October
2004
In 2002, Lydia Sohn helped blow the whistle on an esteemed Bell Labs physicist who falsified data in high-profile scientific publications. (courtesy the researcher) |
A tiny chip developed by a UC Berkeley mechanical engineer is now being tested as a super-fast bioterrorism sensor for the battlefield. The same technology could eventually lead to a disposable disease detector that brings cheap, easy, and incredibly accurate blood tests out of the clinic and into the rural villages of developing nations. To build such a device though, professor Lydia Sohn looked to nature for inspiration. The result is a silicon chip laden with artificial nanopores that mimic the filtration system of human cells.
"My background is in solid state physics and nanotechnology," says Sohn, formerly a physics professor at Princeton University. "But six years ago, I realized that the electronics I was working on were so sensitive that the same concepts could be used for biological detection."
Currently, blood samples are screened for pathogens or diseases mostly through optical detection of antigens or antibodies. Antibodies are formed by the body in response to antigens-- molecules, often foreign, that the immune system recognizes as threats. For every antigen, there is an antibody that binds to it. In one common test, an enzyme is added to a sample that activates a visible colored dye in the presence of a particular antigen or antibody indicative of a particular disease. The problem is that this test requires laboratory equipment not suited for the battlefield or rural villages where even clean water may be a luxury.
On the other hand, Sohn's device--first developed in collaboration with former Princeton graduate student Omar Saleh--is entirely integrated on a single, inexpensive electronic chip. Because the biosensor is fabricated using processes similar to the way integrated circuits are manufactured, the devices can be produced in bulk at very low cost.
This scanning electron micrograph depicts a nanopore made from quartz. More recent prototypes are fabricated from rubber, an easier material to work with. (courtesy the researcher) |
A cell's membrane is riddled with tiny pores, each engineered by evolution to allow a certain substance to pass through while blocking others. Molded out of silicon rubber, Sohn's artificial nanopore consists of a tiny channel just one micron in diameter and around seven microns long. (A human hair is 100 microns thick.) The pore is filled with a conducting fluid so that tiny electrodes on the chip underneath can measure the current across the channel. Small plastic particles of specific sizes, called colloids, flow through the channel. As a colloid moves through the pore, it displaces some of the conducting fluid and reduces the current by a predictable amount. Differently-sized colloids can be coated with specific antigens that correspond to four different diseases or pathogens.
"We then dunk the device into a solution of blood," Sohn says. "If a certain antibody is there, it will be bind to the specific colloid for that disease, increasing the size of the colloid by a few nanometers."
When a colloid with an antibody bound to it flows through the pore, more of the conducting fluid is displaced than if the colloid passed through by itself. The additional drop-off in current signals the presence of the antibody while the amount of the drop indicates which colloid of a particular size is in the pore and, subsequently, the precise pathogen in the blood.
Sohn's laboratory recently shipped 100 of the prototype biosensors to the US Army's Edgewood Chemical Biological Center in Maryland where the technology will be further developed for potential military deployment. Meanwhile, Sohn hopes to expand the device's capabilities.
Employing traditional lithographic manufacturing enables arrays of nonpores to be built on a single silicon wafer, she says, much like a multitude of transistors are etched into a silicon chip. For example, a ten-by-ten array of pores, with each pore capable of detecting a minimum of four diseases, could result in a single sensor that can screen for four hundred diseases simultaneously.
"There's nothing like that on a single chip right now," Sohn says. "And the fact that it's electronic and requires so little preparation means that you can imagine someday having everything in a cell phone-type device that can transmit the results of the test to a physician far away."
The Engineer, the Rat, and the Fruit Fly
by David Pescovitz
Anat Caspi helped sequence the genome of the laboratory rat, only the third vertebrate whose genome has been deciphered. (courtesy NHGRI) |
Engineering student Anat Caspi has an intimate relationship with rats, mice, and flies. She listens to their deepest secrets and poses an unending number of questions about their pasts. Caspi is not as odd as she sounds though. A member of the Bioengineering Graduate Group, Caspi uses novel computer software to unravel the evolutionary mysteries of lab rats, mice, fruit flies, and humans.
