Senniger, Powers, Leavitt, & Roedel
James Barta, an attorney at St. Louis-based IP law firm Senniger, Powers, Leavitt, & Roedel, focuses on software, hardware, and business method patent applications, and lists bioinformatics and computational biology among his practice areas. In a recent white paper, Barta wrote that "the trend towards filing bioinformatics-related patent applications appears to be increasing."
BioInform spoke to Barta last week about some trends that he's seeing in the market, when bioinformatics firms might want to consider patenting their tools, and whether open-source licensing and patenting are really mutually exclusive.
It's still relatively uncommon for developers to patent bioinformatics tools, as compared to other biotech areas, like genes and proteins, and other life science technology platforms like microarrays. Why do you think this is the case?
Bioinformatics spans multiple disciplines, so you often have people who are generally very focused on biotech, but less focused on the tools and how the tools are created, and how you actually get to the end result. And not only the biotech developers, but also the patent attorneys. Typically, at many firms, you just have a biotech [attorney] working on biotech innovation, and that biotech attorney may not be 100 percent familiar with software patents.
At our firm, we take a team approach. We pair a life sciences patent attorney with a computer science patent attorney for all the bioinformatics cases. That way, you've got the life sciences person who can focus on the genes, the proteins, the sequences — all the underlying biotech. And then you have the computer science person who can look at it and say, 'That tool, that algorithm that you created — that's something that's worthy of patent protection.'
So basically it's something that's overlooked [at other IP law firms or at informatics companies]?
I think it is overlooked, and I think it's on both sides. I also think, too, that bioinformatics is a fairly young industry. I don't think you see the same amount of patent filings, the same volume of patent filings in general, that you do in other industries, like the software industry.
It sounds like you think that's likely to change.
I do think that will change over time. I think patents are more pervasive nowadays, I think people understand more about them. I think software patents in particular are gaining some ground. IBM is a huge patent powerhouse, and I know Microsoft is definitely ramping up as well. So you're seeing those big-name companies with software patents, and I think you're seeing some of those companies move into bioinformatics … so I would expect the filings to increase.
IP litigation is very common in the software industry, and in other areas of biotech, but this is not the case in bioinformatics. What factors might account for this?
There are many reasons for that. First off, there's not that many bioinformatics patents [to litigate]. There's not the same level or volume of patents that we see in areas where there is more litigation. Secondly, many of the companies may not have huge litigation coffers. Patent litigation is extremely expensive. Nowadays, you're looking at anywhere from $1 million to $1.5 million just to get to trial to enforce a patent. Few companies can support that, and there are few companies where it would be reasonable to support that, even if they had the money. So most patent disputes get settled outside of court, and that's through licensing.
Also, the field is not that crowded in bioinformatics. In the software industry, there are many, many companies and many big software companies. In bioinformatics, I don't think we see that same size for the companies, and we don't see the same number of companies. I think it's still a growth industry and there's room for people to come in.
No bioinformatics patent, as far as I know, has ever been fully litigated in court. So as far as the validity of bioinformatics patent applications, I don't think there's any question about it, but there hasn't been a court case saying that bioinformatics patents are valid like there has been in the software industry. In 1981 there was a case where the court came down very clear that software patents are OK — you can patent software innovation — so then suddenly everybody started filing software patent applications in the US. And then in 1998, for business methods, in a case called State Street Bank, the court said that business methods are and always have been patentable subject matter, so as soon as that happened, the door opened for that, and now plenty of people are filing business method applications. But I think once a bioinformatics patent is actually litigated and a court says clearly, 'These are valid,' I think that suddenly there will be a lot more interest in the area.
Some bioinformatics companies prefer to protect their IP as a trade secret rather than patenting it. Would that be because they are unsure that bioinformatics patents are valid, or are there other factors that might enter into that decision?
Trade secret versus patent protection is a big tradeoff, and it always has been for any field, not just bioinformatics. I think that there are many reasons that you would want to protect something with trade secrets, and there are many reasons that you'd want to protect something with a patent, and it really depends on the technology itself — the actual invention — as well as the state of the industry. The most famous trade secret is Coca Cola, which back in the 1800s decided to keep their formula a trade secret. Had they actually filed for patent protection, they would have had to disclose the secret formula. After it expired, anyone in the world could copy the exact formula for Coca Cola, and I doubt Coca Cola would be what it is today. So that's one reason why trade secret protection is good — there's an indefinite duration. As long as you can keep a secret, you can remain protected.
