In the middle of the past decade, a neuropathologist at Harvard was frustrated with the slow pace of academic research. Studying the mechanics of multiple sclerosis in the brain, Kari Stefansson wondered if there was a way to hasten the development of genetic-based tests to predict and better understand MS and other diseases.

Like other scientists with an entrepreneurial bent, Stefansson’s solution was to turn to commerce as a way to further his science. In 1996, after leaving Harvard, he founded deCode Genetics in Reykjavik, Iceland, his hometown.

His idea was in part an experiment to see if genomic research–then all the rage as scientists raced to sequence the human genome–could lead quickly to new drugs and thus to profits. This was a subset of the grand experiment of biotechnology itself, which over 35 years has seen scientists sink tens of billions of dollars into mostly early-stage projects that in some cases has provided breathtaking cures but more often has not. As an industry, biotech has suffered losses every year except one.

In 1996, deCode’s plan was to scour the genomes of Iceland’s 300,000 people in search of genetic risk factors for common maladies such as heart disease and cancer. The project was made possible by a controversial act of that country’s parliament allowing the company access (with consent) to citizen’s medical records. In 2000, the company went public on the NASDAQ for $18.00 a share, eventually raising $172 million. In the first day of trading, the stock popped up to $29.00.

At the time, a major venture capitalist and a lead investor in deCode told me that deCode and other genomics companies such as Millennium and Celera represented the future of medicine.

Nearly 13 years after deCode’s founding, Kari Stefansson may be the last man standing from the original genomics crowd–and his company is in deep trouble.

A few days ago, the company announced that it had only $3.7 million left in cash. “The company believes it has sufficient resources to fund operations only into the latter half of the third quarter,” reported a recent deCode press release.

But Stefansson says deCode is not yet ready for an obituary. “We are restructuring to focus on our core activities,” he says, having told investors in a call on August 10: “We are making good progress in recasting deCode as a leading gene-based diagnostics company.”

Perhaps more important is the precarious position of deCode’s genomics research effort. For more than a decade, it has produced a prodigious number of important papers identifying genetic risk factors and published their findings in journals such as Nature and Science.

Most of these papers describe genome-wide association studies (GWAS) that scan human DNA looking for genetic markers associated with common disease. Few labs have produced such high-quality work, a fact that has frequently landed the company’s discoveries on the front pages of the New York Times and in other major media. Just last month, deCode scientists discovered genetic risk factors for heart disease, stroke, skin cancer and schizophrenia.

Lately, the usefulness of GWAS markers in predicting an individual’s proclivity for disease has been challenged in articles in the New England Journal of Medicine and other publications. And virtually every other genomics company founded in the 1990s long ago crashed, got acquired, or abandoned their pure genomics efforts. Yet few geneticists doubt the value of large population studies for providing clues to predict and better understand common diseases–the goal Stefansson originally set out to achieve in 1996.

So far, deCode has spent more than $600 million and has failed to ever turn a profit, even as it moved into drug development and began selling diagnostic tests based on its DNA discoveries. In 2007, deCode launched deCodeme, a consumer genetic-testing website that is a rival to those offered by 23andme, Navigenics and others in this nascent space.

DeCode is partly a victim of the recession and of the meltdown of the Icelandic economy last year. But the company also has failed to find a model for what is still in large measure a very pricey experiment in how to commercialize basic research. Even with drastic cuts in spending and in workers, deCode still suffered losses of $24.3 million in the first six months of 2009. The stock was at 57 cents as of this writing, up from 19 cents earlier this year but far below the company’s heady days.

Stefansson says that the company is planning to sell off its drug development program, including its three compounds currently being tested in humans–two for heart attack and one for arterial thrombosis. “A deal is in the works,” he adds.

As for the possible demise of its research operation, I suspect that deCode’s high-profile collaborators in academia will come to the rescue if all else fails. Recent papers list the likes of the National Institutes of Health and the Wellcome Trust in the United Kingdom as financial supporters of individual studies.

One thing is certain: Kari Stefansson is a survivor and an unrepentant advocate of the power of genomics as both science and a source of potential profit,even if it remains more promise than reality. “At the end of the tunnel, once we emerge,” he insists, “we have a business, we believe, with spectacular potential.”