HEARING AND REQUEST FOR COMMENTS ON ISSUES RELATING TO PATENT PROTECTION FOR NUCLEIC ACID SEQUENCES University of California, San Diego International Center April 16, 1996 Reported By: TERESA BURT, CSR NO. 8607 PANEL MEMBERS: Bruce A. Lehman Stephen G. Kunin Edward R. Kazenske Lawrence J. Goffney Nancy J. Linck 1 MR. LEHMAN: My name is Bruce Lehman. I'm the 2 assistant secretary of commerce and commissioner of patents 3 and trademarks. 4 And joining me today in this hearing are 5 Lawrence J. Goffney, acting deputy district secretary of 6 commerce and acting deputy commissioner of patents and 7 trademarks. 8 To my immediate right, Edward Kazenske, deputy 9 assistant commissioner for patents. 10 To my left, Stephen Kunin, deputy assistant 11 commissioner for patent, policy and projects. 12 Over here on my far right, Nancy Linck, the 13 solicitor, the chief lawyer of the Patent and Trademark 14 Office. 15 I'd like to first thank the University of 16 California at San Diego for providing this facility for us. 17 This is the second time we've been out here in this campus. 18 This is a hearing to receive public comment on 19 a serious problem currently facing the Patent and Trademark 20 Office related to patent protection for nucleic acid 21 sequences. 22 The public was invited to comment on this 23 issue in a notice published on March 12th, 1996, in Volume 24 61, No. 49 of the Federal Register. 25 For over a decade, the Patent and Trademark 26 Office has been examining and granting patents to claims 27 receiting nucleic acid sequences. The scientific and 28 technological advances have permitted the rapid 4 1 identification of large numbers of genes or gene fragments. 2 The ease of utilizing automated techniques for sequencing 3 nucleic acid fragments has resulted in the filing of a 4 growing, although still relatively small number, of patent 5 applications, each of which claim thousands of nucleic acid 6 sequences. 7 Statistics reveal that the number of these 8 applications is growing, and based on the number of 9 organisms in genes still to be discovered, such growth will 10 continue for the near future. 11 In fiscal year 1991, the scientific and 12 technical information center of our Patent and Trademark 13 Office searched about 4,000 sequences. In fiscal year '95, 14 they searched about 22,000 sequences. Currently we have 15 over 200,000 sequences claimed and at least 70 patent 16 applications awaiting search and examination. 17 Our estimates show that the search of 100 18 sequences requires about 15 hours of computer time. But 19 the evaluation of the search results for those 100 20 sequences requires about 65 hours of examiner time. The 21 PTO currently has two massively parallel processor 22 computers and could run the searches in about two years, 23 when the computers are running 24 hours a day, seven days a 24 week. 25 To examine this relatively small number of 26 patent applications only with respect to the prior art, 27 however, would require only 9- -- would require over 90 28 senior-level staff years. Thus, in order to accommodate -- 5 1 in order to process these applications, the entire staff of 2 the Biotechnology Patent Examining Group 1800 would have to 3 work for more than nine months exclusively on these 4 applications. 5 These applications present a challenge to us 6 at the PTO, and we need help and suggestions on how we can 7 address the problem. The United States is a leader in the 8 rapidly growing field of biotechnology, which is a growth 9 industry important to the economic health of this country. 10 The PTO has taken a very active role in 11 working with its customers to try to simplify policies and 12 procedures in ways that encourage and promote the growth of 13 this industry. 14 We are committed to improving the 15 responsiveness of the PTO to its customers and to more 16 effectively address the needs of industry. We must find 17 ways to search and examine the pending applications and 18 provide these applicants with appropriate protection -- and 19 I emphasize the word appropriate patent protection -- for 20 the inventions without creating an imbalance in the 21 appropriation of the resources within and among the 22 technologies and the patent-examining groups. 23 These policies established must permit the 24 timely and thorough examination of all applications which 25 require the same resources for completion. 26 We are currently working in partnership with 27 the applicants of these applications in order to explore 28 innovative mechanisms to accomplish the required work in 6 1 processing them. We appreciate the time each of you has 2 taken to attend this hearing -- all of you in the audience 3 today -- and provide us with input into the solutions to 4 this problem. 5 A transcript of this hearing will be prepared 6 and will be made available for purchase by the public 7 approximately 10 days after this hearing. Copies of the 8 transcript will also be available for purchase directly 9 from the stenographer. The name of the stenographer 10 service today is Peterson & Associates, and their telephone 11 number is area code (619) 260-1069. That's (619) 260-1069, 12 Peterson & Associates. 13 We have received 10 written comments and seven 14 requests to appear orally at this hearing. Any persons who 15 wish to speak and who have not previously informed us of 16 their desire are encouraged to add their names to the list 17 located on the table at the rear of this room. 18 In order to permit all persons requesting to 19 appear orally, including those persons signing up today to 20 present testimony, we would request each speaker to limit 21 their presentation to 15 minutes. Those persons who wish 22 to provide additional comments should submit their comments 23 in writing to us no later than April 23rd. 24 The speakers have been listed, and I think we 25 have a list of them back there, don't we? 26 They've been listed in the order in which the 27 requests were received by us. You may also pick up at the 28 table of the rear of the room copies of the Official 7 1 Gazette publication of the notice of these hearings and 2 request for comments on issues related to patent protection 3 for nucleic acid sequences. 4 When you present your comments, we would 5 appreciate it if you would please give your name and 6 address and tell us whether these comments are your own or 7 whether they are those of a law firm or company that you 8 represent or whether you represent an organization and are 9 presenting comments on their behalf. 10 Speakers are requested to limit their remarks 11 to the questions presented in the Federal Register 12 publication of March 12th, 1996. And I am told that the 13 first person who is here to testify before us -- where 14 is -- oh, right there -- to testify before us this morning 15 is Alice Martin. 16 So, Ms. Martin, would you please come forward. 17 MS. MARTIN: Commissioner, ladies and 18 gentlemen, my name is Alice O. Martin. I represent the law 19 firm of Brinks, Hofer, Gilson & Lione in Chicago. 20 My address there is the NBC towers. I also 21 believe I am speaking consistent with the position of 22 B.I.O., of which our firm is a member. 23 As attorneys for the biotechnology industry, 24 we applaud the cry from the patent office for help in 25 solving what we consider a very common problem. We 26 appreciate that the examiners are hacking their way through 27 this jungle of nucleotide and amino acid sequence databases 28 and that they face a seemingly unmanageable task and, 8 1 indeed, as the commissioner mentioned, the jungle is 2 growing even as we speak. 3 Now, some facts affecting this problem are the 4 database -- databases that exist were not specifically 5 designed for the use of obtaining prior art for obviousness 6 rejections but rather they were designed to accumulate and 7 expand on all information on all sequences that are known. 8 The tools for the search algorithms to hack 9 through this jungle were not specifically designed to find 10 prior art for obviousness rejections, but rather they were 11 designed to quantify and describe similarities, sometimes 12 based on an evolutionary model. And many of them are 13 designed to ferret out even distant evolutionary 14 relationships. 15 Another fact addressing this problem is that 16 examiners are currently, as I understand it, not provided 17 with maps or search strategies with which to find their way 18 through this database jungle. This means that they can 19 take false paths, hit dead ends, waste valuable time and 20 become frustrated. 21 Furthermore, the subjective, somewhat 22 unpredictable paths that are basically unreproducible cause 23 uncertainty among applicants and certainly down the road 24 will cause problems in litigation in the courts. 25 Now, how can we make manageable, yet legally 26 sound, solutions to this problem? I suggest we tame the 27 jungle and provide maps through which we could move through 28 the jungle. Now, the tools or the search algorithms used 9 1 will to some extent depend on the choice of maps. 2 How can we tame the jungle? We're not aware 3 of any legal requirement that the patent office has to do 4 an exhaustive, brute-force search of everything that exists 5 in the current databases. So let's prune the jungle. 6 Let's not search all databases just because they exist, 7 particularly when many of them are repetitive. 8 Perhaps your meeting with experts, as I think 9 has already begun, to explain the construction of the 10 different databases would be helpful so that either you can 11 pick the best database for overall patent searches or pick 12 the best for certain sets of claims. Claims to CD&As may 13 not be using the same database as claims to ESTs, for 14 example. And there are specialized databases that have 15 been developed, for example, by the NIH group for these 16 specialized searches. Perhaps working with these groups 17 already in the field will be helpful. 18 Now, a more extreme suggestion that could be 19 necessary is to plan an entirely new jungle. 20 MR. LEHMAN: Can I ask a question? 21 MS. MARTIN: Certainly. 22 MR. LEHMAN: You said that the NIH has 23 developed these specialized databases for the searches, but 24 the NIH has withdrawn their patent applications. Are you 25 talking about databases they developed for other 26 purposes or databases -- 27 MS. MARTIN: That is my understanding. And 28 I'm not up on the latest but I certainly -- 10 1 MR. LEHMAN: -- they developed for other 2 purposes for their own research purposes? 3 MS. MARTIN: I have read that they were 4 working on specialized databases, and I thought it might be 5 worth exploring, at least whether they have them available 6 or not, what methods they have used to develop those might 7 be very helpful. 8 As I said, a more extreme solution would be to 9 completely plan a new jungle. In this case we would take 10 the old databases and create a new database specifically 11 designed for the purpose of finding prior art upon which to 12 base obviousness rejections. 13 This way perhaps a grouping in the database 14 could be such that you don't have to keep redoing the same 15 kinds of searches over and over, and I'll say a little bit 16 later about the maps that could be used to perhaps organize 17 this database. 18 We also then in the new database could 19 organize it to correspond to different types of claims that 20 are traditionally sent in by applicants. Also it was 21 mentioned, I believe at the April 4th presentation, that 22 there's a problem in determining what is, quote, prior art 23 because of the difficulty in determining what dates various 24 sequences are entered. That information is in the database 25 but is evidently difficult to retrieve, so art can be 26 looked at but really isn't prior art in the current 27 methods. 28 A new database may also help the problem 11 1 evidently of finding extension numbers and then trying to 2 look in the references that are cited and identify which 3 sequence is really referred to in a particular reference. 4 And finally, one of the reasons I think the 5 patent office gave for using its present databases is that 6 there are references that are retrieved along with the 7 sequences. 8 That may not be something that needs to be 9 done for each sequence and, again, can be considered in a 10 new database. 11 I think everyone agrees that redundant 12 sequences should be removed, and this would be solved 13 either by pruning the databases or developing a new one. 14 Not everyone agrees that sequences should be trimmed first 15 before a search of areas that can give hits that, as some 16 of the developers of the search algorithms say, are of 17 dubious biological significance. And here again it might 18 be helpful to reevaluate the tools that are being used for 19 searches. 20 My understanding is that the PTO has selected 21 a search algorithm because of increased sensitivity. This 22 is puzzling to us because for increased sensitivity, you 23 are picking up more distantly related sequences, when for 24 obviousness rejection, what we really have to focus on is 25 structurally similar sequences. The more distant the 26 relationships, the more likely there are to be structural 27 dissimilarities, making it more difficult to prove 28 obviousness because, number one, the structures are very 12 1 different and, number two, to find art that would suggest 2 or motivate all of the different changes between that 3 distant-claimed and the period-claimed sequence would be 4 very, very rare. So it seems to me that may be defeating 5 the purpose. 