Pharmasset is in the business of saving people’s lives.
The pharmaceutical company is racing to develop medicines to treat Hepatitis B, Hepatitis C and HIV and has already made significant inroads. The 10-year-old company in Princeton, N.J., has developed three drugs that are in clinical trials, including a promising Hepatitis B drug in the advanced stages of testing that one financial analyst believes has the potential to become one of the more potent therapies to combat the virus. Pharmasset has also partnered with Swiss drug giant Roche to develop and test a Hepatitis C drug, while its HIV drug is in the second of three phases of testing.
Even with its promising efforts, Pharmasset continues to invest in research and development to find new ways to combat the three viruses, and technology plays an important role. Last fall, the company began pursuing computational chemistry research and purchased four Hewlett-Packard blade servers  to run molecular simulations aimed at finding new chemical compounds that stop the viruses from replicating in the human body.
Historically, researchers have performed antiviral research by testing chemical compounds in laboratories, but that becomes problematic when the tests number in the thousands or millions. It is faster and more economical to use servers to run computer simulations to help find the compounds that are likely to be effective against the viruses, says Ralph Mosley, Pharmasset’s head of computational chemistry. Mosley analyzes the results from millions of simulations, finds the promising compounds and forwards them to the company’s biologists and chemists for further research in their labs.
“The drug discovery process is not fast or simple. It can take 10 years from initial research to the final drug being released, so if I can narrow down a million chemical structures to 1,000 or less, it saves us time and resources,” Mosley says.
Small upstart biotech firms like Pharmasset compete against larger entrenched players, such as Pfizer, Merck and Genentech, for a piece of the exploding biotechnology market, a field that Ernst & Young estimates reached $80 billion in worldwide revenue in 2007.
The potential market for Pharmasset’s planned product portfolio is huge. According to the World Health Organization, 40 million people worldwide have HIV, 350 million people have chronic Hepatitis B infection, and another 180 million are infected with Hepatitis C. People with the two types of hepatitis have an increased risk of developing liver diseases, including liver cancer. In fact, about 5,000 Americans die each year from liver disease related to Hepatitis B, according to the Centers for Disease Control.
Pharmasset, which raised $45 million in an initial public offering in 2007, doubled in size in the past year, from about 35 to 70 employees, and opened a satellite office in Durham, N.C. The company’s growth keeps IT director Jim Anuth busy. His two-person IT team handles everything from help-desk support and configuring computers for new employees to managing the data center and network security.
Pharmasset’s biggest technology investment this past year is its new computational chemistry push, an initiative that’s very critical to the company’s success because it guides future research, Anuth says. Pharmasset established the new technology effort last fall by hiring Mosley and purchasing four HP ProLiant c-Class Server Blades and a BladeSystem c3000 enclosure.
Blade servers are slim servers that feature their own processor, memory and storage, but they slide into a chassis where they share resources, such as power supplies, cooling fans and connections to networking and data storage systems. Blades conserve data center space and are simpler to manage because they reduce cabling and because new software management tools let IT administrators centrally manage and monitor the health of each blade server.
Anuth chose blade servers because of their space-saving design. He opted for HP because the company was the only manufacturer that offered a quad-core, quad-processor architecture at the time he was looking. Other manufacturers offered only dual-core quad processors.
“If we had gone with one of the other vendor’s solutions, it would have required twice as much hardware as well as double the software licenses to get the same performance that the HP solution offered,” Anuth explains.
The company purchased two molecular modeling applications to manage and run the simulations. Each application is licensed per processor, not per core. As a result, more cores on a processor equals better performance for the same software license costs, Anuth says. The company saved $90,000 in software licenses and a similar amount in hardware by choosing the denser cores offered by the HP blades, he explains.
The blade server’s small form factor also requires less server room. Pharmasset’s 600-square-foot server room was tightly packed with 12 rack-mounted servers, and there was just enough space to fit the blades, he says.
The blade servers were fairly easy to install, Anuth says. With the help of HP BladeSystem Onboard Administrator web-based management software, Anuth configured the blade enclosure and the blade servers and connected them to Pharmasset’s storage area network in just 90 minutes. Anuth purchased blade servers with no built-in hard drives; he boots off the SAN through a 4-gigabyte Fibre Channel network. The management software gave him one central view of the servers, network and SAN, making configuration easy.
Pharmasset has access to 250,000 research molecules through a partnership with the University of Cincinnati’s Genome Research Institute.
Using an approach known as “virtual screening,” Mosley relies on a high-powered, Linux workstation to submit jobs to the four blade servers.
The molecular modeling applications convert the 2D drawings of molecules into 3D entities that have volume and characteristics such as positive and negative charges and “greasiness.” The applications running on the servers look for compounds that have a shape similar to an enzyme’s binding cavity and the right characteristics to stick together. For example, positive will attract negative and grease attracts grease. Blocking an enzyme in this way might provide the means to inhibit viral replication.
In addition to virtual screening, Pharmasset conducts research called “high throughput screening.” In this approach, University of Cincinnati and Pharmasset researchers perform real biological testing and send their data to Microsoft Excel tables. Pharmasset then takes the data and analyzes it on its blade servers.
Investing in the technology to pursue computational chemistry is critical because it lets Pharmasset compete more effectively against other biotech firms, says Michael Sofia, Pharmasset’s vice president of chemistry.
“We need to assess the data and pick the most promising compounds for another round of intensive testing,” explains Sofia. “To accomplish this, we need a significant amount of computing horsepower.”
Pharmasset relies on technology throughout its organization. Chemists, biologists and research assistants perform lab experiments using instruments directly connected to standard PCs, so results are immediately transferred to databases on the company’s SAN. The data is stored in a central location so researchers can analyze it and share it with colleagues.
To better protect the company’s intellectual property, Anuth deployed two-factor authentication for remote users. Teleworkers and road warriors have been given key fobs that provide randomly generated authentication codes that change every minute. The users type in the code to access the company’s VPN. For disaster recovery, the company backs up data to disk and then spools it to a 14-tape library.
Anuth, who joined the company in 2005, has new tech projects in the planning stages, including a move to Voice over IP and server virtualization. He hopes to consolidate his 12 e-mail, file-and-print and database servers down to three or four physical servers. As the company grows and needs new applications, virtualization will let Anuth quickly deploy new applications and meet the company’s needs without having to purchase additional servers or build out a larger data center.
If the Hepatitis B drug’s clinical trials end by 2009 with good results, and the FDA approves the drug by 2010, then the company can release the product in late 2010 or early 2011, says Adam Cutler, managing director of life sciences at financial services firm Canaccord Adams in New York.
In fact, in a recent report, Thomas Russo, senior analyst with financial management firm Robert W. Baird, wrote that Pharmasset’s Hepatitis B drug was a “potential game changer” in the industry and that tests to date show “that it can take a place among the most potent agents on the market.”
“Earlier-stage drug companies need to strike partnerships with larger companies to help them spread some of the risk,” Cutler explains. “The trick is to spread risk, validate their R&D success by partnering with a larger company, but not give away the farm. Pharmasset has done that with a nice deal with Roche.”