"The real crux of the research I contribute to is to provide a detailed window not only into the evolution of rodents and flies but evolutionary and genomic processes in general," says Caspi, a graduate student of mathematics professor Lior Pachter.
In March, Caspi was one of more than 200 co-authors of a monumental collection of scientific publications outlining the genome of the common laboratory rat. It was only the third mammalian genome to be sequenced, following the mouse and the human. The ability to read and compare the genomes and evolution of rats, mice, and flies sheds light on our own genetic make-up. Armed with data from these various animal sequences, researchers will be better able to develop new weapons in the battle against human disease. The rat is of particular interest because it's commonly used in biomedical research and drug development.
The research was a massive collaboration that included a team led by Pachter and groups from Stanford University , UC Santa Cruz, and Pennsylvania State University. Three other Berkeley engineering students--Colin Dewery and Sourav Chatterji, graduate students in the Department of Electrical Engineering and Computer Sciences, and undergrad Kushal Chakrabarti--were also co-authors on the papers published in the journals Nature and Genome Research.
"The sequencing of the rat genome constitutes another major milestone in our effort to expand our knowledge of the human genome," said Francis S. Collins, director of the National Human Genome Research Institute. "As we build upon the foundation laid by the Human Genome Project, it's become clear that comparing the human genome with those of other organisms is the most powerful tool available to understand the complex genomic components involved in human health and disease."
In both the rat effort and the ongoing Berkeley Drosophila Genome Project, Caspi focuses on computationally comparing laundry lists of the animals' genes to identify transposons, known as "jumping genes." These segments of DNA can literally hop from one location in the genome to another, commonly causing mutations. For example, transposons are what enable bacteria to mutate and become resistant to drugs.
"We'd really like to understand more about the mechanism that causes these mutations to occur," Caspi says.
In the case of the rat, Pachter's laboratory was provided with a rough draft of the sequence. They then used software they had developed to align the rat genome with that of the previously-sequenced mouse and human. Once aligned, Caspi and her colleagues can generate computer models and sift through the data for similarities and differences, including transposons that are common between the animals.
"The comparative genomic techniques help us identify the elements of a genome that have been conserved by evolution as well as what has changed," Caspi says.
According to the Nature paper, most human genes associated with diseases also have counterparts in rats. Meanwhile, the data also revealed that 40 percent of the genome of rats, humans, and mice is inherited from a common mammalian ancestor.
As more genomes are sequenced, Caspi hopes to improve genetic models to include events such as transposition and other mechanisms leading to mutations. For example, plants and viruses, she says, exhibit many genomic events that scientists are only beginning to uncover.
"Understanding the processes that plants and viruses undergo when they evolve would be a big jump in agricultural and pharmaceutical development," she says.
James O'Brien leads the Berkeley Computer Animation & Modeling (B-Cam) group. (David Pescovitz photo) |
We live in a viscous world of bubbling and oozing fluids. From mud to blood to paint, much of the stuff that surrounds us is neither perfectly liquid nor perfectly solid. Digitally simulating these materials' changing properties is an historically difficult challenge in computer graphics. But UC Berkeley researcher James O'Brien has developed a novel way to animate viscoelastic fluids that could bring movies, videogames, and even surgical simulations much closer to reality.
"For quite some time, we've had mathematical models for simulating idealized solids," says O'Brien, a professor in the Department of Electrical Engineering and Computer Sciences. "And we have other models for simulating idealized liquids. But there are many materials that are in between. They may look like solids, but they also can flow like liquids."
Take clay, for example. It usually maintains a solid form, but squeezing the clay causes it to flow. If a special effects artist wants to add a spray of mud to a scene of a spacecraft crashing into a forest, "treating the computer-generated mud like viscous water just won't look real," O'Brien says.
The same holds true for blood and mucous. The ability to realistically render those materials using a computer won't just up the gore ante in horror films either. Eventually, medical students might practice their skills using "virtual surgery" systems. Creating a realistic surgery simulator though depends on lifelike models of the body's internal organs.