6 The patent office search goals appear to be 7 somewhat different from those we use in the scientific 8 community, and I realize that many of the examiners 9 themselves were established scientists and may have to 10 remove their scientific hat and focus not on the biological 11 relationships but what is structurally similar in terms of 12 the legal sense. And here again I have a seminar from 13 technical experts who have developed the different 14 algorithms. Could be very, very helpful. 15 Also, remember that any of these algorithms 16 will find matches because they're really just looking at 17 the sequences as letters and basically searching through to 18 find a probabilistic model that says this sequence is more 19 similar than would be expected by chance. So in and of 20 itself, there is usually no legal or biological 21 significance. 22 And finally, I suggest we develop maps or 23 search strategies rather than jumping into the jungle first 24 and emerging overburdened and overwhelmed with lots of 25 sequences. We need a strategy because searches in the 26 mechanical arts, if you find an application that has a 27 carburetor and a machine, you don't generally look at all 28 the art in the world that has a carburetor. 13 1 Similarly in the chemical arts, if you find a 2 composition that has a benzene ring, you don't usually look 3 at all compositions that have a benzene ring. 4 One option is to use key words first before 5 going into a strategy of searching the database. Key words 6 can include species, library, protein or function as 7 related in the claims and the specification. I think 8 considering the claim as a whole, as the courts and the 9 board of appeal have instructed us, may suggest that some 10 sequences don't even have to be searched. 11 We could group sequences, and the applicants 12 here should be helpful in this regard in terms of length, 13 core chemical structure, again to relate to the chemical 14 practice art. If we do these maps, this will provide focus 15 using legal standards that already exist. There's nothing 16 new, I think, in the obviousness of DNA sequences. 17 As you all know, this will be less open to 18 criticism and less open to challenges to validity of the 19 patents than using something such as "Let's look at the 20 first 45 best hits that we get out of the search." That 21 seems to me very arbitrary and I think will open the patent 22 up to much criticism in the future. 23 Plus, by using that kind of ab priori and 24 arbitrary criterium, you may have too many sequences 25 examined for some and too little for others. As we saw in 26 the U.S. Supreme Court decision in Deere, there are two 27 criteria that must establish a prima facia case of 28 obviousness. One, is the claim structure different. And 14 1 two, are the differences suggested by the art. 2 Both the courts and the board have said DNA is 3 a chemical. Proteins are chemicals, so we must follow the 4 chemical practice, which is the same as iterated in Deere: 5 is the composition structurally similar? And if so, 6 similarity alone will not be a bar unless there are 7 secondary references motivating those changes. 8 For example, a nucleotide change in two 9 positions must have some reference or some motivation in 10 the art to make those particular differences. As we see 11 it, the existence of these massive databases pose a big 12 risk for hindsight and obviousness to try creeping into 13 obviousness rejections. Looking at the time the invention 14 was made, add a database that exists in the world, the 15 inventor is going to be faced with infinite number of 16 possibilities, and it would be unpredictable to merge or to 17 change sequences without some motivation in the art. So we 18 ask that the examiners take caution in selecting these 19 merely by hindsight. 20 Some other things that might be kept in mind 21 in planning our sequences searches are that the courts have 22 already said that defining a sequence and a method of 23 making it will not support an obviousness rejection. One 24 species will not necessarily make obvious a similar 25 sequence in a second species. 26 Here our big problem, I think, boils down to 27 what is structural similarity in terms of DNA and amino 28 acid sequences that can be applied then to the chemical 15 1 practice standards. We need to decide what is a structural 2 similarity to focus the search rather than getting all 3 sequences that may have some limited degree of similarity. 4 In other words, why look for all sequences because they are 5 somewhat similar if you're not able going to be able to use 6 them in an obvious rejection anyway. 7 Dealing with some of the smaller sequences 8 that we've mentioned that are presented in large numbers, a 9 portion of the sequence is not likely to make a larger 10 sequence obvious because there is no way of predicting that 11 larger sequence, and the courts are certainly going in that 12 direction. If you have a larger sequence, why look for 13 small fragments when there is no teaching of boundaries of 14 the fragment. 15 I think here, too, that applicants may have to 16 take a role in being more second specific in some terms 17 that are now fairly loosely defined such as homologues, 18 analogues, mutants, variants and fragments. These are not 19 mutually exclusive terms. They need to be defined. 20 Now, I know none of us like a piecemeal 21 prosecution. However, there may have to be some 22 negotiation on a 112 section first before a search strategy 23 is planned. 24 In summary, rather than trying to speed up 25 previous ways to do these searches, particularly before 26 encountering these large numbers or large size sequences, 27 we recommend that we reevaluate whether we shouldn't tame 28 the jungle by finding more suitable databases or groups of 16 1 databases for the legal search and by developing maps based 2 on obvious criteria already known to the courts and the 3 board so that we can focus the search and not waste 4 valuable time looking at dead ends. Thank you. 5 MR. LEHMAN: Thank you, Ms. Martin. 6 Are there any questions from other members of 7 the panel? 8 Thank you very much. 9 Next I'd like to call on John Burke. 10 MR. BURKE: Good morning. My name is 11 John Burke. I'm here from Maas Parr Computer Corporation. 12 Thank you for having us. I'm going to offer a more 13 technical presentation that won't have much legalities in 14 it. 15 Maas Parr is a database processing company and 16 in support of bioinformatics of a particular -- 17 THE REPORTER: I need you to speak up a little 18 bit more. Okay? 19 MR. BURKE: Okay. 20 Maas Parr Computer is a computer database 21 processing company. 22 MR. LEHMAN: Could maybe put this closer to 23 your mouth. Okay? 24 MR. BURKE: Okay. Thank you. 25 And we also do bioinformatics. We are a 26 complete computer system with fast processing as well as 27 the infrastructure that's required by central processing. 28 Our internal communication is 28 megabytes per second, and 17 1 more importantly we have audio capabilities which are 2 needed at the U.S. PTO. We have up to 240 megabytes 3 transferred from the disk to the computer core which will 4 be very important doing the number of searches that you 5 have to perform. 6 The advantages to U.S. PTO are a quick access 7 to the database, very flexible access, which is very 8 important of your current model of operation. Because you 9 have many independent investigators working on different 10 patent applications, it's not possible to order the 11 database searches. And with many architectures, it's 12 necessary to preload the database. And therefore to 13 optimize your search time, you would have to perform the 14 searches in a certain order on certain databases. And with 15 our systems, this isn't necessary. 16 As well, the entire range of bioinformatics 17 applications are accelerated, not just the database 18 searching. As you mentioned, one of the problems is the 19 bottleneck in investigator time, and so to accommodate 20 this, we are parallelizing the entire range of applications 21 that support the interpretation of search results. 22 We're a very flexible architecture and very 23 easy to develop one. We have over 260 installations with 24 other 100 computers in universities where research is being 25 performed. Over 26 installations are specific to audio 26 informatics, so there's a very good basis for this. 27 A result of such a large basis is continual 28 application development and development of tools for the 18 1 processing of results. 2 The result of having so many bioinformatics 3 applications running is that investigator review time can 4 be restricted. Not only do we have the database searching 5 tools but we have many database processing tools which is 6 very important to removing the redundancies inherent in 7 biological databases and also to reduce the number of 8 results returned and save the amount of time that an 9 investigator might have to spend going through and 10 examining results manually. 11 In addition to the bioinformatics tools and 12 computer system, Maas Parr Computer Corporation has 13 professional services with experience in designing 14 automated DNA processing for hybrid screening. That would 15 mean that the reviewer time is saved both in setting up the 16 searches but also in the interpretation of the results 17 because much of the interpretation can be automated, and we 18 do have experience in this field. 19 In addition we have educational services on 20 the high level, even on the high level on how to use 21 bioinformatics applications on providing search strategies 22 and also to the lower levels of maintaining hardware, 23 programing, language and such. Additionally, we have 24 worldwide, 24-hour customer service support. 25 Now I wanted to provide some level of 26 specificity in dealing with the problems mentioned in the 27 announcement for this hearing, but there is limited time, 28 so I would just like to talk about one application that's 19 1 very suitable for today, and that is the clustering 2 methodology that we have employed for biological sequences. 3 One of the big problems with biological 4 databases, as mentioned before, is that there are a large 5 number of redundancies, so therefore clustering refers to 6 the attempt to reduce the bulkiness and the cumbersomeness 7 of the database while at the same time preserving all the 8 information content. 9 One possible model for the use of a clustering 10 methodology is the database processing which is done at the 11 systems support level at the U.S. PTO. This would not be 12 done by the investigator but the infrastructures and the 13 support mechanisms available to the investigator. A 14 database would be clustered, processed into another 15 database, then having the redundancies removed and making 16 them a smaller overall size. 17 The benefits are faster search time because 18 there are less entries, but more importantly the results of 19 such a database search would be much more easy to interpret 20 because there are less redundancies and less multiple hits. 21 Another model for use in clustering is by the 22 investigator himself or herself. Clustering can be done at 23 any level of stringency that the investigator decides is 24 necessary, and one possibility is doing prior art searches 25 on multiple databases and then obtaining the results which 26 would be very large and probably very redundant with much 27 repetition. 28 The search results can be clustered to produce 20 1 processed results which will have the redundancies removed, 2 and it will be much easier to go through. Thus, you save 3 time in the interpretation of the results and eye strain on 4 the investigator having to look at multiple hits of the 5 same thing. 6 A very big problem in searching databases, 7 especially EST, is that there are many ESTs for one gene, 8 and really you only need the information of the gene 9 itself. A clustering methodology would allow you to have 10 one hit per one gene. That will greatly reduce the size of 11 the databases. That will greatly reduce the amount of time 12 that an investigator needs to interpret the results of his 13 or her search. 14 The third instance where Maas Parr feels that 15 clustering will be of use is tackling the problem of mass 16 injuries, where instead of the single sequence, you might 17 have hundreds or even thousands of sequences in the one 18 application. The idea here would be for the investigator 19 to run his or her own clustering of the data set that's 20 provided of the patent application that's provided. 21 The clustering would be done at the level of 22 similarity that they choose, and the output will be a 23 processed submission which will contain the original 24 sequence data, but more importantly the sequence data will 25 be put into work groups that are similar in pontiff levels, 26 a similarity that's chosen by the investigator. 