Skip to main content
Premium Trial:

Request an Annual Quote

March 2001: Happy Days at the HAP Factory


Whenever his crew achieves a production goal, taskmaster Jerry Vovis yanks the braided rope of a marine bell that hangs outside Genaissance Pharmaceuticals' SNP discovery facility in New Haven, Conn.

Since becoming the pharmacogenomics firm's chief technology officer two years ago, Vovis, 58, has been running a tight ship, navigating the company in uncharted waters.

Upon his arrival from Genome Therapeutics, where Vovis was senior VP of scientific affairs, Genaissance declared that it would find tens of thousands of nucleotide variations in 2,000 genes from 95 individuals and then organize them into so many predictive markers by the end of 2000.

"There were people who thought, 'This is crazy. What are we talking about?'" says Vovis, a self-professed people watcher whose friendly, engaging style makes him especially suited to the job of eking out subtle differences among people.

At the time, Genaissance had found markers for only a few genes. But with a fleet of 59 ABI Prism 3700 sequencing machines — the largest commercial installation outside Celera — Vovis and team did the deed with three months to spare. By the end of last year Genaissance had amassed data for 3,005 genes — 50 percent more than it had promised its investors.

Within these genes, Genaissance has identified more than 53,000 single-nucleotide polymorphisms and organized them into more than 56,000 haplotype markers.

The Genaissance gamble is that these markers will reveal why people respond differently to drugs. And that drug makers will care.


Genaissance headquarters bears no trace of its previous tenant, the Winchester Repeating Arms factory. Lead has been removed from the walls, and modern curves and angles carve out a bright workspace for the company's 150-plus employees. Still, operations in what is now known as the HAP Factory retain an air of old-fashioned manufacturing.

In the age of high-throughput genomics, Genaissance has turned discovering haplotypes into an industrial process. It starts with the target genes of a group of 93 ethnically diverse humans, including families, plus a chimpanzee and a gorilla, and systematically sequences them. To track down pharmaceutically relevant genes, bioinformaticists harness expression data from Gene Logic's database. Then, to home in on the target genomic DNA, they run mRNAs against GenBank on a parallelized version of BLAST that the company bought recently from its New Haven neighbor, TurboGenomics.

Discoveries of haplotype markers — called HAP markers here — are graded against precise daily and quarterly goals, and employees are rewarded for meeting them. Each clang of Vovis's marine bell, greeted by cheers throughout the building, signals another milestone met and a good excuse for a celebratory break.

Considering the endgame, it's easy to see what all the fuss is about. With its arsenal assembled, Genaissance expects to revolutionize medicine.

To be sure, Genaissance is far from the only firm with faith in the power of SNPs. Celera Genomics, Incyte Genomics, and CuraGen are among those hoping to sell SNP data. And population genomics companies DeCode, Gemini, and Myriad are counting on linking specific SNPs to disease.

But Genaissance says those approaches are misguided. Single nucleotides don't offer enough information to predict how a patient will respond to a drug, Vovis and company contend. Instead, they say, the clues to metabolism are in SNP combinations, or haplotypes.

Fresh off the redeye from San Francisco where he has just pitched this story to investors at a healthcare conference, sharply-dressed CEO Gualberto Ruaño declares, "It's the coupling power of the SNPs."

In other words, using a SNP to predict drug response is like typing a single character into a search engine to find a website. Chances are, it won't turn up what you're looking for. But a haplotype is like a few carefully selected keywords.

"That's what the haplotype really does — it elevates SNPs to an entirely new level of power because it looks at patterns between [them]," Ruaño says. Haplotypes, he contends, have the strength to shrink from thousands to hundreds the number of patients needed to statistically validate a correlation between genomic variation and drug response.

If Vovis is the company's taskmaster, then Ruaño, who with president and CFO Kevin Rakin founded Genaissance in 1997, is its visionary. It was while an MD and PhD student in Human Genetics at Yale that Ruaño says he was struck by the wide gap between genetics and primary health care. "Nobody in medicine was using genetics on a routine basis," he says. And though he admits that he never imagined himself running a publicly traded company, Ruaño says, "That's why I went into the commercial sector.

