Mailing List

Contact Us

Staff

By Tim Stoddard

	Charles DeLisi, Arthur G. B. Metcalf Professor of Science and Engineering and senior associate provost for bioscience, has been instrumental in putting BU at the forefront of genomic research. Photo by Vernon Doucette

Now that the human genome, that vast trove of genetic information encoded in our DNA, is only a mouse-click away, researchers face the daunting task of figuring out how tens of thousands of genes choreograph the life and death of cells. Under the leadership of Charles DeLisi, Arthur G. B. Metcalf Professor of Science and Engineering and senior associate provost for bioscience, faculty in the College of Arts and Sciences, the School of Medicine, and the College of Engineering are developing new ways of understanding the complex machinery inside living cells.

“ We now have the full text for many genomes, but the text for the human genome has not yet been completely parsed into words,” says DeLisi, who is considered the father of the Human Genome Project. “When it is — and we can reasonably expect to be 90 percent complete in the next three to five years — we’ll have a parts list. We’re also developing the tools to understand how those parts interrelate. It’s not like putting together a jigsaw puzzle or a 747. It’s harder than that, because connections between the parts are not fixed — they change in response to the environment. The study and design of such complex adaptive systems is a major research area of modern engineering.”

Following is a look at some of the newest programs in genomic research at the University and specific work being done by Boston University researchers.

Program in bioinformatics

As genomics has changed the biological landscape, Boston University has launched a graduate program to train rising scientists for leadership in the burgeoning field of bioinformatics. Founded in 1999 with a grant from the National Science Foundation’s Integrative Graduate Education and Research Traineeship Program, this University-wide program provides a unique interdisciplinary perspective on the science, engineering, medicine, and ethics of 21st-century cell biology. “To most of us, bioinformatics is a way to understand the cell as a system,” says DeLisi, the program’s director. “It’s the intersection between computer science and genetics. It’s a way to identify the functions of thousands of genes much more rapidly than we could have in the past.”

The program comprises 50 faculty members, and its thrust is integrating biology with information sciences and engineering. But there is also a core course exploring the legal and ethical issues emerging in the field of biotechnology. Advanced mathematics and computation figure prominently in the curriculum, which covers such current topics as genomic and proteomic biotechnology, microarray engineering, and structural biology. “One of the most beautiful aspects of this program,” says Simon Kasif, an ENG professor of biomedical engineering, “is that it’s producing a new breed of students capable of carrying out complex biological experiments and analyzing them with sophisticated computational ideas. In the past there were biologists who did pipetting, and statisticians who did the analysis.”

The sequencing of the human genome was the “first event when all biologists understood that they could not survive without computers,” says Sandor Vajda, an ENG professor of biomedical engineering and one of the core bioinformatics faculty. “It is simply not possible to access all of this genomic information in any other way. You need excellent databases and very good interfaces for searching. As the computing speed goes up, we will be able to answer much more complex questions than we have in the past, going beyond simply searching the database.” The computing resources at Boston University — such as the Biowulf Linux Cluster, an IBM 128-node multiprocessor — facilitate the massive number of calculations that arise when manipulating gene sequences and protein structures. For more information about the Bioinformatics Graduate Program, see bioinfo.bu.edu.

Center for Advanced Genomic Technology (CAGT)

The nascent Center for Advanced Genomic Technology (whose acronym, CAGT, is a play on the four bases of DNA, cytosine, adenine, guanine, and thymine) comprises the research core of the bioinformatics program. Directed by DeLisi, it was founded in 2002 as a direct offshoot of the molecular engineering research laboratory (MERL), which DeLisi established in 1990 when he was dean of ENG. CAGT is divided into four laboratories exploring different aspects of genomics and proteomics, the study of the structure and function of the proteins that genes produce.

Biomolecular systems laboratory

The biomolecular systems laboratory (BSL), directed by DeLisi, is developing computational and experimental methods for monitoring changes in proteins and genes inside a cell as it responds to its environment. To this end, DeLisi’s group has been improving DNA microarrays, a technology now common in thousands of laboratories around the world. Often referred to as a DNA chip, a microarray is usually a piece of glass about the size of a microscope slide that is coated with a grid of thousands of spots of DNA. Each spot contains millions of copies of short DNA sequences, and a computer keeps track of what’s where. The beauty of the array is that it makes it possible to measure the activity of thousands of different genes at the same time.

