When Roche paid $46.8 billion in 2009 to gain full ownership of Genentech, fears were rife that the arrival of the Swiss men in suits would stifle the free-wheeling innovative culture at the California-based biotech.

Three years on, however, the jeans-wearing scientists in San Francisco have proven they are the ones driving the drugs pipeline of the 116-year-old Basel-based pharma giant, casting a shadow over Roche's own research operations.

Roche's four top-selling medicines in 2011 - Rituxan, Avastin, Herceptin and Lucentis - were all cooked up in Genentech's labs and accounted for 55% of total pharma sales last year.

And five of the seven most promising drugs in Roche's pipeline, identified by brokerage Bernstein, stem from Genentech, including "armed antibody" experimental anti-cancer drug TDM-1, which received top-billing at this year's meeting of the American Society of Clinical Oncology.

A decision last week to close Roche's 80-year-old Nutley, New Jersey research facility, which gave the world Valium, underscores how Genentech's Research and Early Development operations, known as gRED, are outshining Roche's own unit called pRED (Roche Pharma Research and Early Development).

After culling 1,000 jobs at Nutley, both gRED and pRED will have around 2,000 staff each, a downsizing of the legacy Roche wing that is prompting veteran Roche employees to fear that a stealthy Genentech reverse takeover is underway.

"In essence, the merger of Roche and Genentech was about Genentech going global," said Andrew Weiss, analyst at Swiss private bank Vontobel. "At the time they probably had the upper hand and I think that hasn't changed."

Nail in coffin

Genentech's approach of unlocking the biological pathway of a disease to find a specific target for its drugs is paying dividends, compared to the old-school method of developing a drug before the biology is truly understood.

While the biotech notches up one success after another, the old Roche labs have been hit by a litany of high-profile and costly failures.

Most recently the Swiss company halted clinical trials of experimental heart drug dalcetrapib - one of the several new medicines aimed at boosting levels of beneficial cholesterol - a failure which perhaps was the final nail in Nutley's coffin.

Roche is also now starting to adopt the Genentech approach by trying to understand the chain of events that cause disease.

The closure of Roche's former U.S. headquarters comes on top of a 2010 programme which aimed to slash 6 percent of its workforce worldwide or 4,800 jobs over two years. In this cull, gRED got off more lightly than pRED, which shed 600 positions.

Manuel Wyss from trade union Unia says employees in Switzerland see no great threat at present as the firm already slashed 770 positions in the country under the 2010 programme.

The family of Fritz Hoffmann-La Roche, who founded the firm in Basel in 1896, still holds a big stake. The company is splashing out 550 million Swiss francs ($576 mln) on a 178-metre-tall office tower that will dominate the skyline.

Schwan has said Roche is committed to Switzerland despite the impact of the strong safe-haven Swiss franc on costs in recent years, but the additional axe-wielding at Nutley raises questions as to whether more cutbacks are on the horizon.

"The announcement definitely indicates that Roche's margins are more under pressure than previously thought," said Birgit Kulhoff, analyst at private bank Rahn & Bodmer in Zurich.

Genentech vs Roche?

The Nutley cull also ensnared Basel-based Jean-Jacques Garaud, Roche's French/American head of pRED, who admitted in an interview last year that researchers from the different operations work very separately.

"We talk to each other, but we don't talk about science too much. We talk about life - books and movies," he told Reuters last September.

Keeping the two research units autonomous was a concession to Genentech which feared the merger would damage innovation and send top scientists and managers heading for the door.

As Roche sought to redefine itself after the Genentech deal, it quit the U.S. industry lobby group Pharmaceutical Research and Manufacturers of America (PhRMA) in 2009 to join the biotechnology association BIO.

It also markets its drugs in the United States under the Genentech brand and Roche's Chief Medical Officer, Hal Barron, comes from the San-Francisco firm.

Genentech staff seem happy with the set-up, stressing at a recent meeting in New York that Roche respected the biotech's culture as well as its science.

Jeans still prevail over formal business attire, while Roche staff who moved to San Francisco when the U.S. headquarters were shifted there in 2009 have integrated well, they say.

It hasn't been all smooth sailing at Genentech since 2009.

The biotech company suffered a blow last year when the U.S. Food and Drug Administration revoked approval of its blockbuster drug Avastin as a treatment for breast cancer. The drug also failed in clinical trials for stomach and prostate cancer.

But Vontobel's Weiss said Genentech's top-notch reputation still helps attract workers.

"Everybody wants to go there and write a paper and put their name on it and Genentech. If you are lucky to find something that is where it's potentially going to happen."

While Roche argues maintaining two research units gives researchers the freedom to tackle problems from different angles, it inevitably pits the two against one another.

"The discovery track record of pRED over the past decade has not been exactly overwhelming, particularly in comparison to its counterpart in south San Francisco," said Fabian Wenner, head of European Healthcare Research at Kepler Capital Markets.

"The question is whether this can be changed or whether something more fundamental needs to be done?"