The in silico argument

Genaissance considers haplotype building such a key part of its marketing message that spokesperson Anita Kaul insists, "We are not a SNP company."

But in truth Genaissance is little more than a SNP company. The HAP Factory, despite its name, actually discovers SNPs. Its HAPtyping facility, too, simply scores SNPs in clinical samples. Haplotypes are calculated in silico, algorithmically generated, a product of bioinformatics.

That's what Taylor Crouch, CEO of Variagenics — Genaissance's closest competitor — points to as his rival's weakest link. "Genaissance has a sort of bioinformatics-driven business model," says Crouch. In contrast, Variagenics has developed technology that experimentally measures haplotypes from the cell tray definitively.

Ultimately during clinical trials for the FDA, you're going to need an experimentally verifiable result," Crouch predicts. "Either Genaissance will find a way to get a hold of experimental haplotype technology or run into some challenges on the ground."

But Vovis counters that comparing results to molecular cloning and conventional family analysis has validated the accuracy of Genaissance's algorithms. In both cases the so-called DecoGen algorithms matched 97 percent of the haplotypes. With the error rate inherent in the traditional methods, that difference is insignificant, Vovis says.

Also, Variagenics' molecular genotyping technology is currently limited to haplotypes that span fewer than six kilobases — yet both companies agree that the markers often stretch 10 times that distance.

Besides, Vovis says, molecular methods are too slow and too resistant to high throughput for his demanding production schedule. "We're capable of looking at about 120 genes per week," he says, proudly pointing to the $18 million worth of sequencing machines that run nonstop to transduce molecular differences into electronic messages capable of predicting drug response.

At its current pace, Genaissance is poised to systematically identify haplotypes for nearly 10,000 genes — more than a quarter of the genome by some accounts — by the end of this year.

To ensure accuracy, Vovis plans to enlist half of the factory's sequencing capacity for diving deeper into last year's genes. "This procedure is as highly accurate and it's very rapid," he says. "So the question is, why would one want to do something else?"

Will big pharma bite?

Sparring aside, suppose Vovis and Ruaño are right. Suppose their haplotypes perfectly predict patient responses for every drug on the market. They still won't be on easy street. Now they have to sell big pharma on their vision, which, by virtue of dissecting the population into "HAP" categories, dampens the future for blockbuster drugs.

Ruaño uses a Hollywood analogy to try to shift pharmaceutical execs' thinking: "How many Titanics have we had?" Pharmaceutical companies would be better off going the way of Miramax—exploit a particular niche and develop a loyal audience, he says.

Besides, Ruaño adds, too may company valuations depend on a single drug. "The Street has been very tough on Merck because it has three blockbusters that are going to expire," he says. "In other words, you depend on the blockbusters — but once they go away you're in deep trouble."

But pharma's resistance to pharmacogenomics goes deeper than a fear of stratified patient populations. Personalized medicine would demand a complete overhaul of the industry's fundamental approach to drug discovery and development.

"The whole process is designed around chemistry and screening, not around genomics and analyzing genetic information," says Michael Liebman, who, fed up with its snail's-pace adoption of genomics technology, recently dumped a pharma career to direct University of Pennsylvania Medical School's computational biology department. "They're chemistry based and that's what their focus is," he says of drug companies. "And they are extremely resistant to change."

Genaissance knows what it's up against. "Every time you have a disruptive technology the first reaction of the entrenched players is, ignore it, it will go away, don't worry about it," says Ruaño. "It's up to companies such as Genaissance to keep pushing the envelope through examples, demonstrations, clinical views. This is how you change the operating system."

If Liebman is right, resistance could ease sooner rather than later. As pharmacogenomic technology becomes cheap and accurate, "the FDA is going to make companies use it," he predicts. "You won't be able to design a drug without it."

Clinical trials and tribulations

To prove its value to pharmaceutical companies, Genaissance applied its data to a drug that one of its desired customers already markets. In a study published last September in Proceedings of the National Academy of Sciences, Genaissance and Stephen Liggett of the University of Cincinnati Medical Center conducted a clinical study comparing the value of HAP markers and single SNPs in predicting an asthmatic patient's response to albuterol, a drug sold by Glaxo- SmithKline as Ventolin.

