Speaker 1  0:00  
Good afternoon, everyone. Welcome to AI and precision medicine, IP and licensing opportunities presented by Autumn. My name is Holly Lundgren autumns online professional development manager and I'll be your staff host for today. All lines have been muted to ensure high quality audio and today's session is being recorded. If you have a question for the panelists, we encourage you to use the q&a feature rather than the chat feature. If you have a technical question or comment, please feel free to use the chat feature for that. Today's session may be eligible for CLE credits. Throughout the webinar, you will you'll see pop up surveys that are simply asking for a yes or no answer. If you do not intend to apply for CLE credits, you may simply dismiss the poll. If you do intend to apply for CLE credits, you must answer the question as it stated, I'll leave it up for about 20 seconds per poll. So you have about 20 seconds to answer if you're going to apply for CLE credits. For more information about CLE eligibility or to find out what states have already pre approved this program, please contact me at h lundgren@autumn.net. I would like to take a brief moment to acknowledge and thank autumns 2020, online professional development sponsors. We appreciate the ongoing support. And now I'll introduce today's speakers. for over 30 years Mary Lou Nakamura has assisted clients with patent preparation and prosecution, copyright registration, and portfolio management. She has particular experience with computer related technologies and systems relating to imaging, video applications, medical systems, mobile applications, virtual currency and other platforms, biochemical analysis and processing including bioinformatics and medical diagnostics, machine learning alternative energy social networks and associate operate associated operations and electromechanical matters with and without a software component. Mary Lou's practice also includes preparing infringement opinions, and conducting due diligence directed to computer and software technologies. She advises on open source software licenses, and other software licensing agreements. Mary Lou is a software engineer with experience in numerous programming languages, data analysis and software application design. She's familiar with various data structures, compression, decompression techniques, encryption technology, Blockchain, digital signal processing, artificial intelligence, neural networks, machine learning, pattern matching, bioinformatics, video gaming and computer network Internet Technology. Dr. Steve Levine is Senior Director of portfolio management and the leader of realistic human simulation initiative at Dassault systems. Samuel similia Sorry. Since 2006, Steve has been responsible for leading strategy and long term vision of enabling science and engineering software to deliver on its full potential. He began his career at Engelhardt Corporation, where he founded and led the simulation group and corporate r&d. Steve subsequently joined a startup molecular modeling company called Bio sim technologies, which grew to become the leading molecular modeling provider for the pharmaceutical, chemical and materials industries. The company eventually became known as accelerace, Inc, and was recently acquired by Dassault systems and rebranded as bio via. During his 16 years at accelerace. He held a number of positions including Director of Product Development, General Manager of materials, informatics, and Senior Director of Corporate Development. Steve holds a PhD in material science and engineering from Rutgers University. And Michael dilling is the director of the Baylor licensing group, the technology licensing team at Baylor College of Medicine, a leading biomedical research institution, and the only private medical school in the southwestern United States. He is responsible for managing the activities of five licensing professionals and industry agreements professional and tube administrators, currently leading efforts at BCM to restructure technology commercialization efforts to increase effectiveness and improve outcomes, guided by the development and launch of an online disclosure application to simply simplify faculty interaction with blg responsible for spearheading collaborative efforts with BCM technology BCM a wholly owned Venture Development subsidiary to identify and catalyze the formation of new startup companies. He serves on the board of directors for met Visa Inc, a new BCM startup company dedicated to the production of monoclonal antibodies. Michael has 15 years technology transfer experience managing a diverse portfolio of biomedical tech technologies and a leading university technology transfer program with a focus on producing licensing outcomes. And without any further ado, at this time, I'm going to turn it over to our distinguished speakers. Welcome Michael, Steve and Mary Lou. Thank you.

Speaker 2  5:19  
Morning all. I am going thank you for that wonderful introduction. We're refreshing this talk from a couple of years ago. So there's some portions that have been updated for this presentation, as we have learned, in the meantime, let's see here. Next slide. So, I will give a quick introduction of the technology. And then we'll move into two case studies, one from industry, Dr. Steve Levine will present that portion and then following that will be the from the T T ellos. perspective of the university licensing office and that portion will be given by Dr. Michael dealing. Next slide. Just very brief introduction, so you can hear our voice with our face. I'm a patent attorney from Hamilton, Brooks Smith and Reynolds. I'm a principal there, the firm is an IP boutique. We have two locations, one in the seaport district of Boston and one location in historic Concord. Our attorneys, many have advanced degrees, or multiple degrees. And we span the gamut of various engineering disciplines and applied sciences including life science. Next slide. My particular background, as you heard is in the All Things software, so software and computer related industries and helping to represent those inventions before patent offices. I'll have Steve introduce himself real quickly, just so we can hear voices and Michael. Next. Oh,

Speaker 3  7:16  
thanks, Mary Lou. This is Steve Levine. And as Holly introduced, I'm leading the human modeling initiative at my company called this whole systems in our life sciences industry, applying the basically computational modeling and simulation to digital health care.