27 This saves a lot of time because instead of 28 viewing the patent application as thousands of individual 21 1 sequences, he or she can view the application as a whole in 2 its work groups and a first-level organization has been 3 provided which will save the amount of time needed to 4 process the application and the amount of searching that 5 you'll have to do. 6 That's only one of the applications available 7 for database processing from -- in the Maas Parr system, 8 but we felt it was the most relevant to discuss today 9 because it attacks the problem of mass submission. It 10 addresses the problem of the investigator time needed to 11 analyze the results. It even allows one to work on the 12 database level to try to clean up databases to remove 13 redundancies, to try to improve some of the -- some of 14 the -- not flaws but particularities of the U.S. PTO 15 situation. 16 So in conclusion, Maas Parr is striving to 17 reduce investigator time as well as provide fast, rapid 18 bioinformatics tools in parallel. 19 MR. LEHMAN: We're using your system right 20 now. Is this the first time some of these -- that we've 21 heard some of these suggestions, or have you been giving 22 them to our -- 23 MR. BURKE: This is the first time that you've 24 heard about clustering, yes. Right now U.S. PTO is using 25 the Maas Parr to do rapid database searching and such. 26 The clustering tools are rather new, but they 27 are working currently. And they're meant for large-scale 28 uses in database and also small-scale use an individual 22 1 investigator would do. 2 MR. LEHMAN: When you deal with the PTO, do 3 you deal with the informations systems people primarily 4 or -- 5 MR. BURKE: Primarily, I believe that we deal 6 with the information systems. 7 MR. LEHMAN: -- as opposed to the examining 8 core directly? 9 MR. BURKE: Yes. 10 MR. LEHMAN: Are there any other questions? 11 MR. KUNIN: You began to mention a little bit 12 about the process of post processing in addition to the two 13 styles of clustering. The first speaker spoke in terms of 14 setting up a process of post processing which might include 15 the formation of an expert system which would kind of help 16 analyze the search results from the perspective of the law, 17 that is, to collect the information not only from the 18 standpoint of identity but also enough structural 19 similarity and relationships that might suggest how some of 20 that -- some of the results might be combined for 21 obviousness purposes. Do you have any comments from 22 -- for Maas Parr on the post-processing solution? 23 MR. BURKE: The post-processing solution, we 24 have some experience, and we have implemented similar 25 systems in the commercial companies. It refers to 26 automatic screening of sequences. I believe that she was 27 speaking of a -- of a set methodology before -- sort of -- 28 I believe she was referring to guidelines for the 23 1 investigators in doing their searches. 2 I think that would be very useful, and an 3 automated schema can be put in place to save time. I 4 personally feel that the investigator should have a 5 latitude to perform any kind of search they want, though. 6 I think such things should be a guideline only and not a 7 straitjacket, but such a thing is very possible and is a 8 very good idea, yes. 9 MR. LEHMAN: You know, you've heard the 10 suggestions of the previous witness and made some of your 11 own, and you also heard my little opening comments. And my 12 opening comments suggest quite a large gap between the 13 resources under our present methodology which would be 14 required to examine these applications and what would be 15 reasonable, given our present fee structure and resources. 16 Can you give me some kind of judgment on your 17 part as to what -- how much of this gap would be reduced in 18 your view by the proposals that you've offered? 19 MR. BURKE: As far as the specific amount -- 20 MR. LEHMAN: It doesn't help us very much 21 to -- to -- to -- to chop it away 10 percent of a problem 22 that is -- is totally debilitating. We need to chop away 23 99 percent of it. 24 MR. BURKE: Yes. It's not just going to be 25 buying more hardware, although that will be necessary, 26 considering how much work you have ahead of you but also 27 the methodologies that you'll need like this clustering. 28 As far as a reduction of database size, I 24 1 don't know. It won't be on the level of 50 percent. 2 However, on the other end by the individual investigator, 3 it could save quite a bit of time in mass applications 4 because there's bound to be a lot of redundancy, and why do 5 a thousand searches when you can do a hundred. I think you 6 will need to buy more hardware. You will need to 7 modernize, but more importantly, you'll have to start -- 8 you'll have to keep up with the latest technology and 9 automated processing and clustering and other such tools 10 that can help you do more than just database searching in 11 prior art investigations. 12 So you asked me for a specific number or a 13 guess as to the amount of time that can be in saved. I 14 don't know that I could do that, but on the top of my head, 15 I can think about it, but I do know that it will -- you 16 have to -- just by buying -- buying more of them will not 17 solve the problem but employing smarter systems of the kind 18 that I'm talking about today will -- will provide the 19 solution. 20 MR. LEHMAN: Thank you very much. 21 MR. BURKE: Thank you. 22 MR. LEHMAN: Next I'd like to call 23 Suzanne Biggs. 24 MS. BIGGS: Good morning, commissioner, 25 gentlemen and lady. I'm Suzanne Biggs. I'm with the La 26 Jolla office of the law firm of Lyon & Lyon, 4250 Executive 27 Square, Suite 660, La Jolla, California 92037. 28 This testimony and comments that I'm giving 25 1 today reflect my own opinions and should not be considered 2 as representing the position or opinion of Lyon & Lyon or 3 of any of its clients. They reflect some of my experiences 4 as a practitioner in this area. 5 Lyon & Lyon is a law firm specializing in 6 intellectual property law. I work primarily advising 7 clients in patent matters, particularly in the areas of 8 biochemistry, chemistry and biotechnology. I've been 9 specializing in biotechnology and biochemistry for about 10 the last six or seven years. 11 We understand the concerns of the Patent and 12 Trademark Office, that they want to issue good patents that 13 are directed to novel, unobvious, useful subject matter 14 that will stand up to challenge in the courts. That's what 15 we want to obtain for our clients as well. 16 We also understand their concern with regard 17 to the burden of examining certain applications. However, 18 I'm not so sure that proposing differential fees is the 19 right way to go about it. These proposed differential fees 20 raise three questions to my mind. 21 Is it appropriate to base fees on the 22 technology type? Is it appropriate to base fees on the 23 estimated difficulty of search by some sort of objective 24 criteria? Is it appropriate to revise fees in a piecemeal 25 manner when the PTO finds that a particular group of 26 applications appear to be more difficult to examine than 27 the norm? I believe that the answer to all three questions 28 should be no. 26 1 With regard to the suggestion that fees be 2 based on technology type, we believe it would be unfair to 3 pick out biotechnology applications to have higher fees 4 just based on the fact that they're biotechnology. 5 Typically, biotechnology companies are small 6 companies. They're money sensitive. Patents are critical 7 to their continued existence, to their continued ability to 8 raise funds, to their ability to get their products to 9 market. Biotechnology companies typically are one area 10 where we seem to be extremely good at competing with the 11 other companies in the world. In fact, so many corporate 12 deals are being done with foreign companies, they must 13 think that our biotech people are very innovative. 14 Differential fees that are much higher would 15 discourage patent filings. It probably would not stop them 16 from filing the ones they see as clear winners, but some of 17 the ones that may be enabling technology, it would 18 encourage them keeping things such as trade secrets. It 19 would decrease the amount of information that enriches the 20 area in general. 21 Also, another thing to consider is not all 22 biotechnology applications are large and unwieldy to 23 examine. Very few contain thousands of sequences that have 24 to be searched. It appears that one of the things that 25 have really sparked these hearings are a small number of 26 applications that seem to have caused particular problems 27 and problems that may not be representative of all 28 biotechnology applications. 27 1 Also, another point to consider is large and 2 unwieldy applications, meaning those that are hard to 3 examine, are not limited to biotechnology. Other 4 technology areas have applications that prove difficult to 5 examine and take a large amount of examiner time. 6 Applications in the computer and software area and other 7 high-tech areas and as time goes on with new technologies, 8 I think a variety of them will have problems that way, 9 particularly in their early stages. 10 As the industry progresses and matures, 11 biotechnology applications will become more focused and, 12 indeed, many of them are becoming more focused now. People 13 are directing them to more specific solutions to problems. 14 And I think this will continue to happen as the art 15 matures. 16 Therefore, it is believed that it would be bad 17 precedent to increase fees for one type of application in 18 general such as biotechnology when there seem to be 19 problems examining them or whenever you get a new 20 technology type which doesn't really have established 21 guidelines for examination. 22 Other fee considerations are -- many -- even 23 focused biotechnology applications often have a large 24 number of claims and as such they are already paying higher 25 filing fees than many applications in the, for example, 26 mechanical arts, where claim numbers of several hundred are 27 unusual. 28 Also, as the industry matures -- and even now 28 1 while it's sort of in its adolescence and companies are 2 looking for corporate partners to help fund clinical trials 3 and the like -- many of these applications are losing their 4 small-entity status, and as that increasingly happens, that 5 will increase the fees the patent office gets, but it will 6 also increase the fee burden on the companies. 7 I have a few comments to make about sequences 8 and numbers. There's -- we've all heard some comments that 9 possibly fees be based on the number of sequences in the 10 application. And one thing, as the rules stand now, 11 anytime a sequence of four or more amino acids -- as long 12 as there are no D amino acids or 10 or more nucleotides are 13 listed in the application, they must be put forth in the 14 sequence listing. 15 In many applications, the majority of these 16 sequences, one, are not claimed, two, need not be searched 17 to determine patentability of what is claimed. Many of 18 these sequences are present in the application to satisfy 19 the requirements of Section 112: how to make, how to use, 20 best mode. There are things like best primers. There are 21 things like plasmid liter sequences, things that the 22 applicant does not expect to get patent coverage on. 23 Often, they will be things that if anyone would be entitled 24 to patent coverage, it would be a company that, say, 25 provides primers, plasmids and the like. So we believe 26 that it would not be appropriate to base these fees on the 27 number of sequences. 28 We can charge higher fees just because there 29 1 are a large number of sequences in the application, where 2 these sequences need not be searched in order to determine 3 whether a patent should issue on the claim subject matter. 4 In fact, complying with the sequence listing 5 requirements can be fairly burdensome in some of these 6 applications just because of the number. We can understand 7 the importance because you all, however you decide to 8 determine your database, need to have all these sequences 9 that are known in it where appropriate, but for determining 10 whether a particular applicant is entitled to a patent, 11 they're just plain not germane. 12 I've heard several suggestions for easing the 13 alleged problem of searching sequences. At present, it 14 seems to be limited to a relatively small number as 15 compared to all biotechnology-type applications but very 16 large applications. And perhaps guidelines can be made to 17 determine what a common focus should be when multiple, 18 large numbers of sequences are being claimed, but that's 19 something for those who are more involved with those sorts 20 of applications to deal with. 21 There's also been some suggestion of 22 applicants making a prefiling search and putting forth some 23 sort of certificate. There is a provision now in petitions 24 to make special, where applicants are requested to put 25 forth -- one of the grounds is a prefiling search and 26 various representations. 27 In view of the sorts of representations and 28 the uncertainties of how those would be construed in a 30 1 later proceeding, I see in my own experience very little 2 use of those provisions. And with respect to a prefiling 3 certificate of some sort of search by applicants, I see 4 some problems. 5 Those problems would be: and how do you 6 standardize what has been searched? What databases do you 7 have to say this search was done within a certain number of 8 days before filing? Also, in determination of guidelines, 9 how close does something have to be before you have to note 10 it in your certification? 