August 2, 2011 | Six years after publishing details of the first commercially available next-generation sequencing (NGS) system, by 454 Life Sciences, Jonathan Rothberg and his colleagues at Ion Torrent have published the first results from a new desktop NGS technology today, also in Nature.
All 44 co-authors are (or were) employees of Ion Torrent or its parent company, Life Technologies, including Kevin McKernan, one of the architect’s of Life Technologies second-generation SOLiD platform, who recently left the company.
The new paper includes an overview of the sequence (at tenfold coverage) of Gordon Moore, the co-founder of Intel and the author of the famous Moore’s Law concerning the growth of compute processing capacity.
Meanwhile, charges continue to fly over the origins of some of the key technology that Ion Torrent licensed from Stanford University’s Office of Technology Licensing (OTL). Recently, two scientists, Stanford’s Ron Davis and his former colleague, Nader Pourmand (University of California, Santa Cruz), complained publicly that Stanford OTL undervalued their technology during negotiations of an exclusive license to Ion Torrent. Now those complaints have sparked a strong response by another former Stanford colleague, Arjang Hassibi.
In an exclusive interview with Bio•IT World, Hassibi, now an assistant professor at the University of Texas in Austin, says he and others played a key role in the development of “charge sequencing” technology. In 2001, Hassibi co-founded a biotech company called Xagros Genomics with Pourmand. But the two men fell out over the demise of the company and the attribution of credit—or lack thereof—for results published in an important paper co-authored by Pourmand and Davis in 2006 (and cited in the new Ion Torrent Nature paper).
Those feelings were resurrected in the past month after Pourmand turned to Bio•IT World and other media to express his frustration with the Stanford-Ion Torrent licensing deal.
“It is very surprising to me that Nader is claiming right now that ‘Hydrogen generation [during DNA sequencing] is my patent, my invention,’” said Hassibi, repeating a quote from Pourmand in a Bio•IT World story. “They want to erase any memory of what happened before that. This I have a problem with.”
They Did Great
For Hassibi, it is a matter of principle and seeking fair credit for past intellectual contributions. Like Davis and Pourmand, he is receiving licensing fees from Ion Torrent, albeit less than the paltry $2,300 that irked Pourmand. “Let’s be clear: Ion Torrent hasn’t done anything wrong. Stanford hasn’t done anything wrong. Some scientists came up with a good idea. We [Xagros] failed. Another company—Ion Torrent—picked it up and they did great.”
The issue of credit for what Ion Torrent calls semiconductor sequencing—and potentially financial compensation down the road—has become more acute following Ion Torrent’s acquisition by Life Technologies in 2010 for $375 million (potentially rising to $725 million).
Among dozens of patents it has in-licensed, Ion Torrent acquired exclusive licenses to two related Stanford patents. The first—Stanford docket S00-157 “Charge Sequencing: A New Technique for DNA Sequencing and SNP Detection” (priority date October 2001)—lists Hassibi and Pourmand as the inventors. The second—docket S04-291 “Charge Alternation DNA Detection System” (priority date November 2004)—has three inventors: Pourmand, Davis, and Miloslav Karhanek.
“Based on the rule of ‘Success has many fathers, failure is an orphan,’ there will be many ‘fathers’ for this technology, including Nader,” said Hassibi. “However, he is completely distorting the story right now by claiming all the credit for himself. I believe this to be neither ethical nor constructive if [as he claimed] he wants to improve Stanford OTL’s licensing processes.”
The sequencing squabble dates back to 2000, when Hassibi, then a Ph.D. student in electrical engineering at Stanford University, first met Pourmand, who was a postdoc at the Stanford Genome Technology Center (SGTC). Pourmand and Mostafa Ronaghi (now Illumina’s chief technology officer) were working with Davis on pyrosequencing, the sequencing technology that underlies the Roche/454 next-gen sequencing platform.
“I remember clearly, one afternoon in the spring of 2000, talking with Nader in the second floor of Stanford Center for Integrated Systems (CIS). He asked me whether we can detect any electrical signal in the DNA structure and if yes, whether I could build the electronics for it,” Hassibi recalled.
Pourmand argued that DNA must have an associated charge, because it migrates during electrophoresis (the basis of Sanger sequencing). The next time the two met, Hassibi proposed to place some DNA near an electrode and connect it to a high-impedance voltage amplifier to see if a length difference resulted in a different charge (or voltage) signature.