Unknown Speaker  7:38  
Thanks. Michael, want to say a couple of words?

Speaker 4  7:44  
Sure. I just wanted to marry Lou. So this is Hi, I'm Michael Daly, I lead the Baylor licensing group and we are a part of a recently reorganized commercialization team at Baylor College of Medicine called BCM ventures. And part of what we're trying to do at Baylor is move beyond the traditional tech transfer model, and actually develop more of an open innovation ecosystem at Baylor College of Medicine. So it's one where we are focusing on if you will, kind of more broader collaborative relationships with industry partners, so glad to be here today. Great.

Speaker 2  8:18  
Thanks, guys. Okay, on to the basics of this technology area. So artificial intelligence, AI is the acronym is a subset area of computer software, where the digital processor is programmed to make decisions that mimic human decision making. So there may be pattern matching certain given certain constraints is what we would say in software. But those are circumstances, the specifics around the situation. And as far as input, and then what the decision as it's made for an output, time and time again, that may build a particular pattern that then can be taught or programmed into a computer. So as that programming or happens, we rephrase that as learning in the sense of machine learning for the computer, we use training data. So on the left part of the screen, we show that that body of information is important to have in order to help refine the programming and decision making being done by the computer. The live analytics portion is what is happening in real time running of an AI machine. And I'll call that the analytics engine. So in the live real world instances, we would feed in certain inputs, the analytics engine would chug along having learned from the training data of what particular patterns or decision making points would be and provides an output. So in, in that output is usable from the AI system as from a technology perspective to either run other machinery or provide a certain user output into a user interface. So, that's the basics on the technology, the hairs, we're going to eventually map that into how to protect those different pieces. So the range of our IP protection tools are shown in this slide. Starting with patents, which are to cover your engineering contributions, you know, advancements in the tech nology arts, and what we call inventions need to be useful, new and non obvious. under US law. Trademarks are protecting your business signature used in commerce. So a name a logo, your domain name, any identifier that connects you as a source of goods or services. Copyrights protects original works of authorship. So in the software world, the computer program, that source code is considered a literary work. As far as copyright protection is concerned, the analytics engine that I spoke about this, there may be some original code going on it with the writing of that, and that would qualify as copyrightable material. Trade secrets is a nother camp of confidential information. So it's the formulas, techniques, customer lists, other beta tests and data that has value giving the owner of that trade secret an edge over their competition. And that's largely in the eyes of the law. protected and administered through confidentiality agreements. So basic contract law. Now, let's map that to our AI systems and components, what pieces are protectable by which type of IP, intellectual property, so the overall AI system, as we saw from the input, and the training data, that analytics engine all put together to provide a particular output there that as a system can be protectable by each of these areas, patents, copyrights, trademarks, trade secrets, sometimes these types of machines have particular project name and coins, a trademark or trade name that people become used to using in the industry. So that's where the trademark protection would come in. If we only wanted just that analytics engine, separated out from a system as a chunk of software, that too could qualify for protection under patents, copyrights, trademarks, or trade secrets. So that's nice to understand that outside of the overall system, as long as it meets the criteria for patentability, the new non obvious and useful or for copyrights, literary works. That's an original writing. That's when these areas of IP would be able to protect that analytics engine. That next separate chunk though, that's always been interesting in this area of AI, especially when in the medical arena, is the training data. Not necessarily will that meet inventiveness, to meet the qualifications for patent protection? Copyrights, it's possible that there be copyrightable protection here as long as there's some originality, maybe the training data is formatted, put together in a certain way that qualifies for a literary work type of copyright protection. Trademarks? Not necessarily is there going to be a connection, an identifier of source and goods services here. So trademark kind of protection is very difficult if we were using that angle on this training data. Because the training data is typically a mix coming from various sources and When put together, it's hard to protect as a trade secret, particular confidential set of information. So, this map is to give you some guidance on where to turn if you're thinking of IP protection, or these different pieces of the puzzle. Okay. Let's move on to some examples then of how AI is being handled in the medical r&d circles. With regard to there are a variety of AI focus groups. I think the a couple of years ago, autumn, I had our friends from University of Alberta, dis describe their joint ventures center, the AMI is Alberta machine intelligence Institute where there's it's a research institutes specific for AI, and how they use an IP holding company and various licensing to put this consortium together and help grow this area of AI. Specific to medical research. There are a few database areas that have hit the radar and are of interest. There's the gamut Corp out of MIT holds a particular medical research database. There are data aggregators, tried net x and iqvia are two that are patient repositories in the Health Research Networks area. Even Amazon Amazon data marketplace has what's called June tau ju n t o, which helps companies find and buy relevant data. So it's a marketplace for data. And then in the pharmaceutical or drug discovery space. There's a European led effort. They go by an acronym i m i innovative medicines initiative. There's a they have a project called melody project that aims to bring pharmaceutical companies together to create machine learning platform for drug discovery. So again, has all those different components the training data, along with the machine learning the analytics engine. So there's a number of consortiums and conglomerates to keep in mind of how those partnering agreements are happening and growing, to bring researchers and data sources together.

