Scientists push the publishing of code powering genetic research

By Paul Elias The Associated Press
Thursday November 22, 2001

SAN FRANCISCO — Before computer whiz Steven E. Brenner accepted his tenure-track research post at the University of California-Berkeley last year, he demanded that the school’s intellectual property police leave him alone. 

Brenner prevailed. He’s now one of the few experts in the emerging field of bioinformatics with the freedom to distribute his work, software used in gene research. 

“It’s vital to what we do,” says Brenner, who supports a movement to force universities to allow “open source” publishing of gene research software code. 

It’s a somewhat quixotic movement, since universities’ 2,000 yearly patents now provide 10 percent of their budgets, about $5 billion. With government funding on the decline, schools say they need to profit from faculty research. 

The movement also runs counter to U.S. laws that permit publicly funded schools to enter into exclusive licensing agreements with private companies. 

Unlike scientists who keep research secret until it is published in a peer-review journal, some software developers — who get little credit when their code leads to a genetic breakthrough — want to share their work as soon as it leaves their keyboards. 

It’s an old debate in the world of computing — and a new culture clash in bioinformatics, the practice of using computers to search genetic material for potential cures that has caused such excitement in the medical community. 

The problem with not sharing the software code of bioinformatics programs, say researchers like Brenner, is that bugs can go unnoticed, hindering scientific advances. 

With collaboration, open-source advocates say, the quality of bioinformatics software will improve. 

But universities — and some programmers — oppose the open source movement, fearful valuable trade secrets could be lost. 

Private corporations that increasingly fund academic research are no different. 

Berkeley, for instance, is in the middle of a five-year, $25 million deal with Swiss-based agriculture giant Syngenta, a Novartis Corp. spinoff that funds research in Berkeley’s department of plant and microbial biology. 

For $5 million a year, Syngenta gets to license whatever is invented by most of the department’s scientists. Researchers who accept Syngenta’s money are barred from showing their software outside of the university without permission. 

Open source advocates object to such private agreements, along with universities using federal grants to create private intellectual property. 

“If taxpayer money is used to create the software, then it should be publicly available for free,” said Harry Mangalam, of tacg Informatics in Irvine, Calif. “The public is being billed twice right now.” 

Mangalam is one of three bioinformatics developers circulating a petition calling on the National Institutes of Health and the National Science Foundation to require that bioinformatics scientists who receive federal grants make their code freely available for peer review. 

The request runs counter to the Bayh-Dole Act, which three decades ago allowed federally funded universities to enter into exclusive licensing agreements with private companies and profit from patented research. 

“I think the Bayh-Dole Act is one of the great economic success stories in the nation,” said Terry Young, executive director of the Texas A&M Technology Licensing Office. He says the law should remain untouched. 

But the petitioners have a sympathizer in Gary Strong, the acting executive officer for computer science at the National Science Foundation. 

Strong thinks the NSF and other government agencies could make publicly posting bioinformatic code a requirement in their grant specifications.  

Whether this will happen is up to Bush administration appointees yet to be named. 

Even the petitioners concede that Bayh-Dole is untouchable. They’re hoping for a minor rule change or other legislation that could carve out an exception for software. 

“This can be very difficult to deal with when universities believe that every patent they hold is a winning lottery ticket,” said Karin Lohman, a Democratic staff member with the U.S. House Science Committee. “Only a very few patents bring in significant revenues.” 

One of the few bioinformatics experts who has made money for his university is Phil Green. His patented, widely used code generates about $10,000 per license for the University of Washington. But he says the money isn’t the issue — it’s respect. Open source publishing devalues what they do, he said. 

“I don’t think computer programmers should be treated any differently than other scientists,” Green said. “It sort of diminishes the stature of the science.” 


On the Net: 

Open source site:http://www.open-bio.org 

Petition site: http://www.openinformatics.org 

National Institutes of Health: http://www.nih.gov 

National Science Foundation: http://www.nsf.gov