11 The other issue that I can see is a problem 12 is: Who would make the certification data that a 13 particular search that fits certain criteria is made? 14 Would it be the attorney? If it's the attorney, that would 15 further increase costs for the small clients because it 16 would require a fair amount of attorney time to review the 17 search and the protocol. 18 If it's a nonattorney, how would we know what 19 the quality was of the people doing such a search, that 20 they really understand the databases, they understood the 21 criteria? I can't -- these are just issues to consider in 22 going forward with any sort of the suggested changes, and I 23 thank you for your time and your consideration. 24 MR. LEHMAN: Thank you. 25 Are there any questions? 26 MS. LINCK: I have one question. 27 You mentioned that we may not search all the 28 sequences disclosed in the application, but what happens 31 1 then when the claims are amended to claim one of the 2 sequences that wasn't searched initially but was disclosed? 3 MS. BIGGS: You would have searched at that 4 time the sorts of sequences. I was particularly commenting 5 on, for example, you have an application where you've 6 isolated a new protein from a source. You have cloned the 7 protein. Many of the sequences that are involved in the 8 cloning are probes, primers, things like that that under 9 these regulations in the 37 CFR 2nd, 821 and so forth will 10 have to be listed. However, these are things often that 11 are either known sequence or they're things that you just 12 are not claiming. They are just not related to what is 13 being claimed. They're not related to the proteins, 14 sequence of the protein that are being claimed. They're 15 not related to the CDNA. 16 MR. LEHMAN: But we would have to identify -- 17 these are in the prior art already, which is what the 18 search accomplishes. Would you offer an admission in the 19 application that these are prior art sequences? 20 MS. BIGGS: Well, if we're not claiming the 21 sequence, many times we will say where the sequence comes 22 from, that they come from a particular provider of liters 23 that one puts in a mass mix or they're just things that we 24 don't claim. Of course, if we later claim them, they would 25 have been to searched, but just as one pays if you start 26 out with 20 claims and you add another 40, you pay for 27 them. If one were to amend one claim to add an additional 28 50 sequences that the Patent and Trademark Office could do 32 1 if they were to base a fee on the number of sequences that 2 were claimed as opposed to presently the application they'd 3 deal with that, but many of those sequences don't need to 4 be searched. 5 MR. LEHMAN: So would you support an increase 6 in fees if in fact an amendment would be made that required 7 additional searches to cover the additional costs of the 8 search? 9 MS. BIGGS: Right. Now, first of all, I don't 10 support increased fees for biotechnology applications. I 11 would have to see some sort of proposed regulation. I 12 think in general where they're focused, it's without a full 13 review of all sorts of applications that may be difficult 14 to search. I don't think it's fair to pick out one area of 15 technology as having -- that may be easier to quantify. It 16 would be like in pharmaceuticals if you have a group that 17 claims more than 1,000 compounds, you have to pay extra. 18 And some of those, I'm sure, must be very difficult to 19 search appropriately because of -- even with the 20 restriction practice as it is now. 21 MS. LINCK: But if there was a method for 22 adjusting fees to make it fair across the board, have the 23 fees more aligned with the actual costs to the office, 24 would that be something you would consider? 25 MS. BIGGS: I'd have to see a concrete 26 proposal I think. 27 MS. LINCK: Thank you. 28 MS. BIGGS: Thank you. 33 1 MR. LEHMAN: Thank you very much. 2 I'd next like to ask Richard Klobuchar, 3 please. 4 MR. KLOBUCHAR: Good morning, commissioner, 5 members of the trade office and members of the public. My 6 name is Dr. Richard Klobuchar. I'm currently vice 7 president of the corporation of Science Applications 8 International Corporation. SAIC is headquartered here in 9 San Diego. My office is at 200 North Glebe Road in 10 Arlington. I'm currently also the chief scientist for 11 SAIC's Systems Engineering and Technical Assistance support 12 contract for the Patent and Trademark Office and currently 13 co-principal investigator for a pending distributed 14 supercomputer project with ARPA and the PTO which is 15 referred as to the Metacomputage Project. These comments 16 represent the comments of my corporation. 17 SAIC is an employee-owned, diversified, 18 high-technology research and development services company 19 focusing on information systems, telecommunications, 20 security, transportation and health. Our 22,000 employees 21 are organized across autonomous sectors and transform 22 emerging information technologies into customer-focused 23 solutions for literally thousands of government and 24 commercial customers. Under the competitively awarded 25 System Engineering and Technical Assistance contract for 26 the United States Patent and Trademark Office (USPTO), 27 we have been performing such technology infusion tasks 28 since June of 1993. 34 1 Since September 21st of 1995, SAIC, as the 2 PTO-SETA contractor, has been supporting the USPTO in the 3 selection of hardware, software and approaches to 4 facilitate their examination of biotechnology patent 5 applications. Efforts to date have focused on the analysis 6 of requirements and has resulted in a Preliminary Molecular 7 Sequence Searching Systems Requirements Analysis document 8 dated December 19th, 1995. 9 We anticipate continued efforts for the USPTO 10 through a new Task Order entitled "Automated Support for 11 Biotechnology Patent Examination." Under this Task Order, 12 the SAIC PTO-SETA staff will provide technical and 13 analytical support to enhance USPTO capabilities for 14 automated biosequence searching in support of PTO 15 examiners. 16 On the 15th of March 1996, SAIC's Laboratory 17 Sensors and Automation Division, located here in San Diego, 18 entered into a contractual relationship with Incyte 19 Pharmaceuticals, Inc., for the development of a 20 next-generation robotic system in support of Incyte's DNA 21 sequencing operations. 22 On the 18th of March, SAIC notified the USPTO 23 that via the Internet, we had discovered the work being 24 performed by SAIC's Laboratory Sensors and Automation 25 Division in San Diego. In our OCI notification letter, we 26 described the work being performed by SAIC's Laboratory 27 Sensors and Automation Division. We went on to explain 28 that the role of that laboratory is strictly in the area of 35 1 robotic enhancement for the Incyte laboratory equipment 2 and, therefore, we felt that no Organizational Conflict of 3 Interest (OCI) existed. Nonetheless, concerns have been 4 raised by the USPTO that there is the potential for a 5 perceived OCI by the public, especially for those 6 corporations who compete with Incyte in the filing of 7 biotechnology patent applications. 8 As a result of these concerns, we agreed to 9 two actions; one, to develop an OCI Mitigation Plan and, 10 two, to make a public statement, this statement, regarding 11 our activities. The OCI Mitigation Plan was delivered to 12 the USPTO on the 2nd of April. That plan fully identifies 13 and addresses all possibilities for a perceived OCI in task 14 areas pertaining to support for biotechnology patent 15 examination. Our plan establishes Organizational Controls 16 in order to isolate the SETA contract from all the other 17 contracts being performed by SAIC, both physically and 18 operationally. Further, the plan establishes an 19 Information Control System consisting of Physical, Data, 20 and Administrative controls in order to protect sensitive 21 information from public disclosure. And, all SETA 22 personnel assigned to any biotechnology-related task will 23 receive special, targeted OCI training in order to 24 reinforce the execution of this Mitigation Plan. We 25 believe that this Mitigation Plan addresses USPTO OCI 26 concerns, ensures the confidentiality of USPTO information, 27 and has removed any remaining OCI obstacles for the 28 continuing PTO-SETA support. Thank you. 36 1 MR. LEHMAN: Thank you. 2 Are there any questions? 3 Ms. Kepplinger, do you have the other 4 witnesses who signed up? 5 Next I'd like to call Barbara Luther, please. 6 MS. LUTHER: Good morning, gentlemen and, 7 Mrs. Linck. I am Barbara Luther. I am here to represent 8 the biotechnology industrial organization. My address is 9 at Incyte Pharmaceuticals, 3174 Porter Drive, Palo Alto. 10 These comments are the summary comments from Bio patent 11 committee meetings. There were a number of other 12 suggestions that I won't mention today that we could not 13 come to agreement on, but we certainly appreciate the 14 opportunity to come and talk to you again. 15 MR. LEHMAN: This reflects the Bio Industry 16 Trade Associates Patent Committee? 17 MS. LUTHER: Right. 18 First of all, I think we have to look back to 19 October 1994 and applaud you on the progress made. We are 20 also here to propose future goals and even some specific 21 suggestions on how to accomplish them. 22 Biotechnology comes of age: 11 major 23 biotechnology products qualified for marketing in the last 24 year alone. The list of problems solved by biotechnology 25 drugs grows longer. 26 First of all, recombinant insulin, a protein 27 manufactured for diabetics; recombinant human growth 28 hormones to help children grow to normal size, recombinant 37 1 hepatitis B to protect people from liver destruction; 2 recombinant t-PA to stop heart attacks in their tracks and 3 return people to productive lives in days, as opposed to 4 weeks; recombinant Factor VIII to stop the crippling 5 bleeding in hemophiliacs -- a pure drug which alleviates 6 obtaining Factor VIII from the commercial blood supply, 7 which formerly resulted in many bleeders getting hepatitis, 8 then AIDS; Epivir for AIDS; Doxil for Kaposi's sarcoma; 9 Vitasert for CMV retinitis. 10 To fund this tremendous research effort, 11 companies have needed the promise of getting a patent to 12 protect their efforts and give them enough money to pay 13 back their investors. The companies themselves only seek 14 to patent isolated and purified DNA and proteins, as they 15 exist in our laboratories after the hard work of our 16 scientific research teams. Patents did not cover DNA in 17 your own body. Patents are only wielded against 18 competitors which try to profit from our inventions. 19 Many more biotech products are in the works 20 and in the news. The early drug pipeline has literally of 21 hundreds of drugs in Phase I and Phase II studies. 22 Many new products and services are used in the 23 emerging biotech market arena. It is highly computerized. 24 As you heard from our gentleman at Maas Parr, now 25 sequencing machines can prepare sequences for easy loading 26 into computers. Computers programs have been designed to 27 compare sequences and permit the scientist sitting at the 28 computer terminal to make realistic predictions about the 38 1 structure and function of recombinant protein. 2 Computer-modeled drugs are going into clinical testing. 3 Companies interested in leveraging 4 biotechnology expertise can turn to a number of genomics 5 companies, including my employer Incyte Pharmaceuticals, to 6 out-source their biotechnology discover. We genomics 7 companies provide new therapeutic proteins, diagnostic 8 probes, human receptor targets for drugs. Breast cancer 9 and obesity genes are examples of Incyte's competitors' 10 successful collaborations. But enough of our success, 11 let's look at the PTO accomplishments in detail. 12 In October of 1994, you took quite a beating 13 over the utility issue. At that time, you insisted that 14 you wanted to give us quality patents we could take to the 15 bank. And you instituted changes. You issued guidelines 16 and legal rationale for changes. The PTO followed up to 17 assure that these changes would be implemented by the 18 examiners. You drafted explicit examples which showed 19 examiners how to reject claims but also how to allow claims 20 when we met the appropriate standard of proof. You 21 implemented quality control and achieved value on the 22 utility front. 23 Another great accomplishment that is just 24 beginning, the patent office has been training a new crop 25 of examiners to propose allowable claim language on the 26 first action. This revolutionary change, if sustained by 27 the office, could save biotechnology companies millions in 28 patent attorney fees, and these millions will be used to 39 1 further research. This would eliminate unnecessary 2 narrowing of claims. It would eliminate prolonged file 3 history battles. It would eliminate some of the problems 4 we have encountered in celebrated cases such as Genentech 5 versus Novo Nordisk. 6 A third accomplishment, you also appointed 7 biotech practice specialists, who monitored the performance 8 of individual examiners and specifically retrained 9 examiners. Though I haven't had to resort to them yet, 10 they being there holds great promise in helping assure 11 quality and fairness throughout the biotech examining 12 group. 