Hassibi designed a setup in the IC design lab of his then advisor, Thomas Lee, using iron needles as electrodes and micro-titer plates, while Pourmand prepared magnetic beads with primed DNA attached to them. “We placed a small refrigerator magnet on top one of the needles to immobilize the beads (and DNA) on one electrode. We placed the electrodes in the polymerization buffer and added dNTP. What we saw wasn’t conclusive, but there were distinct fluctuations when polymerization was supposed to happen,” said Hassibi.
Hassibi and Pourmand, who both hail from Iran, began a close collaboration. Based on those early data, they filed a provisional patent in October 2001, which became a full patent the following year.
“We initially called the technology ‘charge-sequencing,’ inspired by pyrosequencing,” said Hassibi. [The Stanford docket S00-157 and provisional patent both had this title.] But Hassibi argued that “what we are seeing is not the DNA charge but essentially some perturbation in the charge equilibrium near the DNA (and near the electrode).” This, he says, is why the title of Hassibi and Pourmand’s US patent 7,223,540 is “Transient electrical signal based methods and devices for characterizing molecular interaction and/or motion in a sample.” Hassibi later called the technology Charge Perturbation Signature (CPS).
The core of Hassibi and Pourmand’s original patent, said Hassibi, is this: “If polymerization happens near an electrode, you see an explosion of ions. ‘Ion Torrent’ is a perfect name for a company commercializing this specific technology, although I am not sure if they named it because of this. Now, to simplify it, one can say it is pH. Initially, when we were marketing Xagros, we said, ‘It is negative charge.’ But the explanation is more complicated than that. These ions move, you have diffusive processes, then you detect it if they get to an electrode.”
In 2001, Hassibi and Pourmand decided to jump on the biotech start-up bandwagon and form a company called Xagros Genomics (named after the Zagros mountains in Iran, but with the obligatory ‘X’ instead). Xagros exclusively licensed docket S00-157 from Stanford and got funding from Tempo Ventures.
Hassibi’s task was to create the sequencing hardware, sensor and the semiconductor chip, while Pourmand was in charge of assay development. A strong advisory board included Davis, Lee, and Berkeley’s Richard Mathies.
Hassibi said they obtained funding because “CPS was semiconductor-compatible... the story was music to the investors’ ears.” Pourmand, a family man, remained affiliated with Stanford, but Hassibi, still a graduate student, opted to join Xagros full-time. Sia Ghazvini joined the company from Combimetrix as CEO, with Dean Hafeman, a founding scientist at Molecular Devices, becoming chief technologist.
According to Hassibi, as the assay development lagged other aspects of the technology, management decided to put Hafeman in charge of that project. “Nader first agreed and the project got on track partially and we all were very hopeful that we would pass this speed bump,” says Hassibi.
One day, however, Hassibi came to work to find that Pourmand had cleared out his desk. “He said that due to health reasons, he could not work in a start-up anymore,” Hassibi said. Hassibi and Pourmand later met in person at Stanford, but Pourmand was unhappy with the way Xagros was operating, and had decided to rejoin Davis’ group at Stanford. Pourmand encouraged Hassibi to do the same, which infuriated Hassibi.
“I told him I had quit my Ph.D. and put 2.5 years of my life without back-up plans or getting any academic credit—and now he wants me to come back? I also mentioned that the investors relied on us two and had put serious money into Xagros and we were responsible for the other employees.” The two agreed to try to co-exist. “Our last handshake was that if we ever decide to publish this work, we would do it together,” said Hassibi. But that did not happen.
Hassibi said Pourmand refused to hand over government-funded projects on which he was listed as the PI, or to negotiate the status of his outstanding shares. A second round of company financing fell apart, and Xagros finally went out of business in 2004. Ultimately, the Stanford patents went back to Stanford, while other patents were abandoned (some related to CMOS chips), as there was no money to support them.
Hassibi eventually returned to Stanford to get his Ph.D. (designing CMOS chips for biosensing and sequencing), but had no contact with Pourmand and Davis. It was during his last semester in 2006 that a colleague showed him a paper by Pourmand and colleagues in the Proceedings of the National Academy of Sciences, contributed by Davis, entitled “Direct electrical detection of DNA synthesis.”
“I was appalled and could not believe what I was reading in the paper,” said Hassibi. “There was no reference to me or Xagros—that pissed me off. Much of the data and the methods were developed at Xagros by me and others and [Pourmand] simply published it without acknowledging any of it. It seems that they came up with technology by themselves and they are trying to erase [our contribution].”
