PALO ALTO, California -- "We want theoretical computer scientists."
The declaration is perhaps not as stirring as Uncle Sam calling patriotic men into the army in World War I, but Vicky Markstein is just as determined in her pursuit and convinced of its importance to the nation and the welfare of humanity.
As chairwoman of the inaugural IEEE Computer Society Bioinformatics Conference, held at Stanford University last week, Markstein is trying to recruit the leading minds of computer science into what she calls "the industrial revolution of biology," an anticipated period of discovery resulting from the analysis of genomic information.
To spark the revolution as she intends, however, Markstein must first bridge a gap between two academic fields that do not always speak the same language.
"In math, everything is calculated and exact," says Lidia Martin, a chemist at Barcelona's Prous Science. "In biology ..." she adds, and shrugs her shoulders.
Jesus Salillas, a computer scientist who works with Martin studying and modeling pharmaceutical molecules, agrees that they look at problems differently and have had to work at establishing a common language for collaboration.
Martin and Salillas, and others at the conference, talk about a type of culture clash that arises because computer scientists like to see problems in the abstract -- in terms of sequences and string theory -- whereas biologists are very empirical.
Sven Rahmann, a mathematician at the MPI for Molecular Genetics in Berlin, notes, "A biologist says, 'That doesn't make sense from a biological point of view,' and I say, 'Well, but it's a nice algorithm.' You want to formalize something, but then (in biology) there are ten thousands of exceptions."
Markstein contends that these differences must be overcome, as computer scientists will be essential for deriving useful knowledge from the large databases of genomic data that have recently become available. "(Biologists) are going to need these computer people because these (genomics) problems are so large," she says.
According to Dr. Bernardo Huberman, a researcher at HP Labs, the problem is not necessarily attracting computer scientists to the field, but motivating them to do work that is of interest to biologists. He says that computer scientists are eager to try their methods and algorithms with the large data sets in biology, but they tend to enter the field with "a solution looking for a problem."
As a result, Huberman concedes that "biologists find a lot of work being done rather uninteresting," and notes that most bioinformatics articles are often read only by computer scientists.
Biologists say that for the academic partnership to work, computer scientists must learn more of the underlying physical science, which they have been loath to do. Dr. Leroy Hood, a keynote speaker at the IEEE gathering, says, "I think most bioinformaticists at this conference don't understand biology very well."
"Bioinformatics has reached the stage where further biological input is necessary to go forward," agrees Dr. Abraham Lempel, another researcher at HP Labs. Lempel helped develop the compression algorithm for creating zipped files, which he thinks can be applied to DNA matching, but he says he needs to understand the science better before pursuing that research.
Lempel says that to collaborate effectively, computer scientists and biologists simply need "to interact and discuss and not look in a downward direction (at each other)."
For Vicky Markstein, the importance of merging the two fields is worth the effort involved in getting everyone on the same page. She says, "The holy grail for all of this is that we'll really get a handle on human disease."
"If you throw enough smart people at the problem, then you have a better chance of getting something done."