13 And two weeks ago you demonstrated, for the 14 first time in my memory, your sequence searching 15 techniques. You have developed a fine system as the art 16 and science have evolved. You are being complimented on 17 bringing your sequence source problem to the attention of 18 the public so we as customers can help. 19 Future issues. Sequence searching. This is a 20 really big problem. We hope to continue to work with you 21 to build a fair and efficient searching technique. These 22 are difficult and complicated computer techniques and 23 bioinformatic techniques. They can help make your job a 24 lot more efficient. We would like to see you craft an 25 internal quality standard or set of standards for 26 searching, much as you did with utility. I will provide 27 some detailed suggestions on that in a moment. 28 Another big issue that you need to address is 40 1 the scope of claims for patents with DNA and amino acid 2 sequences. It varies considerably. Some examiners present 3 precise sequences. Others will allow broader scope. In 4 fact, it is relatively easy to substitute DNA codons, as 1 5 to 6 codons specify each amino acid. Similarly, 6 conservative substitutions of amino acids have been known 7 for more than a decade; one amino acid with another 8 comparable one, so that there is no change in structure or 9 function. This state of recombinant technology art is such 10 we must change the human sequence for more efficient 11 cloning in bacteria. 12 There are computer programs now which will 13 convert an efficient human gene into an efficient E. coli 14 gene. If the examiner limits the claim to a particular DNA 15 sequence, competitors could avoid such a narrow claim. If 16 we are to have a patent we can go to the bank on, we need 17 adequate claim scope. We propose continuing discussions on 18 this topic. 19 Section 112 issues: Sometimes it seems like 20 examiners have converted their utility rejects from "It's 21 incredible that you could do this" to "You haven't taught 22 me how to do it." If we cannot get a patent, we'll still 23 be delayed years in getting biotech patents. The 24 development of drugs today is literally an industry, with 25 regulations that stipulate when and how to do certain 26 testing. Pharmacologists are familiar with establishing 27 dosing, whether for antibiotics or other recombinant 28 proteins. Obviously, the examiners had experienced drug 41 1 developers, and so those topics are largely unknown to 2 them, but not to those familiar with the current state of 3 the art. Again, this is something we need to work on 4 together. Perhaps I should go back to B.I.O. and see if 5 members would have scientists to provide seminars on these 6 topics. 7 Now, for some specific steps. Step one, I 8 don't know if you realize this, but the examiners are 9 asking the inventors to send in any sequence that appears 10 in the patent application literally. If we have a new 11 hoopla of human growth hormone factor and we show a 12 comparison in the figure, they tell us to send in a 13 sequence listing, the sequence for the known human growth 14 hormone factor that already appears in Genbank and is no 15 doubt in your database already. And we see these day to 16 day. We know that your database may be very redundant. We 17 could work together on this. If we simply cite to a known 18 database location, called G.I. number for Genbank, that 19 should be sufficient. 20 Step No. 2, because of this redundancy, we 21 know that you already have a redundant sequence. So as the 22 gentleman from Maas Parr said, we need to work on that 23 redundancy. It may be double, triple, quadruple the 24 sequences it may be already. And from the pace the 25 research is going and the pace it looks like, you'll be 26 receiving a lot more sequences in the future. Streamlining 27 your database will be essential. 28 No. 3, convene a meeting on the sequence 42 1 analysis technology with industry, academic and government 2 experts. Definitely Maas Parr has some new information. 3 And hopefully working together, we can get to the goal of 4 helping you search. It would be helpful if we started an 5 information exchange, which sounds like it could be 6 beneficial, and I know you'd like to continue that on an 7 open forum, too. 8 The state of the art does appear to offer 9 software tools and help subjects to come up with all the 10 relative but not the superfluous prior art. From the PTO 11 presentation, it appears likely that inefficient searchings 12 is excessive and bogs down examiners with excessive 13 printouts and diskettes. 14 Although such a meeting should take place as 15 soon as possible, there's a big bioinformatic meeting in 16 Baltimore in early June. Maybe your schedule would allow 17 so we could help arrange or follow on or present a meeting 18 of that. All the experts are meeting in the neighborhood 19 at that time. 20 Step No. 4. Draft internal quality standards 21 for sequence searches by patent examiners just like you did 22 with the utility examiners. We should have an appropriate 23 search standard, then we'll get something we can go to the 24 bank on. Providing these examples to the public will 25 encourage biotechnology inventors to ask for rational 26 coverage and support it with the searches they routinely 27 perform. 28 And Step No. 5, after we get that expert input 43 1 and quality standards, let's reevaluate the cost to do 2 searches and examiners' time. And then that will be a good 3 time to have a meeting and again request the investors' 4 input to help. That will be the time to discuss additional 5 approaches, such as providing tailored searches. 6 Again, we applaud the progress that you've 7 made. We really appreciate the way you've been working 8 with us, and we look forward to cooperating on these 9 initiatives in the future. 10 MR. LEHMAN: Thank you very much. 11 Any questions? 12 I have a question relating to Incyte. I 13 appreciate what you had said about the utility guidelines, 14 and obviously as ICF being the essence of utility 15 guidelines, that you have to show that there is going to be 16 some utility, for example, in the case of biotechnology 17 inventions, that relates to the pharmaceutical, ultimately 18 what would be the disease that would be cured and so on and 19 so forth, and you have provided four mechanisms far short 20 of human clinical trials to show that utility, and so on 21 and so forth, but you've indicated Incyte is a -- I think 22 you referred to it as a genomic -- what was the word? 23 MS. LUTHER: Genomics company. 24 MR. LEHMAN: Genomics company. And the 25 applications that we're talking about are largely 26 applications that are flowing out of genomics companies 27 rather than other kinds of biotechnology companies. And so 28 first I'd like you to explain for me a little bit more what 44 1 the difference is between a genomics company and other 2 people in the industry. Can you do that? 3 MS. LUTHER: We have Randy Scott here who is 4 our vice president and chief technology officer who does an 5 excellent description of that. 6 MR. LEHMAN: He's on the list? 7 MS. LUTHER: Yes. 8 MR. LEHMAN: I'll ask him then, since he is 9 going to be preparing an answer for that question. And 10 maybe I should just wait until he comes to testify to 11 follow up on that line of questioning. 12 MS. LUTHER: We believe by issues of state of 13 the art bioinformatic and using the biology, we do know 14 about the sources of its sequences and the multitude of 15 comparison that we can make among all different libraries 16 that we have representing multitudes of different tissues 17 in the human body, that we can come up with a good argument 18 for utility for many of the sequences. 19 MR. LEHMAN: Thank you. Are there any other 20 questions? If not, we'll move on to Randy Scott. 21 MR. SCOTT: Good morning, my name is Randy 22 Scott. Hopefully a good part of my comments will address 23 exactly some of your questions and issues. I would really 24 like to structure this in two different parts. One, first 25 of all, is just to step back and talk a little bit about 26 what we think is the future of the vision of this whole 27 field of bioinformatics and genomics because I think if you 28 want to deal with practical problems today, it's incredibly 45 1 important to be envisioning what this world is going to 2 look like in two years so that you don't have to be solving 3 today's problems two years from now and realize that all 4 you've done is put yourself behind the eight ball in terms 5 of a whole new set of problems. So I'd really like to talk 6 a little bit about the vision that Incyte has and the way 7 that we believe the world is going to look in a few years 8 because of bioinformatics, genomics, which is now a very, 9 very important and vital sector of the biotechnology 10 marketplace. 11 To sort of give you an introduction, I'd like 12 to go back about a year ago when I was with my son, who is 13 an 11-year-old, and we went to the movie "Apollo 13," and 14 we were sitting there watching the point in "Apollo 13" in 15 which the projectory is off coming back into Earth, and so 16 they had to change that trajectory slightly. And we all 17 sort of think of NASA as this fantastic, high-tech 18 organization, and I was amazed in the movie, which was 19 quite accurate, in which the -- one of the astronauts 20 pulled out a slide rule and calculated the trajectory on a 21 slide rule. 22 Of course my 11-year-old son punched me and 23 said, "Dad, what's a slide rule?" because he's been 24 playing on computers now for most of his life and using 25 calculators, and so he had no recollection of the level of 26 complexity that he has to work with in today's environment 27 versus what they had in the 1960s. And it struck me right 28 as he asked that when I was growing up in the '60s, my 46 1 parents and everybody at that time used this phrase, "Why 2 is it that we can put a man on the moon, but we can't cure 3 the common cold?" 4 And it sort of struck me right at that moment 5 that the reason is very simple. The complexity of the 6 human body and of biology and disease is in fact millions 7 of times greater than the complexity of putting a man on 8 the moon. So I think what we've looked at in medicine as a 9 very high-tech field and sector for a lot of years and 10 we've patted ourselves on the back, in fact, we at Incyte 11 believe that we're in the very, very early stages -- in 12 fact we're not high tech at all. 13 Imagine trying to play a game of cards with a 14 deck of cards in which you only have five of the cards to 15 play with. There's not many games that you can get 16 involved with. And in fact what Incyte is doing and what 17 the worldwide genome effort is doing is first and foremost 18 just trying to take apart the human body and figure out 19 what are the 100,000 genes that make us us. And in fact 20 the mechanic of your car is not going to do very well if he 21 doesn't even know all the parts that exist in your car. 22 And so until we have all of those parts and all of those 23 pieces in place, our fundamental understanding of biology 24 is always going to come far short of the absolute 25 complexity with which it exists. 26 Just to give some examples, most of the drugs 27 that are on the market today were probably developed 18 or 28 19 years ago at which time maybe only a few hundred genes 47 1 were even identified and characterized to any great extent. 2 As of 1990, we estimate there are probably 3 only 2,000 genes within the genome for which we had a 4 complete structure, much less much in the way of downstream 5 biological information. And yet in the last five to six 6 years, because of the advent of hybrid dna sequencing 7 technology, we've now been able to identify at Incyte 8 what's probably approaching 100,000 genes within the human 9 genome. In fact our expectations are that the level of 10 complexity we're still seeing, that there may be as many as 11 150,000 genes or more within the human genome. 12 When you think for a second that means in five 13 to six years time, we're going to go from 2,000 to 100,000. 14 That's almost two orders of magnitude increase in the 15 number of genes we now have available for researchers to 16 study. And so I think throughout the world, that's been 17 brought up as an enormously complex issue. We have to be 18 able to deal with these types of problems, not with paper 19 lists but rather with very sophisticated computer analyses 20 to be able to help us to examine that data. 21 But in fact that's only the beginning of the 22 complexity. Because the second order of complexity is not 23 just having a list of all 100,000 genes in the genome, 24 which in fact we will have and many different companies and 25 the public domain will have within the next few years, but 26 knowing which of those genes is expressed and turned on and 27 off in different phases of biology. 28 Now, if you take my prediction of 150,000 48 1 genes within the genome and you just take the position that 2 each gene is either on or off in any given cell or tissue 3 or biological environment, that already gives you any 4 number of possible expression states of the genome of 2 to 5 the 150,000th power. While that number is not infinite, 6 but it might as well be from our practical aspects of how 7 we can look at that data and yet that's precisely the goal 8 of companies like Incyte Pharmaceuticals, is to understand 9 not just the list of the 800,000 genes in the genome or 10 more but also which of those genes -- and typically it 11 maybe anywhere from 10,000 to as many as 30,000 genes that 12 are expressed in a single biological setting. 