldier Nanotechnologies—an Army-funded research institute at MIT—and two for bioengineering and the BioImaging Center,” explains Matsudaira. The top five floors are occupied by a major private biotech company, and the first contains the server room for MIT’s Computational and Systems Biology Initiative.

Only three of the 10 MIT faculty reside at the BioImaging Center. The rest use the imaging equipment, then access the massive database of information remotely from their own offices and labs throughout campus.

“It is both a virtual site and a real site,” Matsudaira says of the Center.

Not the Biology You Learned in High School

The human genome project changed the study of biology, explains Matsudaira. Biologists no longer simply study cells, they break them down into their component parts and reassemble them in a discipline called systems biology. Matsudaira likens it to the work a systems engineer does on an airplane.

“A Boeing 747 consists of roughly six million parts, half of which are rivets,” he says. “An aerosystems engineer at Boeing makes sure that all of those parts fly in formation.”

Systems biologists don’t have the luxury of a blueprint to build the cell, but they do know that roughly 24,000 genes encode more than 24,000 different proteins in cell. Those proteins comprise the machinery that make up the cell, making it do what it does.

“Systems biologists are trying to carefully tease things apart and figure out what each of these proteins is doing, not only individually but as part of a system,” he says. “A biologist goes in and diddles with the cell. We can manipulate any gene at will. We have total control over the cell.”

For example, systems biologists are investigating which genes are turned on in a cancer cell, or what changes a cell into a tumor. In doing so, they uncovered why certain cancer treatments are effective in only half the cases: They discovered that small-cell carcinomas, all of which look identical to a pathologist, actually are present in two forms, only one of which responds to the drug.

Scientists manipulate cells at the wet benches using molecular genetics, chemical genetics, and cell engineering; then measure the result using array technology, bio-devices, and tools such as imaging equipment. The next step is to collect the data and organize it using sophisticated computational methods, such as bioinformatics, data semantics, and databases, which spot trends in the data.

That is where genomics and proteomics come into play. Proteomics is the study of the protein machinery of the cell, essentially assembling a parts list of the working parts of the cell. Genomics is assembling a parts list of genes, which is where the information for the cell is stored. Using those disciplines, scientists then extract or develop knowledge based on the new data, and then go back to the wet benches to test the knowledge through experimentation using network, mechanical, and biochemical models.

“You have this repeating loop of manipulate, measure, mine, and model,” he says. “The ideal systems biologist is both an experimental biologist and a computational biologist. We are trying to create a new kind of student, a new kind of researcher at MIT.”

Tools of the Trade

One of the most important tools for biologists remains the microscope, first invented about 400 years ago.

“Every few decades or so it gets reinvented and becomes a much more powerful tool,” says Matsudaira. “At MIT, we have set up the BioImaging Center around imaging platforms or microscope modes in which we can push the envelope of state-of-the-art, whether it is in high resolution nuclear magnetic resonance (NMR), looking at the structures of tissues; or at the other end with the atomic force microscope (AFM) pushing to be even faster and faster so we can look at dynamic events in cells or in isolated solutions.”

Accommodating this equipment in a renovated 30-year-old building was a challenge. Most of the light microscopes use lasers, because lasers produce a very defined wavelength of light. But they also produce a lot of heat. And noise is a big factor with all imaging equipment, because noise causes vibration. To quiet the space, the floors are padded, and the corridors are carpeted. The equipment itself is installed on air-cushioned tables.

The imaging suites are clustered at one end of the floor to minimize foot traffic, and minimize noise and vibration.

“Microscopy is not exactly something that you do cheek and jowl with somebody else,” he says. “When you’re taking pictures, you want to be isolated from everyone else.”

Student carrels are at the end of the building opposite the imaging suites; in between lie the faculty offices and the wet labs.

The wet labs were designed as generic open lab space to make them adaptable to ever-evolving science. Some of the less vibration-sensitive microscopes are located right in the wet labs.

“It was important for us to have the students and the scientists who are manipulating and staining the cells, and putting those cells in the microscopes, be adjacent to the microscopes, which are self-enclosed and automated,” says Matsudaira.

“We installed about $2 million worth of microscopes and software,” he says. “It allows us to looks at cells in a fairly industrial way.”

But the imaging equipment is only half the story. All the data collected by the scientists is uploaded directly to a bank of servers in a 1,000-sf room on the first-floor, part of MIT’s Computational and Systems Biology Initiative. The servers are used not only by the BioImaging Center but also by many other departments and institutes at MIT, connected by a high-bandwidth network.

“Some faculty are collecting on the order of a terabyte of image data a day,” he says. “We bypass the PC, obviously, and data goes straight from the microscope down to servers and storage on the first floor in a purpose-built computer facility with 50 terabytes of storage.”

The server room is equipped with a raised floor, and an HVAC system handling 80 tons/hour of air conditioning. The servers have a 30-minute UPS backup to allow for an orderly shutdown in case of a power outage. The power going into the room is about 150 kilovolt amperes.

Front-end processors, with several terabytes of storage on their hard drives, handle very large image data sets and can retrieve data from storage efficiently.

“I think the ability to support all of the high-end imaging was possible because we were able to build out high-end computation,” says Matsudaira. “Usually, computation either comes as an afterthought, and then you scramble to create space for it, or the computing budget was never part of the construction budget at all. We insisted right from the beginning that we couldn’t do one without the other.”

Fostering Collaboration

MIT has worked to foster a culture of collaboration and interaction among scientists and students. For example, faculty offices are concentrated along a single corridor.

“It is important for faculty to actually see each other and talk to each other on a daily basis,” he said. “Faculty offices usually are in their labs, hidden behind doors.”

Students are likewise put together in adjoining carrels.

“Our students are such that they are the ones initiating collaborations,” says Matsudaira. “They are talking with friends who say, ‘Why don’t you talk to so-and-so; she’s doing this experiment; maybe she can help you.’

“From the very beginning, that was the key to the success of how our science progresses,” he continues. “The Center is interdisciplinary by definition, created to address research problems across departments and across schools. The BioImaging Center and the Computational and Systems Biology Initiative involve faculty from biology, bioengineering, mechanical engineering, electrical engineering, computer science, chemical engineering, chemistry, and physics. There are not that many departments that aren’t a part of this.”

Matsudaira says that the Center has been a success in its first two years; his only regret is that it isn’t bigger.

“I could use another floor and we would fill it up with students and equipment in no time at all,” he says.

By Lisa Wesel

We welcome your Questions and Comments

Copyright 2008 Tradeline Inc.
All Rights Reserved
ISSN: 1096-4894