The problem with that is that it's really hard to keep a secret, particularly in today's information age. And, secondly, if it's something that can be reverse-engineered, there's no point in keeping it a secret because it won't be a secret for long.
In general, what we've seen is that very few companies go for trade secret protection over time. Initially, if you're a small company and you don't have a lot of money to actually pursue patent protection, trade secret protection is reasonable. Also, if you're not sure if your company is going to be in existence in a year or two, trade secret protection is a reasonable way to go, but I wouldn't recommend it generally.
Open source licensing is very common in bioinformatics, and some commercial companies, like Affymetrix, have begun releasing bioinformatics software under an open source license. Should developers interested in releasing their code as open source still consider patenting?
This is actually a very exciting area. There is a lot involved here. There's been a lot of open source litigation; there's been a lot of big money moving around. SCO versus IBM is one of the big ones, and that's all about open source code, who takes it, is it enforceable, is it valid, and have you violated this open source license? And it's laying the groundwork for other industries that also have software that is being released under open source licenses. In general, I would say, 'Yes — you can definitely pursue patent protection for something that you're going to release openly.'
There are several open-source licenses out there, but probably the most popular one is the Gnu Public License. They have a provision that says if you take some code and protect it under the GPL, it sort of infects your product so that when you then take that code, put it into your product, and release it and sell the product, technically, under the terms of the license, you are supposed to release the source code upon request. It can be kind of nasty and some companies have questioned the validity of open source licenses, but in general it has not been fully litigated yet.
People don't doubt the enforceability of it, but they doubt whether people are actually conforming to the license, or, in most cases, violating the license. I think many companies may not know what their developers are doing, so I find that to be an interesting area, because if you're a company and you have a series of software developers — whether it's in bioinformatics or any other industry — I think you run the risk of having those developers download some free code and click though the license without reading it, not realizing that they've just infected the whole product with the GPL.
I think that happens a lot, and I think that happens particularly in bioinformatics because many of the people involved have come from academia, and it's pretty easy to download some freeware or shareware code and not read the license, use it, and move on without thinking twice.
You mentioned Affymetrix. If you do a search on the USPTO website, you'll see that Affymetrix already has hundreds of patents, and hundreds more pending, so they're pursuing that dual strategy as well.
That's what made me ask about that, because they have many software patents, and then more recently they've begun releasing some of their software under the LGPL, but I think the perception in the open source community is that you can't do both.
You can do both. My understanding of the sentiment in the open source community is that they don't like patents — it stifles innovation, it keeps people from creating with this constant threat of being sued or shut out of your business — and I understand those positions and there's probably some merit to that. But in general, there are two ways you can use a patent: You can get a patent for defensive reasons and you can get a patent for offensive reasons. Defensive reasons means you can protect your business and your technology. You may have little interest in suing somebody else or squashing some other competitor — you just don't want to be sued yourself. If you get a patent for an offensive reason, or you decide to enforce a patent, that's when you actually go out and send these cease-and-desist letters to potential infringers, and approach competitors in that way. But many companies don't do that. As I said, it's a very expensive thing to do. There are companies that are out there just to enforce patents, but I haven't seen that in the bioinformatics world. Most of what we see are people who are trying to protect their innovation and protect their research.
The last thing a company wants to do is ignore the patent process, and then get sued later for infringement by a competitor, and not have anything to back up their own technology with. In litigation, typically, if you're sued by one company, you open up your patent portfolio and say, 'What are they infringing of ours?' So you can go back and forth. If you have absolutely nothing in your portfolio, you have very little bargaining power. So pursuing patents and open source is, I think, not just reasonable, but almost necessary to keep a company alive.
Also, I'm sure there might be a concern that with open source licensing being litigated more, if any or all of a certain license is rendered unenforceable, then what? If the restrictions that that license put on a product would no longer be valid or enforceable, you have very little control over the person that you just gave your product to. If you have a patent, however, that could always be a second avenue of protection.