“ The problem,” DeLisi says, “is that the arrays currently don’t work very well, and more and more people are losing confidence in them. Probably 90 percent of what we see on an array is in error.” He and his colleagues have developed software to dramatically improve the accuracy of the microarrays.

BSL researchers are also developing high-throughput methods for identifying sets of proteins that are activated and deactivated as conditions change inside a cell. With assistance from BU’s Fraunhofer Center for Manufacturing Innovation, BSL is forming Boston Array Technology, a company that will produce a new kind of microarray for identifying thousands of proteins using only the genome sequence of the organism being studied.

Flipping the switch

Genes are blueprints for proteins, but proteins called transcription factors are responsible for turning genes on or off by attaching to stretches of DNA called cis-elements, which are toggle switches that tell the cell to express nearby genes. Zhiping Weng, an ENG assistant professor of biomedical engineering, is developing computational methods for mapping the precise locations of these switches in the genome, and is also trying to find out what combinations of transcription factors turn genes on and off. As the director of CAGT’s laboratory for protein engineering (called Zlab, for Zhiping’s lab), Weng wants to understand how proteins control the behavior of cells. “In the human genome, there are maybe 30,000 genes,” she says, “and for each one, we want to know what cis-elements are functional. The ultimate goal is to find out for each gene how many transcription factors are controlling that gene, when they become functional, and how the cell is controlled in terms of transcription.”

Protein landscape

Transmitting a message from the surface of a cell to its nucleus, where genes reside, involves a daisy chain of interactions between messenger proteins. Sandor Vajda is using high-speed computation to understand how proteins physically interact with one another and with certain small molecules (aka drugs). As director of CAGT’s structural bioinformatics laboratory (SBL), Vajda is developing computational methods for predicting which proteins will stick together, and what the married proteins will look like as a couple. “Determining the structure of two interacting proteins is extremely complex,” he says, “because you have to generate billions of relative conformations of the two proteins. And they’re not rigid bodies, so we have to take into account that they are moving. To follow all of this movement is an immense amount of computation.” Vajda also wants to find small molecules that can cut in and interrupt the dancing pairs of certain proteins. Studying the docking of small molecules is key to the development of rational drug and vaccine design strategies.

Using computational models of the human genome, Zhiping Weng, an ENG assistant professor of biomedical engineering, is searching for the precise locations of so-called cis-elements — stretches of DNA where proteins attach and turn nearby genes on and off. Ultimately, she wants to understand how signaling proteins choreograph the behavior of cells. Photo by Kalman Zabarsky

Computational genomics laboratory

The genomes of over 100 species of bacteria are now easily accessed online, but in most cases researchers do not know what over a third of the genes do in the genome. “One of the great surprises of the genomic revolution was discovering how little we actually know about organisms that we previously thought we understood very well,” says Simon Kasif. “We can identify the genes, we know they’re translated into proteins, but we have no idea what they do.”

The main thrust of the computational genomics laboratory (CGL), directed by Kasif, who is also the codirector of CAGT, involves developing new technologies for classifying undescribed genes into broad functional categories. To do this, Kasif’s group compares genomes from different organisms or different strains of the same organism. In this way, they isolate interesting genes that are either changing quickly or not at all, and they are able to infer the function of mystery genes. “There’s been a tremendous push to understand what these genes might be doing,” Kasif says. “All of the new genes recently discovered open up a whole set of opportunities for discovering new diagnostic procedures or identifying drug targets and vaccines.” The lab has focused on groups of genes that appear to play a crucial role in infecting hosts, because they are natural drug targets. In addition to identifying new genes and assigning them biological functions, Kasif and his colleagues are building computational frameworks that use experimental data to predict the behavior of cells in normal or perturbed conditions, which promises to have an impact on drug design and screening. For more information about CAGT and its associated laboratories, see www.bu.edu/cagt.

14 November 2003
Boston University
Office of University Relations