Accelerating lessons on life

THE BALTIMORE SUN

The crimson heart beating inside Raimond Winslow's tiny office has begun to twitch. The diagnosis: heart failure. In a few moments, it will be all over.

Yet the 44-year-old Johns Hopkins University engineer appears unconcerned. In fact, he seems to be enjoying the show.

That's because this organ beats not inside a living carcass but an IBM SP3 supercomputer. Winslow, a pioneer in the fast-growing field of computational biology, is aiming to simulate a flesh-and-blood heart "from literally the gene on up."

The project is one of many examples of how computer science is revolutionizing the life sciences.

In laboratories around the country, computers are becoming the test tubes of 21st century biology. Life scientists who a decade ago were content to putter around with Petri dishes and pond water are flocking to desktop PCs and even more exotic hardware to unlock the secrets of life.

The most vivid example of how this marriage is taking shape occurred Monday, when public and private research teams announced they had pieced together most of the 3.2 billion chemical building blocks that make up human DNA.

Solving this complicated molecular puzzle was a triumph of machine as well as man. Celera Genomics in Rockville, the first company to decode the human genome, relied on more than 300 gene-reading robots and one of the most powerful supercomputers in private hands to do the grunt work.

Life-saving work

Technology like this, scientists say, will be even more crucial for the first challenge of the "post-genome" era: the race to sift through the gobbledygook of As, Cs, Ts and Gs - the initials of the chemicals that form DNA - to discover lifesaving drugs and therapies.

"The human genome is not the end point - but the starting point of the next race," said Jesse Lipcon, vice president of Compaq Computer Corp.'s high-performance computing group, which supplied Celera with its computers.

Cracking open the human genome has opened a wealth of new data to biologists. In fact, they're drowning in it.

The National Institutes of Health last year estimated the typical biomedical lab churns out as much as 100 terabytes of data each year - enough information to fill 1 million encyclopedias. And it's going to increase.

Doubling of information

Thanks to the spread of robotic sequencers, biologists are quickly assembling vast digital DNA zoos containing the genomes of creatures ranging from microbes to man. More than 30 living creatures - now including Homo sapiens - have had their DNA fully decoded. Thousands of others reside in these databases in bits and pieces.

GenBank, the first and largest online public DNA database, contains 8,604,000,000 genetic letters from more than 70,000 organisms. And the National Library of Medicine in Bethesda, which manages GenBank, says its holdings are doubling every 14 months.

'A revolution'

"The practice of biology has changed dramatically in a way that amounts to a revolution," said Eugene Koonin, a molecular biologist at the National Center for Biotechnology Information in Bethesda.

Koonin recalls when he decoded DNA by hand, and then punched the mind-numbing strings of As, Cs, Ts, and Gs into his computer one letter at a time.

He regularly swaps genes with collaborators through e-mail and spends hours wading through online DNA databases looking for patterns and clues to the functions of new genes - a hot new science known as "bioinformatics."

Eager to tap the growing amount of genetic information showing up online, some biologists spend 90 percent of their time at the computer, according to a recent report from the National Institutes of Health.

Path to discoveries

"Going through genome databases is sometimes the only way to make discoveries," said Malcolm Gardner, a molecular biologist at The Institute for Genomic Research (TIGR) in Rockville, where scientists are compiling an online DNA library of microorganisms responsible for diseases ranging from stomach ulcers to syphilis.

Gardner, who has spent the past 15 years trying to unravel the secrets of malaria, knows the importance of these databases first hand.

After he decoded a new malaria gene and put it in TIGR's database, German scientists trolling this library of bad bugs discovered that the malaria gene matched a plant gene they were working with.

The serendipitous connection, Gardner said, might lead drug companies to a new weapon against the malaria, which kills 1.1. million people a year.

Bioinformatics is only one of the important new fields to emerge from a marriage of bytes and biology. Another is computational biology, or using computers to simulate the intricate inner working of everything from human organs to tiny proteins.

"Computational biology is going to be at the forefront of biological science in the coming century," said Hopkins' Raimond Winslow.

Checking proteins' shapes

One of the hottest problems in computational biology is protein folding.

"Almost anything you can think of in the body is done by proteins," said Michael Levitt, head of Stanford University's department of structural biology. "And what really matters is their shape."

Just as a hunk of metal functions differently depending on whether it's molded into a knife or fork, the function of proteins in the body also depends on their form.

Building supercomputer

In December, IBM launched an effort to build a $100 million supercomputer, nicknamed "Blue Gene," that will attempt to predict how proteins curl, wad or bend based on their DNA blueprint.

Unraveling this biological origami on today's computers would take centuries. Blue Gene, which is expected to be 500 times more powerful than the fastest existing computer, is being designed to complete one protein a year. This supercharged machine is expected to throw off so much heat that IBM scientists have bought a gas turbine the size of a jet engine to cool it.

Hunting for drug 'targets'

High-speed computers are also being used by pharmaceutical companies to hunt for new drug "targets" on proteins, which are riddled with crevices and grooves. Finding a molecule that slips into these nooks and crannies much as a key fits a lock is the key to stopping disease.

In more than a century, scientists have identified fewer than 1,000 targets. Until very recently this was mostly done by white-smocked chemists swirling chemicals together, hoping to stumble across a cocktail that did the trick.

But drug industry scientists estimate the completed human genome could mean as many as 20,000 new drug targets up for grabs. The problem is: How to find the ones that work?

Developing medicines faster

Increasingly, pharmaceutical companies are turning to technology for help.

It takes on average four to six years to identify a promising new drug. Many drug companies are using high-speed computers to quickly accept or reject molecular drug candidates.

"You can now screen more in a day than we used to screen in a year," said Martin Haslanger, executive director of research technologies at Eli Lilly & Co. "The computers are finding all kinds of experiments to do that we didn't know existed."

To see the future of drug discovery, peek inside the laboratory at the Bristol-Myers Squibb's research center in rural Connecticut, where scientists sporting what appear to be oversized wraparound sunglasses sit around a conference table, seemingly poking at thin air.

But those glasses are part of a cutting edge virtual reality system that allows researchers to visualize the molecule they're working on in 3-D. Through the glasses, a molecule appears to spring to life in the middle of the conference room.

"The molecule is live and rotating," said Jonathan Mason, director of the company's computer-assisted drug design program. "So you can point to a particular spot and say, 'What if I modify the molecule here?'"

Said Eli Lilly's Haslanger: "It's like the discovery of a new world."

Copyright © 2019, The Baltimore Sun, a Baltimore Sun Media Group publication | Place an Ad
36°