Speaker 2  17:49  
So today, I would like to then segue into two very specific case studies of this AI in the medical industry and the living heart project, certainly by Diso systems, has caught everybody's eye and we'll have Dr. Levine talk about that. And then following that will be Dr. Michael Daly's speaking about the Baylor College of Medicine's delving into a handling of AI and precision in the Precision Medicine world.

Speaker 3  18:33  
Great, thanks, Mary Lou. So thanks, everyone, for giving me this opportunity to share with you this really interesting case study. This is a unique situation with a project that I launched. And I'll take you through that experience of basically creating what we call virtual twins, for precision medicine, and ultimately, the connection to AI. So virtual twin is a term that's be used in the manufacturing world, to represent a digital copy, essentially, of the real item. So if you're building a car, you first build it on the computer, and you can actually replicate its behavior on the computer, using the basic laws of nature. We understand how metal bends and how plastic breaks, then we can program that into the computer and understand how the automobile will behave. And therefore optimize its behavior, maintenance, etc. And the living heart project was really inspired by the idea that, well, the human body also obeys the laws of nature, we have to maintain it. We devise devices and we test them not on digital versions of humans but on actual humans. And if we had digital twins of humans, for example, a human heart, then we could actually do testing and develop techniques where we could use the digital twin instead of our real twin. And then medicine could be more precise, and then ultimately use that as our source of intelligence. So thanks, sorry, jumped ahead a little bit. So here's where the interesting intellectual property issues crop up. So unlike building a car, where you could go to a manufacturer and ask them, you know, they know they can tell you exactly how they build their car, can't go to a cardiologist and ask him, well, how do I build a heart, he may know how to fix it, but he doesn't know how to build it. That knowledge is distributed over a wide variety of different systems, we break down something as complex as the human heart. We look at it at different scales. So we look at the tissue level, the cellular level, the molecular level, we look at the drugs. And those manifests themselves in different kinds of physics, you have your electrical system, you have the structural heart, tissue, and your muscle. And then of course, the blood flow. And all of those are very complex systems. And so that knowledge is distributed over a wide variety of different people. Most people exist in different domains, there's, of course, a scientific researchers who, who try to understand exactly how things work. There's those out in engineering, who put those concepts into practice and build new devices and new techniques to help us. With new therapies. We have, of course, practitioners who learn how things work by observing their behavior. So they learn a lot, in their minds, try to communicate that as best as possible. And then of course, you have their regulatory folks. And they're all doing their own independent research. And maybe they publish their information, but they're all taking a piece of it. And so that knowledge is scattered around and fragmented amongst all of this community. And so living our project was really a question of, if we were able to unite all these groups together, and share the information on a single representation, because they all have the same foundation. We know they're all referring to the human heart. If we all work together, do we know enough to actually build and bring to life a human heart? The way we can do that in manufacturing, and then we could actually use it for precision medicine, as training, education, etc. So that was really the foundation for the living heart project, could we really build a physical model of the human heart? So the goal here will to bring together these different communities, which, as any of you tried to bring together, different disciplines? No, it's very difficult to get, for example, researchers to be able to speak the same language as clinicians or engineers. And of course, the regulator's have their own way of working in bringing them all together to get a common vernacular was one of the key goals, because I thought if we could get a common language and use the computational model, a virtual reality version of the living heart, actually as the common language between them, then maybe we could actually achieve the goal. So the purpose of this was to bring these together, and then ultimately, not just create a an interesting thought exercise, but actually use the fact that we have researchers now collect connected directly to clinicians, through the support of regulators, could we translate new innovations directly into clinic? Can we shorten that cycle, which we all know, is quite short now are quite long, and we need to shorten it.

