Reinventing the Sequencer

Focused on the End User

For Efcavitch, working on a high-risk project is nothing new. For two decades at Applied Biosystems, he led team after team through the development of products, including the model 3700 that ultimately delivered Craig Venter's genome in 2002. "We were under a tremendous amount of pressure to get that platform out the door. It was a very aggressive timeline," he says.

The 3700 may be the instrument for which Applied Biosystems is best known, but Efcavitch says he's more proud of his other accomplishments at the company, which he joined in 1981 after doing a postdoc in Marvin Caruthers' laboratory at the University of Colorado. At the time, Efcavitch was among just a handful of employees at the infant biotech, leading a team of scientists in DNA synthesis technology. Steven Fung, Applied's senior director of genetic analysis research and development, says it was Efcavitch who brought the synthesis technology from Caruthers' lab to the company. "He championed it here and turned it into a very successful business," says Fung.

For the next several years Efcavitch's team worked to make the availability of synthetic oligos more efficient and inexpensive. In 1987, as Applied Biosystems grew, Efcavitch left the oligo group to develop high-performance electrophoresis chromatography, and later, bioanalytical instrumentation, capillary electrophoresis, and high-performance liquid chromatography.

Part of Efcavitch's success in the early days, says Fung, was his connection to customers. "He would go out and talk to customers along with the sales force," he says. That focus on the end user stayed with him as the company grew. In 1991 Efcavitch became the leader of the DNA sequencing group, which developed the 377 sequencer in 1995, the 310 in 1996, and the 3700 in 1999. Joe Smith, Applied's former senior vice president of business development, says that he remembers during the development of the 3700 sequencer that Efcavitch had the software completely rearchitectured. Efcavitch had commissioned a study of the software's complexity. "He came by my office one day with this complexity chart and said, 'Do you believe this?' He said, 'There's no way to fix it without redoing it.'"

Such decisions to start over, or to challenge current technology and improve it, were not always greeted with enthusiasm, says Smith. "He made a lot of unpopular moves at ABI, but they paid off," he says. "The gel business was one of the major ones." Efcavitch proposed making a replenishable electrophoresis gel that could be used in DNA sequencers, rather than casting a new gel every time the user made a new run. "Nobody thought the gel would work," says Smith, but with Efcavitch's insistence and confidence, the replenishable gel is a technology still in use, as is Efcavitch's ability to push through unpopular ideas.

The High Risk of Intuition

In 1999 Efcavitch moved from research and development to business management, taking responsibility for more of the commercial side of developing microarrays. "It was tremendously interesting, but I realized I didn't have revenue lust," he says. "I enjoyed every single minute at Applied Biosystems, but it was time for me to move on."

He left the company in 2004. Helicos' CEO, Stan Lapidus, contacted Efcavitch to interest him in the company. "He has an extraordinary skill mix," Lapidus says, and Efcavitch's work at Applied Biosystems matched just what Lapidus was looking for in a research and development leader.

"My first response was no, sorry. After 10 years of DNA sequencing I didn't want to get into another sequencer," Efcavitch recalls. Nevertheless, he looked into Helicos' platform, a new type of sequencer based on findings from Stanford professor Steve Quake (Proc Natl Acad Sci, 100:3960-4, 2003). "I was struck by the simplicity of the project," he says.

The HeliScope's technology starts by fragmenting a genomic sample into 100-base pair chunks, which adhere to a surface. Then, the sequencer introduces one of four fluorescent analogs to the strands of DNA. If the analog is the correct counterpart to the next available nucleotide on the strand, it will bind to the DNA template. Using fluorescence microscopy, the HeliScope records which of the four analogs binds to that nucleotide, washes the template of any unused analogs, and starts over on to the next nucleotide down the line.

The technology was bold and challenging enough to stimulate Efcavitch, and he started with Helicos in October 2004. Efcavitch was leading a small team of scientists who quickly replicated Quake's results, producing five-nucleotide read lengths, and they soon extended that to 12 nucleotides. Progress seemed to move quickly.