Hassibi has not spoken to Pourmand since the publication of the PNAS paper. Hassibi subsequently learned that Pourmand and Davis had filed an incremental patent “to overshadow” their 540 patent. “They should have involved us as co-inventors,” said Hassibi.
Pourmand had also resubmitted Xagros’s SBIR grants to receive NIH funding, which further upset Hassibi, but as he was still an international graduate student on an F1 visa, he decided not to pursue any further action.
Contacted by Bio•IT World, Pourmand says he “highly respects” Hassibi’s work and down plays any talk of a disagreement. He points out that the original 2001 patent “didn’t mention anything about hydrogen [ions]. We saw the signal generation based on incorporation of nucleotides. We solved electrical detection. [The 2001 patent] showed that we could detect electrically dNTP incorporation—but still at that time we couldn’t understand where that comes from.”
Pourmand says he left Xagros because the company was increasingly interested on bioluminescence, prompting him to return to Stanford to work on electrical detection. “After Xagros went belly up, I came back to Stanford, and continued working on that without Arjang.”
“In 2004, we realized the signal we’re detecting is hydrogen [ions]. In the original work with Arjang, we thought it was pyrophosphate actually… Even the sensors, they say they originally designed, we didn’t use that in the 2006 [PNAS] paper. We used commercially available polarized electrodes, amplifiers and so forth. Basically, we started from fresh.”
Pourmand insists he was “not ignoring his [Hassibi’s] work, absolutely not,” in the PNAS paper. “It’s completely different.” He adds he would “absolutely” have cited a paper that referred to his earlier collaboration with Hassibi if one existed—but it didn’t.
“I understand he feels I’m saying ‘it’s my patent,’ but I’m particularly referring to hydrogen detection and hydrogen release, not the electrical signals,” says Pourmand. “In the 2006 patent and paper, we’re clearly claiming it is hydrogen [ions]. I don’t really care which system you’re using to detect it, the release of hydrogen is important.”
Hassibi is now based at the University of Texas at Austin, where his research focuses on building new semiconductor chips for life sciences applications.
Hassibi was late learning about Ion Torrent’s interest in his intellectual property. In September 2009, Stanford OTL sent a “conflict of interest” memo to the Stanford Dean’s office on the proposed licensing deal of the two aforementioned dockets to Ion Torrent. (Davis was already a member of Ion Torrent’s SAB at that time.)
According to the OTL, the earlier S00-157 technology provided “a faster and cheaper alternative to current methods of DNA sequencing” by detecting variations in the charge of immobilized DNA. The S04-291 invention more specifically focused on sequencing “by detection of electric charge perturbations of polymerase-catalyzed reaction by the electrochemical detection sensor with immobilized DNA.”
Stanford’s OTL concluded that Ion was in “a strong position to successfully commercialize” both technologies. Despite widespread marketing to dozens of companies, only the 157 docket had been previously licensed (to Xagros). Several companies subsequently expressed interest in both technologies, but none took a license until the deal with Ion Torrent.
“I expected Stanford to take some equity, but I think they were convinced by Rothberg et al. that this is the maximum that they can get for it,” said Hassibi. “Stanford OTL didn’t have an obligation to involve me in the negotiations. I was not the Stanford PI involved in this project at the time. Who would they talk to? It would be Ron, but he had a conflict of interest [as a member of Ion’s SAB]. My anger is not at OTL, although they should have got equity.”
“I want to give a lot of credit to Rothberg,” Hassibi added. “We never perceived, including Ron or Nader, putting things in microwells. This is very important. Everything that was done in Stanford, as far as I know, has been based on immobilizing DNA on a gold electrode. Ion’s embodiment in terms of sample loading and interfacing is quite different, because they came out of 454, which is a perfect match for this technology.”
Indeed, Hassibi may have reason to be especially grateful to Rothberg. Last year, he started raising money again. “This was music to my ears: Rothberg on cover of Forbes magazine was the best marketing we needed! Companies like Life Technologies are not going to have a division of integrated circuit designers creating these CMOS chips. They’re going to outsource it.”
Hassibi’s new company, Insilixa, is a fabless semiconductor company that might count the next-generation of sequencing manufacturers among its future clients. •
This story also appeared in the 2011 July-August issue of Bio-IT World magazine.