Speaker 3  24:12  
Next, sorry. So in short, we set this out and you see the project timeline here. So as you can imagine, bringing together all of the experts, we started out I had the idea back in 2013. And brought together a few experts that I reached out to with this idea. The had confidence from bringing them together that they would support the effort, basically Foundation launched the project in May of 2014. And actually by May of 2015, we've been able to assemble together the knowledge that already existed, research that had been funded over the years and years It was actually out there. And we were able to put it all together and had a commercial product. And actually a year later, which, if you've worked, done anything in the computer industry to get a product out from concept to market in one year, particularly in a domain, like digital health is quite impressive. And it really demonstrates how much knowledge is actually out there. And I think it's key, what we learned was that there really is a lot of intellectual property that exists in the minds and hard drives, etc, of people around the world. What's really novel about, excuse me, our project was we found a way to collect up that intellectual property. Using this crowd sourced model where people could share with us their knowledge, maintain ownership of that intellectual property, they would give us a piece of it, and then we would return to them that piece in the context of an entire working system. So if I knew if I knew something about the electrical system of the heart, they shared that with me, I built that into our full model of the of the human heart and give it back to them in a working model of an entire human heart. And so the knowledge they had in their own domain expertise, they could then leverage, they could share the results of that intellectual property without actually sharing the adult intellectual property itself. And that turns out to be a really fantastic formula, particularly in the domain, very challenging domain, with respect to who owns the knowledge of the human body. And I think for the most part, what we what we found, and you can see in the left hand corner how much the membership has grown over the years. So it's a little small, so you probably can't see the numbers. But we have, this is a little older now. We have over 140 organizations around the world contributing, all working together on the same representation, some are helping to build it, some are helping to test it. And you can see here on the on the right hand corner, the range of applications, so being able to use it, to design new medical devices, to do personalized 3d printing, machine learning, etc. And then moving on to drug safety. So we now have this incredible piece of technology that is made up of intellectual property. And here's a representation of the the actual universities that are participating in the living heart project. So they all are working on the same model. But using using it in their own unique way. Some are bringing their their novel technology, some share code and ideas. Some are doing testing, as Mary Lou indicated, we have input information, we have output information, we have testing information, everybody's bringing a piece, I show up just an image from the one of the meetings. So every year we get together, up until this year, we've managed to get together physically, this year, we get together virtually like. But there's advantages. When you have such a large, broad community being able to get together virtually means more people can contribute. But because we've been able to create a common foundation, so people who had complementary parts of understanding the human heart, when they get together, they understand now how their piece fits into the hole. And that creates a bridge to the knowledge that everyone else has. And so we've created this very unique synergy between the knowledge of one group and the knowledge of another, particularly the researchers and, and the the practicing clinician, nurse, clinician. So this idea, this concept of bringing together the virtual twin is not just an interesting topic, it's as I mentioned, the regulators are key to the success of translating this technology into practice. The FDA has been part of the project since the beginning. Last year, they signed an extension of five year extension in the collaboration agreements, specifically, now that we actually have a digital twin of the human heart to challenge us to see if we could use it as a clinical trial to create a virtual patient population so that we could ultimately reduce the number of physical patients, certainly the number of animal studies. And so if we could create a new way to analyze treatments on the computer, time you take them to a clinical trial, you already have run the trial two or three times on these virtual populations. And so the risk is much lower, the knowledge base that you go into the trial is much higher. And so you can shrink the size of the cohorts, you can reduce the timeline, and certainly reduce the number of ambiguous endpoints, which often happens to you. The results are inconclusive, and therefore, you end up having to run yet another trial. And you see here, a quote from Jeff Sherman, who is the director of the CDRH. When in charge of all Devices and Radiological Health, at the kickoff meeting for this group, he laid his expectations out which is not small, he's expecting us to really change the world. And the idea that by bringing together not just the intellectual property, the knowledge based around the world, but bringing together the community so that people can actually understand how to use it. So we've been able to both protect their intellectual property and share it at the same time, I actually put the image of Iron Man Down in the corner is Dr. Shuren, when we were talking about this project, he said, I want to leave this image in your head, what I want is my reviewers, when someone comes to me with a new innovation, I want them to be like Iron Man get into that Iron Man suit, and have immediate access to every information, everything that really could be known about that treatment, I want them to have access to it. And by opening up the digital world, we actually potentially could have that. So that's our inspiration. And we have this community now of hundreds of people working together, we're actually going to try to look at a mitral valve. For those of you not familiar with the anatomy of the heart, the mitral valve is the most important valve in your body, it keeps the blood when your heart pumps at high pressure back into your body. It's the valve that keeps the blood from going backwards. And it has to operate under very high pressure and not leak over your lifetime, the minute it starts to leak, it's a steady progression for heart disease. And so there are 10s of millions of Americans that suffer from this. And it's a very underserved population so. So we feel like if we can actually help create a pathway to bring these new to market using these digital models, we can really make a difference. And we bring with us, we put this idea out to the living heart project community, and more than 50 members in the project have signed up to be part of the working group, that's going to help us build not just the best in class, mitral valve, but one that's completely parametric, so that we can, we can very rapidly make it represent the entire population, and all different disease states. And so by creating this community, we're actually changing the game with respect to understanding the the foundation of the human body. So where does this go? Well, here's an example of how we, we use the intellectual property and manage it. And this is really one of the more exciting things I think about using these digital twins, as I articulated, if you can provide people with a testing platform, essentially a cohort of virtual humans, you can imagine all the things you can do with it. So as Mary Lou talked about building, an artificial intelligence system really is a process of taking a training set, you take those inputs, and in this case, instead of collecting data from actual patients, we actually run it through the living heart. And we use that as our training set. So the living heart represents our patients, we collect the data, and we use machine learning to actually build an artificial intelligence system. We then take our test data, that test data then is run through the same cohort. So it's a different set of inputs run through the same cohorts, and then we use the artificial intelligence system to see if we get the signals and if we can get the same output between the artificial intelligence system and the real patient then we know the artificial intelligence system is valid And so we can go from a cohort of patients. So imagine this cohort of patients is validated against real patients, we use that because it's very precise and clean data, unlike what you often get in the clinic, which can be very complicated. So the data is very precise and well understood and controlled, we can create an artificial intelligence system that can replicate that entire cohort. And now we have something very powerful. And how we manage it, as shown here. And I apologize, it's a little bit small. But this is, again, key to the scalability of this project. And certainly something that has been the secret to our success is that each piece of it can be owned by a different organization, and it still works. So the training data, which is often clinical data, or, or knowledge that comes from the medical field, that may be owned by one organization, that another organization may actually build the artificial intelligence system. So they set up the configuration, and they may have their own code, or their own way of building the AI system. And they can own that, and they can patent it and copyright it. And we retain ownership of the analytical software, which is the living heart system. And that continues. So it can be used to build multiple AI systems on different training data. And so we create this platform that then scale and everyone in the project that has the ability to use it for their own application. And that creates a certain amount of scalability that's been really the secret. And hopefully a formula that we now are expanding to the brain, and lungs and other parts of the human anatomy. And so it's been a really interesting journey. We've learned a lot about getting people to work together. And and unfortunately, we all come together when it has to do with human health. So we've been fortunate to be able to give get people to create trust that will protect their IP. And so far, we're on our seventh year now and hundreds of members and so far, so good, and really pleased to share our experience and happy to answer any questions you might have at the end. So with that, I'll turn it over to Michael.