"One of the place where you'll see a ton of fluids with all sorts of very interesting and different properties is inside the human body," O'Brien says. "If you're able to model the human body with a high degree of accuracy, a surgeon can train on a simulator like a pilot trains on a flight simulator."
The algorithm O'Brien developed with graduate students Tolga Goktekin and Adam Bargteil takes existing fluid simulations and adds the ingredient that gives materials like toothpaste, motor oil, and dish soap characteristics of both liquids and solids. The key difference between basic fluids and solids, O'Brien explains, is the presence or absence of elastic forces.
For example, if you gently try to bend a spoon just a little bit, elastic forces cause it to spring back to its original shape. But apply too much pressure and the molecules flow into a new configuration, leaving the spoon bent.
"The same thing happens with viscoelastic fluids," O'Brien says. "If you squirt a little ketchup on a plate, it doesn't flow into a puddle but rather sits there as a glob. That's because there's a very small amount of elastic force."
O'Brien's algorithm is rooted in the mathematics of Level Sets, a method developed in 1988 at UC Berkeley and UCLA to track and simulate the shifting boundaries of many dynamic materials.
"When the strain in a viscoelastic material is too large, it starts to flow," O'Brien says. "By setting those thresholds low and modeling the flow properly, the simulated viscoelastic fluid behaves correctly."
The researchers presented their work at the SIGGRAPH 2004 computer graphics conference in August. A short film demonstrating the algorithm was shown as part of the popular Electronic Theater program. Applying the technique to interactive simulations such as virtual surgery are still several years away though. Faster computer processors and more efficient algorithms are necessary to generate viscoelastic animations in real-time.
More advanced versions of the algorithm will likely be seen on the big screen before the operating theater. Already, several movie studios have expressed interest in the technology. Once the basic fluid simulation framework is complete, O'Brien says, it will be surprisingly easy for special effects artists to use the tool.
"If special effects are done right, you shouldn't even notice that they're there," he says.
James Hong and Jim Young, as photographed for People magazine. |
Four years ago, UC Berkeley Electrical Engineering and Computer Sciences alumni James Hong (BS '95) and Jim Young (BS '94, MS '97) were drinking beer in their apartment and talking about, what else, girls. Jim commented that he thought a girl they recently met was a "perfect 10." That discussion led to a half-joking plan to launch a Web site where visitors could submit their photos for other people to rate from one to ten. A few days later, HOTorNOT.com went online and quickly became one of the, well, hottest sites on the Web. Mentions in Entertainment Weekly and People followed, along with a Webby Award and Hong and Young's appearance on the Sally Jesse Raphael Show.
Since then, undeniably addictive HOTorNOT and its parent company Eight Days Inc. have expanded into a multi-million dollar online dating service with more than 600,000 active members. Meanwhile, the number of user-submitted photos to the rating site have topped 13 million.
Now, Young, 31, and Hong, 30, have put up $200,000 of their own assets to encourage people to rate the presidential candidates, not by clicking the mouse but actually visiting the polls. On Labor Day, the two launched the VOTEorNOT sweepstakes. Of course, it's illegal to pay people to vote, so the contest goal is simply to raise awareness about the importance of casting a ballot. Anyone who signs up on the VOTEorNOT Web site is eligible for a post-election random drawing for a $100,000 prize. Also, if an entrant refers someone to the site who goes on to the win the prize, the referrer also pockets $100,000.
The aim of the contest, Hong says, isn't to push for any particular candidate over another.
"Part of this is inspired by a book Jim and I both read, The Wisdom of Crowds. The basic premise is that the crowd in many cases is smarter than individuals, based on having a large diversity of people inside who make their own judgments," Hong told Wired News. "What that means is, especially since this election is so close, the more people we can get to vote, the more likely the right answer will appear."
Even with their newfound political mission, Hong and Young haven't lost site of their original idea. Photos of the 2004 presidential candidates in their younger days are available for HOTorNOT ratings.