13 We don't want to know just those genes that 14 are expressed in the central nervous system. We want to 15 know the genes that are expressed at a particular time and 16 state of the central nervous system that might be 17 associated with disease. So as you begin to think about 18 the complexity that I've described here, you begin to 19 realize that computer tools and informatics are absolutely 20 the only way that we'll be able to understand this 21 complexity, so they're absolutely pivotal to how we move 22 forward as an industry and I think how the patent office is 23 going to be able to deal with the increasing complexity of 24 pharmaceutical inventions because we're going to be 25 providing you with data not from four or five genes, which 26 we've measured by a lyse analysis that predicts that one of 27 those genes is associated with, for example, Alzheimer's 28 disease or rheumatoid arthritis, we're going to be 49 1 providing you with electronic, computerized information, 2 databases that suggest we've done exhaustive studies of 3 these 10,000 genes, and we can show you which of these 4 10,000 genes are up-regulated during this biological event 5 and which are down-regulated. And we can correlate that 6 with structure and with function. And we can make very 7 good predictions about which of those molecules are going 8 to be important, not just to cure the disease but to 9 understand the disease to begin with, which we think is one 10 of the primary reasons for our existence, is to be able to 11 identify markers, genes that could be used either in a 12 clinical setting, in a diagnostic setting to look at a 13 particular disease or as clues for the pharmaceutical 14 industry to be able now to exploit some cured disease. 15 Genomics, in fact, we believe -- or I 16 believe -- is simply a commercial term that is risen to 17 describe what the public sectors called the worldwide human 18 genome effort. There are two different parts to what I 19 believe are genomics. The first part is hytrobid 20 [sic]tools for studying biology thousands of genes at a 21 time. And in the early days most people recognized that as 22 hytrobid [sic] DNA sequencing. We have come up with some 23 smart ways to do the DNA sequencing, some very sufficient 24 ways to do the DNA sequencing that allow us to identify 25 expressed genes in a very, very rapid rate. But that's 26 just the beginning. 27 Once we have each gene, when we sequence it 28 the second time, we can identify the gene that we've 50 1 already seen, and we can be able to compare where we've 2 seen that gene before. And so we're not just doing this at 3 the genomic level. We're doing it at the level of biology. 4 We're understanding which gene is terminal and in a 5 prostate cancer sample versus a normal part of that same 6 prostate just a few millimeters away, being able to do an 7 exhaustive comparison of those two samples of which genes 8 are up-regulated and down-regulated which leads us to a 9 fundamental understanding of which genes are involved in 10 the prostate cancer process. So that's the first step 11 deciphering DNA sequencing. 12 As you may know, there are many, many 13 technologies that are being developed now which will allow 14 us to take any excessive genes that we can identify on a 15 database, put those onto literally a silicon chip or a 16 glass slide and to be able to rate thousands of clones and 17 now to be able to process multiple biological samples 18 across those thousands of clones. 19 So now after we do our initial studies, we can 20 take a thousand genes that we believe are involved in 21 prostate cancer and we can take biological specimens from a 22 variety of different individuals, and we can test that 23 hypothesis, and we can demonstrate further with clinical 24 samples which of those genes are regulated only in a 25 portion of prostate cancers and which are regulated in all 26 prostate cancers. And in fact in many diseases, we'll find 27 that there may be no single gene that correlates with that 28 disease but rather patterns of different molecules of 51 1 different family's members which correlate with that 2 particular disease, and that will allow us to explore that 3 information further. In fact we believe that that's the 4 future of diagnostics, is not one molecule at a time but 5 thousands of molecules at a time. 6 How would you look at that data? How do you 7 look at data derived from searching 200 patient samples 8 across a thousand or more molecules? 9 A simple list or even a computer printout 10 profile is not enough. You have to be able to use very 11 large, very sophisticated relational databases to store and 12 allow you to search and retrieve just the bits of 13 information that you want to utilize for the particular 14 studies or for the particular analysis that you want to do, 15 which brings me to what we believe is the second part of 16 the genomics effort is not just creating a hytrobid tools 17 for analyzing biology thousands of genes at a time but 18 information tools and databases, tools for mining those 19 databases that allow us to get the unique data sets that we 20 need very, very quickly. 21 This is not unique to biology. It's not 22 unique to bioinformatics. This is a wave that is taking 23 over the entire economic sphere of our country -- database 24 mining. As we're able to store larger and larger amounts 25 of information, the tools now for mining those data sets 26 and being able to create usable information are enormously 27 valuable. And I would submit that one of the things that 28 the patent office needs to be thinking very seriously about 52 1 is moving towards electronic filing. 2 As we file large number numbers of patents 3 with large numbers of CDC sequences in those patents, that 4 paper stack cannot even begin to explain the complexity of 5 the database from which it was derived and our ability to 6 search that database to compare the sequences against each 7 other and get results within seconds, not hours or days. 8 And our ability to compare where each gene and each 9 individual biological setting is as expressed, that 10 information is only going to explode over time. 11 Our database currently houses several 12 megabites of information, but we see that whole field 13 expanding by orders of magnitude in complexity over the 14 coming years. You never use all of the data all the time. 15 No biologist is trained to be able to sit down and use all 16 of that information. You're only able to mine that 17 information for the observations that you want to follow up 18 on. 19 So the implication for biologists, and I think 20 for many scientists, is that our whole world is going to be 21 changing over the next several years. We don't think 22 bioinformatics is simply a tool that a few specialists will 23 be using. We believe it will be the primary basis in which 24 biologists will be able to retrieve, do searches, analyze 25 and set up their individual experiments. 26 At Incyte, we draw a lot of analogies for our 27 business to the computer industry with its emphasis in both 28 hardware and software. For us, hardware means that first 53 1 part of genomics, hytrobid tools for analysis. Our 2 hardware is DNA sequencing and tools for analyzing DNA 3 samples or sequences at a very, very rapid rate. 4 Today at Incyte, we're sequencing between 5 6,000 to 7,000 cDNA clones per day. That's an enormous 6 array of sequencing. In fact, we believe that there's 7 probably a corollary between Moore's law and the 8 semiconductor industry and what we're doing today in DNA 9 sequencing. That is, we double the rate of sequencing at 10 Incyte and have doubled it almost every six months since we 11 started the company down this path in 1991. 12 If I'm correct with that, that means from 6- 13 to 7,000 sequences that are being analyzed a day, we'll 14 have to process 12,000 or more in six months, 24,000 per 15 day six months after that. And algorithmic expansion, of 16 course, means we have to build all the tools to do that. 17 In fact I would argue that we've done that in a very, very 18 satisfactory fashion. So let me just give a brief 19 explanation of the bioinformatics tools that we use in 20 house. 21 And I think the philosophy or the theory comes 22 back to database mining: Don't try to create a large set 23 of data about every single molecule that you're going to 24 look at. Select your questions very carefully. We tell 25 our scientists all the time, "Be careful what you ask for 26 because you may get it." So if you ask for an 27 all-encompassing report of every sequence, you will get 28 that, and you will be so overloaded with information, you 54 1 could never analyze it, and the whole system will grind to 2 a halt. 3 So in our system we take some very 4 straightforward approaches. First, do the easy searches 5 first with the fastest tools and the easiest tools, ones 6 that don't take the computer intensity of time. So within 7 our process we can search and analyze about 6,000 sequences 8 a day in a pretty straightforward process in which, first, 9 we take every one of those sequences, and we go through a 10 series of masking programs, and we eliminate all the things 11 that we know we're not interested in like repetitive DNA 12 elements, alpha sequences, things that are not useful 13 information because they've been sequenced many, many times 14 over -- there's nothing inventive about that process. But 15 there's a key reason to do that is because when we now turn 16 around and search our sequences against the public domain, 17 if we don't remove the so-called junk sequences out of our 18 information, what we'll find is that our junk will match up 19 with public domain junk and you end up with junk squared. 20 And that's not good. So you want to get rid of at least -- 21 have at least one data step in your comparison set in which 22 you've removed all that and we go so far as to mask things 23 like dinucleotide or trinucleotide repeats, so we get down 24 to the sequences, just the things that we're really 25 interested in. 26 Then we go -- 27 MR. LEHMAN: How do you distinguish between 28 the public domain and what is yours? 55 1 MR. SCOTT: Well, that actually comes in a 2 second phase of the analysis. In the first phase of the 3 analysis, we've worked with numerous academic 4 collaborators, probably many of the same people that you 5 would find available for consulting roles to set up 6 databases of all of the mitochondrial genome, the 7 ribosomial genome, almas, repetitive almas. We've 8 established what those outside collaborators -- Temple, 9 Smith and Boston University was one of the founders of 10 Genbank, Jean Claude Raboret [sic] of the CNRS and many, 11 many other scientists' parameters that are fairly widely 12 agreed upon in the bioinformatics community to be able to 13 eliminate those elements first. 14 MR. LEHMAN: Well, is your distinguished -- is 15 what you distinguished between being the public domain and 16 what is yours then simply what has previously existed -- 17 MR. SCOTT: That's -- 18 MR. LEHMAN: -- prior to your -- 19 MR. SCOTT: That's correct. That comes in the 20 second stage -- phase of the analysis. Once we've 21 eliminated those, probably for our purposes, more 22 uninteresting molecules, we do a very fast search, which is 23 called a blast search. That's -- we can set up multiple 24 sequences processed in batch mode. It's not the best 25 searching algorithm, but it is the fastest. And you can 26 search huge volumes of sequence information very, very 27 quickly. And the only question we're asking is: "Do we 28 have an exact match to Genbank primate?" 56 1 We don't need to look at C elegance. We don't 2 need to look at yeast. We don't need to look at bacteria 3 because we know that our feed sequence comes from humans, 4 and so we simply look first for an exact match in a very 5 small portion from Genbank. And over half of our sequences 6 are immediately caught through that process as exact 7 matches. And so we can eliminate those right off the bat. 8 And then we can take the remaining sequences 9 and just start to do more in-depth searches. We can say, 10 "Okay. Now let's look for homology in the human primate 11 database or" -- first what we do is we go to rodent 12 database and we say, "We have something that is very 13 closely aligned to a mouse or rat sequence" because, again, 14 that allows us to identify "This looks like the human 15 correlate to a known mouse or rat gene." 16 Then we can go back and use more sophisticated 17 tools to look for homologies. And I would suggest that 18 that in fact is one of the approaches that the patent 19 office would want to take, is to do the easy things first. 20 Come up with ways that will allow you to screen and sort of 21 immediately eliminate half of your work load or more. Then 22 only come back to the narrow subset of things for which you 23 want to go in much greater depth and use the more 24 sophisticated algorithms such as the Smith-Waterman 25 algorithm I believe is running under Maas Parr instruments 26 to do that second process. 27 So by going through these stages, Incyte 28 actually does, I think, a very high-quality job at cleaning 57 1 up, comparing things to the public domain -- first simply 2 identifying what's an exact match and what is not and then 3 going through a process of trying to annotate those 4 molecules that are not exact matches to show what are they 5 most structurally related to and then from then on we can 6 select from molecules that we think are highly related or 7 are highly up-regulated in certain biological cases to 8 really hone in on the individual value of those molecules. 