This cadre of receptors insert themselves into the surfaces of the immune cells, waiting to spring shut on tell-tale proteins from bacteria or other invaders, to trigger the immune system to attack.

Cowell, a postdoctoral fellow, recently received a $500,000 "Career Award at the Scientific Interface" by the Burroughs Wellcome Fund. The award recognizes her as exemplifying a new breed of interdisciplinary scientists who are joining with biomedical researchers to make sense of the vast amount of data emerging from the sequencing of the human genome and that of other animals.

"What weâve begun to do is develop theoretical models of these binding sites in DNA that are necessary for a very important recombination process," said Cowell. "Itâs important for generating diversity in the antibody repertoire and in the repertoire of receptors important for immune function." Also, said Cowell, malfunctions in such rearrangements can produce runaway growth genes responsible for lymphomas and leukemias.

The problem in recognizing such signal sequences, said Kelsoe, is that while parts of the DNA sequences seemed to be "conserved" across most such signals, other stretches of DNA seem to be so variable -- or "degenerate" -- as to be little more than genetic gibberish. That is, while certain segments had more standardized sequences of the genetic units called nucleotides -- abbreviated A, T, G and C -- others segments seemed random.

Said Kelsoe, "Biologists had previously characterized these sequences in a way that hid a great deal of information. The variable sequences were simply ignored, and the assumption made essentially by the entire field, that there was nothing of interest there.

"It was known that the recombinase enzymes could recognize recombination signals that were so degenerate that even an expert looking at the sequence would not be able to recognize them," he said. "And yet, occasionally, the enzyme would show us these restriction sites by producing a break at exactly the point. So, it was clear that there had to be more information in the signals than was appreciated broadly by the field."

Working from such clues, said Kelsoe, Cowell and her immunologist colleagues developed statistical methods to recognize previously hidden patterns within these variable regions. These patterns went beyond the simple sequences, to more subtle relationships among the units.

"It was Lindsayâs insight that led to a statistical approach that allowed her to understand that there was a great deal of information hidden in these variable sequences, but in a way that immunologistsâ training did not allow them to recognize," said Kelsoe. Such information, explained Kelsoe lay in the "higher order" relationships among related nucleotides in the sequences, as opposed to "simple sequence order in which the fourteenth nucleotide was always an A and the seventeenth a T." To Kelsoe, a key to Cowellâs success has been her understanding of both mathematics and immunology.

"Many mathematicians who run three hundred thousand simulated experiments over a coffee break donât appreciate what itâs like to do real bench work -- for example laboriously isolating a gene from a cell," said Kelsoe. "Lindsay does, since she has experience in the laboratory, but she also brings to biology something that we badly lack -- a strong sense of quantitation.

"While an immunologist or cell biologist might be happy just getting a single given result, sheâs more interested in the frequencies of such a result. And that way of thinking led her to the mathematical analysis of the recombination signaling sequences that yielded this new model."

For Cowell, her work with Kelsoe and his colleagues represents the beginning of a career of bringing the tools of mathematics to bear on the medically critical questions of immunology.

"In the long run, I would really like to significantly increase the interest by immunologists in collaborating with scientists with more theoretical backgrounds like mine," she said. "And, Iâd like to help recruit people from math, statistics and computer science to be interested in immunology, because immunology is so incredibly fascinating and complex.

"Just about every question in biology can be asked in the context of immunology, because these immune cells are so incredibly complicated," said Cowell. "And so, not only does the complexity of the immune system provide a wonderful context for applying these statistical methods, but I also believe that bringing these methods into immunology will significantly increase the amount weâre able to learn." Duke, in particular, is in a prime position to foster such collaborations, said Cowell.

"With the universityâs Center for Bioinformatics and Computational Biology, as well as an extraordinary medical center, I think Duke provides the perfect environment to develop a group thatâs very focused on precisely these kinds of problems," she said.

Also, said Kelsoe, Dukeâs open research culture, as exemplified by the Department of Immunology, encourages such research partnerships.

"I think our department is extremely special, even among Duke departments, in that itâs very open in what it believes is important research that contributes significantly to the body of knowledge," said Kelsoe. "We believe that there are many ways to answer questions essential to immunology, and weâre pretty open to anybodyâs approach as long as itâs rational, testable, and eventually turns out to be true."