Genaissance found 13 SNPs in the ß2 adrenergic gene, which codes for the drug's target. Theoretically, these SNPs could be combined in 213, or 8,192, haplotypes. But using its proprietary bioinformatics system, Genaissance showed that only 12 combinations actually exist it in nature. The company calls these, the final products of its HAP Factory, HAP markers and stores them in its Isogenomics database.

The researchers further narrowed the markers down to the four most common haplotypes. Liggett then measured responses of 121 asthmatic patients to Ventolin. Genaissance genotyped each patient with its Sequenom MassArrays, a process it calls HAPtyping, and reported a direct correlation between a patient's response to the drug and specific haplotype markers, but not to individual SNPs. The company has filed a patent for this correlation.

Despite its pharmaceuticals moniker, Genaissance has no plans to develop its own drugs. Instead it will piggyback its IP to existing drugs or drugs already in clinical trial.

For instance, in its first deal, secured last November, Genaissance will help Johnson & Johnson companies Janssen Research Foundation and R.W. Johnson Pharmaceutical Research Institute generate pharmacogenomic data for a drug in the pipeline. J&J gets what Genaissance plans to offer all such customers: access to HAP markers for several pre-selected genes each year and the DecoGen informatics tool that algorithmically links clinical drug response to the markers.

Janssen would not say for what drug it would use the HAP technology, but analyst Dennis Harp of Deutsche Banc Alex. Brown surmises "the program will be directed in the neurology, anti-psychotics area, because that's a big area of focus within Janssen."

Though the terms of the deal weren't disclosed, Genaissance CFO Rakin says they were "in line with our expectations." Following the company's August 2000 IPO, he told investors that three-year deals would be signed at annual rates of $3 million to $5 million.

Genaissance is mum about other deals, only promising to announce one by the end of 2001. But according to Solomon Smith Barney, Pfizer is beta-testing Genaissance technology at its Groton, Conn., and Sandwich, UK, sites.

Bellwether for new medicine

Rakin says the company will increase its spending while maintaining a $40-million-a-year burn rate. Without additional deals, its $120 million of reserves, including $90 million from its IPO, will evaporate within three years.

But Genaissance is so confident its data will be in high demand that it plans to charge future partners $10 million a year. On top of that, its contracts allow Genaissance to retain rights to all HAP markers discovered jointly.

Big payoffs, however, will not come from pharma partnerships, but from data the company generates within, Rakin says. Immediately following its IPO, Genaissance enlisted Ken Kashkin — Ruaño's former professor at Yale medical school who has since worked for Knoll, Abbott, and Bayer — as chief medical officer to conduct clinical trials on the top 20 selling drugs.

So far Genaissance's Mednostics program has amassed HAP marker-linked response data on the cholesterol-lowering statins: Bristol-Myers Squibb's Pravacol, Pfizer's Lipitor, and Bayer's Baycol — altogether a $12 billion market. The company plans to auction the data to the highest bidder.

Later this year Genaissance will also begin collecting data on the top-selling asthma, diabetes, schizophrenia, and obesity drugs. "I can go to a company and tell them, look, I have here a picture of the people who have the highest response," says Ruaño. "This is your niche, and you should use this to displace competition. In this niche you are the king."

According to Salomon Smith Barney, Genaissance expects its first Mednostics customer late this year to cough up $7.5 million.

Ruaño lists a number of other ways he intends to leverage the IP. For a pharmaceutical company to use these data to conquer a targeted marketing kingdom it must be able to identify its subjects. And diagnostics have the potential to diverge into two revenue streams for Genaissance: "One, when a test is done," says Vovis. "And, if our markers are associated with the drug, we can also [get] royalties on the prescription of the drug."

In addition, Vovis and Ruaño believe the managed care establishment would love to get hold of data that point to who is likely or unlikely to benefit from a particular drug. Ruaño envisions patients carrying haplotype-encoded cards. Before prescribing a particular drug, a physician would slide the card through an ATM-like machine to read the patient's likely reaction. Genaissance sees itself as the owner of the data that will drive this new way of practicing medicine.