Then the "dark days" set in. "It seemed like no matter what we would do we wouldn't get past 12 nucleotides," says Efcavitch. The goal was to get to 25 nucleotides, but for months the team was stuck. At one point the company was ready to cut its losses and abandon the project altogether, but Efcavitch's knowledge across chemistry and engineering, and his confidence in whatever technology he's developing, had come through for him before.

Finally, a formulation change jumped the read length to 18. The company decided to go ahead with product development, confident that 25 nucleotides would be reached soon, but again the team stalled.

For months the read length wouldn't budge, and Efcavitch had to rally a tense bunch of frustrated scientists. Not only was the read length at a plateau, but also Efcavitch came to the conclusion that the fluorescent analogs Helicos was using for sequencing essentially needed to be reinvented to ensure accuracy. "When I walked in the door, my intuition was telling me we would have to go this route," says Efcavitch.

The decision was not a popular one; Efcavitch says his colleagues were skeptical about the high-risk project. "They had powerful arguments, I have to admit," he says, but without getting new analogs to work accurately, the sequencer might fail. "There was no question," Lapidus says. "Frankly nothing else worked." Efcavitch had worked with organic chemists at Applied Biosystems, and he had a fairly good idea that compatible analogs could be designed.

To make sure, he turned to members of Helicos' "rock star" scientific advisory board, which includes Steven Chu, the director of Lawrence Berkeley National Laboratory, Harvard genetics professor George Church, and Eugene Myers at the Howard Hughes Medical Institute. Lapidus says his experience in building companies taught him the importance of a strong advisory board, especially to support the ambitious goals of Helicos' technology. When starting the company, Lapidus, Quake, and investor Flagship Ventures appealed to people they knew and had worked with in the past to recruit "the best people with relevant experience," says Lapidus.

Efcavitch says board member David Liu, a professor of chemistry at Harvard University, confirmed the feasibility of the analogs project. Lapidus and the rest of the scientific advisory board backed Efcavitch, and he hired a team of organic chemists to make the new analogs. "The thing I like about Stan and the board of directors is they encourage a Plan B type of mentality," says Efcavitch. "Once the concept was confirmed by the scientific advisory board members, that gave it credibility and we took the risk."

Galvanizing the Troops

Taking that risk finally broke the logjam in 2006. One day at the end of that summer, a Helicos scientist noticed a change in her data after altering a step during the sequencing process. It seemed like it might be just the change that could extend the nucleotide read length. Efcavitch turned the group in that direction, and soon the 25-nucleotide read length was reached. Shortly following, the organic chemists presented the newly designed analogs to the molecular biologists, and the entire single-molecule sequencing technology came together.

The company forged ahead with optimizing the HeliScope and getting it ready for market. In early November 2007, the company announced the price for the sequencer: $1.3 million, and $18,000 for reagents. To get down to the $1,000 price per genome will take about two years of optimizing the chemistry, says Efcavitch, who is pushing hard to get there. Patrice Milos, the chief scientific officer of Helicos, says, "I think that he leads his group with a real sense of urgency. I see that in every meeting - that the time is short and we have a lot to do. He does a lot to galvanize the troops."

Efcavitch is just as confident about the market for the HeliScope as he is about the technology, which is the basis of everything Helicos does. Stock analysts don't appear as confident, however. In November 2007 Helicos' stock dropped from $14 per share to less than $10, after an analyst downgraded it to "sell." Since the company went public in May 2007 the stock has gone as high as $15. As of late November, the market capitalization was roughly $209 million.

The analyst who downgraded Helicos' stock cited competition from other companies, including, perhaps not surprisingly, Applied Biosystems. But while other instruments require amplification or multiple sequencers, the HeliScope is designed to do it all with one step. "Some say, gee, you guys are late, meaning we're not the first out with genome sequencing technology. But I don't think we're too late," Efcavitch says. "It's not about the existing two billion dollar sequencing market, but about the next generation."