Brown Researchers Use Wildcards To Develop Better Way To Sequence DNA

When sequencing DNA, scientists look for the arrangement of bases (amino acids represented by the letters C,T,G and A) contained in a particular strand of DNA, which is too small to be seen under a microscope.

In the decade-old hybridization method, short sequences of DNA bases – six to 20 bases long – are used as probes to find the bases contained in a long string of DNA. The DNA to be sequenced binds to some of the probes on a tiny glass chip called a microarray.

The target DNA binds to probes that match its sequence (T always pairs with A, and G always pairs with C). Then the DNA can be detected and read. Using the information obtained from the probes on the microarray, a computer program is then used to construct the long sequence of bases in the target DNA strand. Current techniques can construct only a few hundred bases.

“What we show is that you would be able to sequence a substantially longer piece of DNA,” Upfal said. “Instead of a few hundred bases, you could do tens of thousands.”

Based on a novel mathematical insight, Preparata and Upfal have designed a new pattern for the probes on the chip. In the new pattern, some of the DNA bases in each probe are replaced by gaps that act like wildcards. The new design extracts substantially more information about the DNA than previously designed probes.

Together with the new probe design, they have also developed a new algorithm, tailored to the new probe pattern, that uses the information obtained from the probes to reconstruct the original sequence.

Preparata and Upfal want to test chemical compounds called universal bases for use as the gaps in DNA probes. A perfect universal base would attach itself to all four bases in DNA.

“If the universal bases behave closely to ideal, this is by far a superior method,” Preparata said. “It would be a fundamental change in the way of doing sequencing.”

For this work, he and a group of colleagues have received a National Science Foundation grant of $850,000 over two years. During the research, the computer scientists are designing and analyzing algorithms, and Brown biochemists are using probes that consists of mixtures of DNA bases and universal bases to test how well the mathematical models work.

“The potential power of our findings may change the way medical diagnosis is practiced,” said chemistry professor Kathlyn Parker, who is leading the chemistry part of the research. The research will also help scientists learn about fundamental interactions of DNA, which will have many applications.

The research project is unusual because it integrates biochemistry and computer science, two disciplines that don’t easily mix. “They are two really different languages,” said Upfal.

Preparata and Upfal have presented two major papers and published two journal articles on the work. Information is online at www.cs.brown.edu/research/sbh/.