Signal Amplification

As he exits the HAP Factory, Vovis stops and admires the bell. When he hung it here in the fall of 1999 Genaissance's entire operation was on one floor of one wing of the building. It now seems like a quaint symbol of the old days when the mechanical clanging did not have to be broadcast electronically. "Since we're now occupying three floors and two wings, not only is the bell rung," says Vovis, "we send out an electronic message."

Sidebar: Identifying Polymorphisms in a Business Plan

Both Genaissance and Variagenics are pitching pharmacogenomics as a tool for smarter clinical trials, but their business plans differ on a single central point. Genaissance focuses on hawking information, Variagenics on a proprietary genotyping technology

True, Variagenics is also generating IP — it has applied for SNP patents on more than 5,000 genes. But it points to the distribution of this technology through CROs and diagnostic kits as its main strength. "You've got to have interesting technology in the center," says CEO Taylor Crouch.

On the other hand, Genaissance insists that not being married to a particular technology is to its advantage. "We're not in the picks and shovel business," says CEO Gualberto Ruaño. "We're in the information business." His philosophy is "explore the market for the best picks and shovels, and buy it."





DNA Samples

Index Repository
Proprietary collection of cell lines from more than 350 unrelated individuals of diverse backgrounds, and members of extended families for haplotype validation

Has purchased more than 100 cell lines from unrelated individuals and some families from the Coriell Institute in New Jersey

SNP Discovery

59 Applied Biosystems ABI 3700 Sequencers

"A bank of ABI 3700s,” number not specified

SNP Scoring

Uses Sequenom MassArray to score SNPs;does not measure haplotypes directly

Proprietary technology that uses mass spectrometry to measure DNA fragments produced by chemical cleavage at points where it has inserted synthesized nucleotide analogs. Can score both strands at the same time

Compiling Haplotypes

DecoGen Informatics System
Uses population data from Index Repository and proprietary algorithms to build haplotypes
Correlates clinical drug response to haplotype markers

Uses algorithms to model haplotypes for use in clinical trials. But directly measures haplotypes, limited to 6 kb, in clinical samples with proprietary NuCleave technology.


As of the end of Dec. 2000 contains: 53,677 SNPs and 56,426 haplotypes representing 3,005 genes

As of the end of Nov. 2000 contains: 20,000 SNPs and 7,000 haplotypes representing 2,000 genes

Pharma Partners



HAP2000 Program
J&J's Janssen Research Foundation and R.W. Johnson Pharmaceutical Research Institute
Three-year partnerships for annual access to HAP markers of selected genes with annual subscription rates of $3 million — $5 million for first few partners. $10 million annually for subsequent partners. Royalties: 2 percent — 4 percent. Genaissance keeps markers.

Variagenics Impact Program (VIP)
Boehringer Ingelheim, Amgen $1 million — $3 million in fees, possibly single-digit royalties, for applying technology in Phase 2 and Phase 3 trials

Other business developments

Internal clinical trials to generate response profiles for top-selling drugs on market
Gene Logic
Subscribes to gene expression database to prioritize gene targets for SNP discovery
Licensed SNP data to Gene Logic

Exclusive distribution deal with CRO for outsourced genotyping for clinical trials
Non-exclusive co-marketing deal with this CRO
Development and marketing of NuCleave kits to be used with Bruker Daltonic mass specs

The Scan

Genome Sequences Reveal Range Mutations in Induced Pluripotent Stem Cells

Researchers in Nature Genetics detect somatic mutation variation across iPSCs generated from blood or skin fibroblast cell sources, along with selection for BCOR gene mutations.

Researchers Reprogram Plant Roots With Synthetic Genetic Circuit Strategy

Root gene expression was altered with the help of genetic circuits built around a series of synthetic transcriptional regulators in the Nicotiana benthamiana plant in a Science paper.

Infectious Disease Tracking Study Compares Genome Sequencing Approaches

Researchers in BMC Genomics see advantages for capture-based Illumina sequencing and amplicon-based sequencing on the Nanopore instrument, depending on the situation or samples available.

LINE-1 Linked to Premature Aging Conditions

Researchers report in Science Translational Medicine that the accumulation of LINE-1 RNA contributes to premature aging conditions and that symptoms can be improved by targeting them.