9 That system is now in place and has been 10 running since September of 1995. So well over six months 11 it's actually been put in place in collabortion with the 12 collaborators, as I mentioned, many of them who are free to 13 consult and work with other groups and in fact we would 14 quite encourage you to have some of those groups in because 15 our daily life is surrounded by, you know, the process of 16 information. We don't do it -- if we don't do it in a 17 rapid time frame, we won't exist because we can't present 18 this information in our database for pharmaceutical 19 partners. So I think I can probably going on for some -- 20 MR. LEHMAN: I'd like to ask. What is -- how 21 long has Incyte been in existence? 22 MR. SCOTT: Since April 1991. How far along 23 was our corporation. 24 MR. LEHMAN: And do you have sales right now? 25 MR. SCOTT: We do have sales. We have six 26 database subscribers, which are some of the six -- some of 27 the best pharmaceutical companies in the world -- Pfizer, 28 the pharmacy at Upjohn, Nomen Ordiz, Johnson & Johnson and 58 1 Abbott Pharmaceuticals. Our current annual revenue rate 2 from sales of the database is in the range of $25 million 3 per year. So in fact as I look more and more at the 4 software and informatics sector, there's an awful lot of 5 software and database companies out there from other fields 6 that don't yet add up to 20- to $30 million in revenue. 7 Our marketplace, we believe, is the top 50 8 pharmaceutical companies in the world, to get this 9 information into their hands as rapidly as possible. They 10 have accessed our intellectual property portfolio. All of 11 those companies have nonexclusive access to the use of our 12 sequences for research purposes and then exclusive access 13 on a first-come, first-serve basis to individual gene 14 product patents that might derive from our work. So in 15 fact we are creating sort of a unique resource for the 16 pharmaceutical world. And in many ways, patents are almost 17 more important to us because what we're selling is ideas. 18 MR. LEHMAN: When you talked about how you do 19 your work, you said that you start out comparing your 20 genomic mapping to what's in the public domain already, and 21 you identified that as basically work that's already been 22 done by other people, largely academic scientists. 23 What kind of reciprocal relationship is there 24 between you and them? Let's say, you know, the scientist 25 at Harvard wants to follow up on research that he has done. 26 Do you have relationships with them? Do you share 27 information with them? And if you do, what kind of terms 28 and conditions do you apply to them? 59 1 MR. SCOTT: We actually do so, but we do so 2 indirectly because we don't really regard this database as 3 ours. We regard it as our pharmaceutical partners. So 4 almost all of our pharmaceutical partners in fact have 5 programs by which they interact with the academic 6 community. We quite encourage that. So while we do have 7 some individual academic collaborations -- one, for 8 example, is an institution by the name of Mayo Clinic by 9 which their researchers at the Mayo Clinic can get access 10 to full-length clones that derive from Incyte database 11 through collaboration there. Most of our academic 12 collaborators go to the pharmaceutical companies themselves 13 because it often turns out that they, like we, are looking 14 for capital to fund their research. And capital, you'll 15 find, is not within the biotechnology companies but within 16 the large pharmaceutical companies that are really 17 developing and dropping off of this business. 18 So we think in fact there is a tremendous 19 synergy here. In essence, the pharmaceutical industry is 20 paying Incyte to very rapidly try to create some of this 21 information ahead of time to come into the public domain. 22 And so by funding that program, they're both accelerating 23 the pace of research, and they, as they look through the 24 database and work with their own academic collaborators, 25 can identify individual molecules that they now want to 26 pursue more aggressively and encourage -- 27 MR. LEHMAN: Would you say that what -- 28 primarily what you're doing is creating a database on a 60 1 human genome? 2 MR. SCOTT: Absolutely. I think Incyte is an 3 informational technology company. 4 MR. LEHMAN: And that database is used as a 5 tool, or when you speak of "tools," are you speaking of 6 something different from the use of the database? 7 MR. SCOTT: No -- well, I think the tools are 8 many fold. They're not just the database. They're -- 9 they're the database, the proprietary software we've 10 created around the database to be able to help them mine 11 hat database and the actual molecules themselves. Because 12 what's unique about Incyte is that not only can you search 13 through the database to find a molecule, it's not coming 14 out of an archive like Genbank, where now you have to go 15 track the original reference source to find out if they're 16 even willing to give you material. Our pharmaceutical 17 partners immediately have access to every gene for which we 18 have a cDNA stored away that we provide within days after 19 that. So they're also requiring access to new receptors, 20 new targets, and they have access also to the intellectual 21 property portfolio and the molecules for which we're filing 22 patents on. 23 MR. LEHMAN: Are you filing patents in other 24 countries besides the United States? 25 MR. SCOTT: In other countries, we're filing 26 patents on full-length clones only and being very selective 27 about how we filed on EST sequences, primarily because of 28 the publication rules because we believe that our economic 61 1 value and what we're being paid for primarily by the 2 pharmaceutical industries is for lead time. And that's 3 true of almost all pharmaceutical research. They know 4 that ultimately every pharmaceutical company will have the 5 same genes and the same products, but if they could get a 6 lead time past their competitors, then they can move very, 7 very quickly. So anything that of course puts all of our 8 information into the public domain simply means that the 9 capital resources for doing this will dry up because -- 10 MR. LEHMAN: One of the -- you -- sounds to me 11 like fairly accurately compared what you're doing to the 12 more classical software industry and that in fact there is 13 sort of a merger going on here between software and biotech 14 industries and what you're doing, and of course the 15 software industry has for a long time kept the light on 16 multiple forms of intellectual property protection and not 17 simply patent protection, to rely on copyright protection 18 as well. 19 Copyright does not protect what is just pure 20 information, however, and thus the Supreme Court has 21 decided that pure information in the form of database -- a 22 database -- even with that database is very difficult to 23 create and requires a lot of investment that has no 24 protection. 25 This is a global problem, not just a U.S. 26 problem. And in fact the European commission is in the 27 process of promulgating directive to all the members states 28 to the European union to enact laws which would provide a 62 1 new kind of intellectual property protection for -- 2 specifically for databases, even though they don't meet the 3 test of what we call originality in our copyright sense. 4 MR. SCOTT: Right. 5 MR. LEHMAN: But it would be more of a 6 copyright-like type of protection as opposed to a patent 7 type of protection. Really protects the information in the 8 database and from unfair extraction or unfair exploitation 9 of that information. 10 Would that be something that would make your 11 life -- and it's easier -- it would be easier to get, by 12 the way, because then you wouldn't have to apply for 13 patents and, you know, spend all the money and spend a 14 couple years trying to deal with the issues that we're 15 talking about here today. Would that sort of protection be 16 something that would be welcomed for you? 17 MR. SCOTT: I think on the databases as a 18 whole, certainly would allow me to sleep easier at night to 19 know that we have some level of protection should that 20 database ever be stolen or, you know, other sorts. 21 MR. LEHMAN: Particularly important in a 22 computerized environment because in the old days when you 23 had databases that were contained on paper basically, you 24 know, just physically there was a certain protection. You 25 know, you couldn't -- 26 MR. SCOTT: Right. 27 MR. LEHMAN: -- you couldn't go in and copy 28 stuff out of it but obviously now where databases are 63 1 available electronically, you can -- if you can enter into 2 the database and enter into the security that may have been 3 provided by incryption, you can, you know, extract 4 information from it, and of course nobody knows where you 5 got the information. 6 MR. SCOTT: I think absolutely that would be 7 welcomed. It probably wouldn't change the patent stance 8 because, I mean, we really are a merger between computer 9 science and microbiology, and the value, the information 10 now leads you to which molecules you want to take forward 11 into product relationships, whether it be diagnostics or 12 therapeutics. So I don't think it would be looked at as a 13 solution to the patent dilemma of the patenting of 14 thousands of molecules at a time. And the other thing, 15 this dilemma that we're now seeing in the patent office 16 with the number of molecules that are being discovered is 17 really -- it's the embarrassment of riches or the fact that 18 such a wealth of information can be created, everyone knows 19 it's so enormously valuable in such a short time, I think, 20 is much of the -- much of the shock. 21 So still individual molecules are something 22 that we would continue to pursue. We have a whole group of 23 biologists that we call the Group in Silico, biologists -- 24 for years people have done in vitro and in vivo biology, we 25 think the state of the art now is so good with 26 bioinformatics that -- certainly not for all, not even for 27 most, but for a large number of molecules there is a clear 28 structure-function relationship that's been hammered out in 64 1 the laboratory. And now we can identify members of this 2 family. And even more importantly than that, it's highly 3 up-regulated in the joints of arthritis patients. So we 4 believe this is a clear-cut observation and be able to 5 provide this molecule in a way in which the pharmaceutical 6 sector can both have intellectual property protection on 7 the molecule and be able to pursue those aggressively, I 8 think, is going to be tremendous value to society. 9 I think people will be shocked at the amount 10 of knowledge of disease processes, of the most complex 11 disease processes we haven't been able to get at by doing 12 biology one at a time like cancer, AIDS, writing different 13 viral settings as well as arthritis. These are complex 14 diseases probably with multiple etiologies, multiple 15 genetic makeup and background and will not fall into one 16 disease category. And it won't be until you break those 17 apart into 10, 15 separate disease categories and different 18 treatments and regiments that we're really going to be 19 successful in curing them. 20 So I see tremendous value added over the 21 coming years, not just by Incyte or our colleagues and 22 competitors in the genomic sector but now as pharmaceutical 23 companies really come into this field in a big way to be 24 able to exploit the information and develop drugs. 25 MR. LEHMAN: Are there any other questions? 26 Thank you very much. 27 Next Michelle Duran. 28 MS. DURAN: I signed by mistake. 65 1 MR. LEHMAN: Oh, so you're not planning to 2 appear. 3 Ned Israelsen. 4 MR. ISRAELSEN: Good morning. Thanks for the 5 opportunity to testify at these hearings. My name is 6 Ned Israelsen. I'm a partner with the law firm of Knobbe, 7 Martens, Olson & Bear. I head up our biotechnology group. 8 The address of our local office here in San Diego where I 9 work is 501 West Broadway. 10 And I'd like to focus my remarks specifically 11 on the questions that were proposed in the Federal Register 12 in August and try to provide the views of my law firm, and 13 I'll emphasize that these are not necessarily the views of 14 our clients on the problem facing the patent office right 15 now. 16 Seems to me that it would be unfair to make 17 all biotechnology applicants pay for a problem that's the 18 result of a handful of patent applications. And so at the 19 risk of offending those companies that are filing the 20 patent applications, I think that the cost should be 21 allocated based on the patent office resources that are 22 consumed by these patent applications with large numbers of 23 sequence. 24 I have some personal experience in looking at 25 this issue. I was involved in filing some patent 26 applications for the National Institutes of Health and 27 became rather notorious with thousands of EST sequences, 28 and I'm not saying that I was -- I was very surprised that 66 1 we did not get a thousand way restriction requirement back 2 from the patent office. 3 It seems that the patent office when faced 4 with a large number of sequences that are related only in a 5 general way, already has a tool for -- for dealing with 6 collecting fees commensurate with the amount of work 7 required for examination. 8 Another alternative that I could suggest, 9 aside from restriction requirements, would be a surcharge 10 if considerable resources are going to be consumed by a -- 11 for examination of a particular patent application. A 12 surcharge could be imposed, but I think only after giving 13 the applicant an opportunity to amend the claims. 14 So what I would envision is that the examiner 15 takes a first cut at looking at the claims, looking at that 16 amount of searching that will be required, coming up with a 17 total amount and an explanation of why considerable 18 searching would be required and then giving the applicant 19 an opportunity either to amend the claims or to pay the 20 additional search fee. 21 I have been surprised, again, with some patent 22 applications I filed where a claim that we really don't 23 expect to get is thrown in for sake of completeness that 24 says, for example, an 8 MER or an 18 MER or 15 MER of the 25 full length sequence that we really want to patent. My 26 understanding is that considerable amount of search 27 resources can be used looking for those short sequences 28 when ultimately those have only an ancillary importance to 67 1 the applicant and if given the choice between paying a fee 2 to have a massive search conducted or forgoing the claims 3 for the short sequences, I think a lot of applicants would 4 forgo that broader search. So the search could be 5 attenuated even before it started by giving the applicant 6 the opportunity to amend. 7 Echoing another approach the patent office 8 could take -- and it's already been suggested -- is to keep 9 the database size down by eliminating the need to submit 10 known sequences and unclaimed sequences, in other words, to 11 allow reference, for example, to a Genbank sequence instead 12 of submitting that sequencing electronically once again and 13 providing multiple copies in the patent office database. 14 And again the requirement that sequences of 10 15 bases or four amino acids be submitted is a burden to 16 applicants and I think, again, clutters up the database. 17 As to a long-term solution, I've given a lot 18 of thought to patenting of naturally occurring sequences, 19 and I concluded that the current patent system is not well 20 suited for handling naturally occurring sequences that are 21 cranked out in the thousands or tens of thousands at a 22 time. I know there has been concern expressed by 23 scientists working in a field that the rewards of a patent 24 for -- are not necessarily commensurate with the value 25 provided when a genomics company is allowed to get a patent 26 on a sequence based on a very minimal utility. 27 So one suggestion that I have espoused in 28 recent years is a registration system similar to copyright 68 1 for providing some sort of protection for sequences that 2 come out of massive sequencing efforts, naturally occurring 3 sequences. Perhaps that registration system could provide 4 only a defensive right, again, subsequent patents and could 5 also provide perhaps a transferable exclusion of that 6 sequence from the prior art. So that when a utility is 7 determined -- that fits into the more traditional 8 categories that one would anticipate for a patent such as 9 the use of a particular molecule for treating a particular 10 disease or a particular diagnostic application, the 11 prior -- the exclusion from prior art would provide the 12 ability to subsequently patent that sequence for its 13 biological use or for a use related to the protein encoded 14 thereby and the prior publication, the publication of that 15 sequence subsequent to its discovery and registration but 16 prior to the patent focusing on the biological utility 17 would not provide a barrier to getting more traditional 18 patent protection. So those are my thoughts. If there any 19 questions, I'd be happy to entertain them. 20 MR. LEHMAN: Thank you. You heard me talking 21 about the European database with the previous witness, and 22 that sounds like that's sort of the direction of what you 23 were talking about there. 24 MR. ISRAELSEN: It's a similar idea, yes. 25 MR. LEHMAN: Of course, one of the -- that is 26 more of a copyright-based approach, and that doesn't give 27 you any exclusivity vis-a-vis independent creation of the 28 same database. 69 1 MR. ISRAELSEN: That's correct. 2 MR. LEHMAN: Should it? 3 MR. ISRAELSEN: That's a -- that's a loaded 4 question, and I'm probably not the best person to answer 5 that. With my proposal for a registration system for DNA 6 sequences that are naturally occurring, it would not give 7 exclusive rights. It would not give the right to exclude 8 others but would help preserve the option of obtaining a 9 patent at a subsequent time. 10 MR. LEHMAN: Okay. Are there any other 11 questions? 12 MR. GOFFNEY: Yeah, I wanted to ask him 13 whether he had any criteria when a search -- a surcharge 14 might be required. 15 MR. ISRAELSEN: I don't know the economics, 16 but I would expect that it would be imposed only on a very 17 small percentage of applications so that an application 18 containing -- just to pick a number -- 50 sequences would 19 not incur the surcharge. 20 MR. LEHMAN: Are there any other questions? 21 Thank you very much. 22 MR. ISRAELSEN: Thank you. 23 MR. LEHMAN: We appreciate it. I want to 24 thank the -- particularly the practitioners who came and -- 25 oh, we have one more. 26 Amy Hamilton. 27 MS. HAMILTON: Good morning. My name is 28 Amy Hamilton. I'm from Eli Lilly & Company. My views do 70 1 represent those of the company. Our address is Lilly 2 Corporate Center, Indianapolis, Indiana 46285. 3 I hadn't planned on speaking today until I sat 4 there and heard nothing but restriction requirements, and 5 then I was thankful to hear Mr. Israelsen mention that 6 because I really think that's incumbent upon the PTO to 7 take applications where there's no overriding, nontrivial 8 utility to unify the sequences that clearly the law 9 requires that they be treated as separate inventions. And 10 then when that is done, GACT, and its ramifications will, I 11 think, cause these companies to properly file these as 12 separate applications. 13 I would also recommend some kind of a 14 surcharge for a huge number of sequences. We have thought 15 about, again, some charge appropriately determined by the 16 PTO, something similar to what is done with claims above 17 some reasonable number of sequences. I think 10 is a 18 reasonable number because for a single invention, it's hard 19 to imagine why you would have more than 10 sequences in 20 that application. I'd also like to challenge what was 21 stated in the OT about requiring completeness of 22 examination. 37 CFR Section 105 allows for incomplete 23 examination when there is misjoinder invention or 24 fundamental defects in the application. And I would 25 suggest that cases could be reviewed for a threshold 26 utility before ever making a search, which I think of all 27 things should drastically decrease the amount of computer 28 searching. Then this could be challenged in the courts, 71 1 and we could get the utility question resolved in the 2 courts, as we are going to eventually need anyway. 3 One last thing, 1800 examiners continue to 4 make rejections based on potential methods of making genes 5 and proteins. This form of rejection has been explicitly 6 stated as improper by the CAFC in a series of cases ending 7 in in re: Bell and in re: Dual. This is prolonging 8 examination and taking up valuable examiner time when there 9 is no prior art chemical structure to be used in the 10 rejection, and I would suggest that valuable examiner time 11 could be saved by eliminating this type of rejection. 12 That's all I have. Thank you. 13 MR. LEHMAN: Thank you very much. Did you 14 give us your address where you were located? 15 MS. HAMILTON: I did. I will again. Eli 16 Lilly & Company, Lilly Corporate Center, Indianapolis, 17 Indiana. 18 MR. LEHMAN: You came all the way out. 19 MS. HAMILTON: I wanted to come to the first 20 one. 21 MR. LEHMAN: Thank you very much. We really 22 appreciate it. 23 I believe that completes -- unless someone 24 else signs up, that completes the list of persons who've 25 asked to testify. Does anybody else like to -- 26 MR. O'HARA: I have maybe one comment. I 27 didn't sign up. My name is Patrick O'Hara. I'm from a 28 company called Zymo Genetics, and that's Z-y-m-o Genetics 72 1 in Seattle, Washington, and we're a customer of Incyte's. 2 And I just wanted to make a reality check on 3 the amount of time that you're thinking of that it's going 4 to take to do the searches because we do quite a of 5 searches ourselves, and I think that you're talking about 6 having 200,000 sequences that you have to search and it's 7 going to cost about $24 million and take two years and nine 8 staff years to deal with it, and I think that Randy said 9 they can do 6,000 a day using blast as an algorithm, and 10 the question, I guess, is blast sufficient for the kind of 11 things that you have to do? 12 The person from Peterson & Associates said 13 that very sensitive kinds of searches may not really be 14 necessary for the kinds of obviousness criteria that you 15 have to look at. So it seems like what might be possible 16 to do would be that the PTO simply develop a very specific 17 criteria of what it is that is invent- -- invent- -- sorry, 18 invented with regard to the ESTs, and we at Zymo Genetics 19 do about -- we don't -- we don't do 6,000 a day but we do 20 about 1500 a day. And so if you are able to eliminate 90 21 percent of your problem, instead of being 24 million would 22 be 2.4 million, instead of two years, would be 4.8 months 23 and at about 1500 a day, that would kind of do it, and I 24 think that with blast and related algorithms, plus some 25 post processing of the results, you'd be able to come up 26 with a very nice output that would pretty much 27 automatically determine what it is that you're looking at 28 in most of these EST applications. And I think it would 73 1 decrease the number of staff years that you're talking 2 about as well. So while with regard to the kinds of 3 algorithms that you would have to use, ESTs are parts of 4 genes, and there are specific issues with ESTs that mean 5 that certain kind of algorithms have certain kinds of 6 problems. So, for instance, there's partial overlaps that 7 will give you very different scores than complete overlaps. 8 There are certain issues with regard to accuracy in the 9 database and with regard to the ESTs, and those can cause 10 drastic differences in scores, but, however, those things 11 can all be handled pretty much automatically at a rate of 12 at least 1500 a day. 13 So it seems to me those problems should be 14 manageable by developing criteria very specifically, and 15 then perhaps contract out the searching so that the PTO 16 with the many kinds of searches or your type with 10 17 sequences or so, but when you're talking about those mass 18 ESTs, contract it out, develop the criteria and contract it 19 out. 20 MR. LEHMAN: Thank you very much. I 21 appreciate that. 22 Are there any other people who wish to be 23 heard? If not, again, I want to thank the people who took 24 the time to come here, particularly individual 25 practitioners who obviously could be doing other things, 26 and note that we're having one further hearing on this. We 27 came here to San Diego because this is a center of the 28 biotech industry. We wanted to make it easier for people 74 1 to come here as opposed to come all the way to Washington, 2 but we will be having a hearing on April 23rd in Arlington 3 for anybody who has some comments between now and then 4 wants to get on the plane and come to Washington, but of 5 course we're also happy to receive any further written 6 comments, and we would appreciate it if those would be 7 submitted by April 23rd. 8 A transcript of this hearing will be made 9 available as soon as it's ready as well as all of the 10 comments received. And we think that's going to be on or 11 about May 13th, 1996. And it will be available for 12 inspection in Room 520 of Crystal Park One, 2011 Crystal 13 Drive in Arlington, Virginia. And it will also be 14 available on the Internet through a nonfile transfer 15 protocol. The address is ftp.uspto.gov. 16 All written comments and oral comments made 17 here today will be taken into consideration before 18 implementing policies relating to this issue. And any 19 written comments that are received after the 23rd can't be 20 assured of consideration. I have to say that we really try 21 hard to give people a chance to have input, and then after 22 we're six months into a new policy and somebody says, 23 "Well, you didn't consider this or that," we have not yet 24 had a patent application that gets us access to 25 impressions, so we really need to hear from people if we're 26 going to take their views into consideration. 27 So I'd like to remind everybody that another 28 hearing will be heard on April 23rd starting at 9:00 a.m., 75 1 and that will be in the commissioners conference room, 2 Suite 520, Crystal Park One, 2011 Crystal Drive, Arlington, 3 Virginia. And that concludes the hearing. And I thank 4 everybody that participated and shared their views with us.