Speaker 3  37:41  
Sorry, just advancing. I had one more slide. One more slide. Yeah, this is just a summary slide. So if you want to learn more, we actually publish we, as much as we can on a website, three ds.com/heart. You can learn more and review the technical papers. So with that, I'll turn it over to Michael.

Speaker 4  38:08  
Fantastic, Steve, thank you so much. And thank you for the opportunity to be a part of this. So I'm going to come at this from from a different perspective. And the perspective that I that I'm bringing to the table is the perspective of of a technology transfer office director from a biomedical research institution where we manage intellectual property that covers if you will, kind of the biomedical research waterfront. So we deal with everything ranging from from therapeutics to vaccines, to medical devices, to educational materials. And now increasingly, we're starting to see this field of Artificial Intelligence impact our practice and impact what we do. And so I expected in that respect, I'm probably similar to many of you, where you've seen you've seen AI related technologies starting to filter into your jobs and filter into what you do on a day to day basis. And that's certainly been been the case at Baylor College of Medicine. And so my team is organized in a way that each of the project managers on my team manage a certain they manage intellectual property arising from certain academic departments at Baylor. And so there are some departments at Baylor that are becoming if you will, kind of AI development hotbeds certainly that I would say the Department of Neuroscience at my institution is one of those areas. And so I thought, though, that just in terms of giving an introduction for tech transfer personnel to the to to the AI space, that it might be worth just stepping back and taking a look at what's happening in AI. I mean, this is this truly this this is an area that, as you probably know is it's going to profoundly impact medicine and all aspects of our lives. It is just in just to give you just an An idea of the scope of of the change that we're witnessing, at the USPTO. Now, AI related inventions comprise about 26% of all annual patent filings. It's exploding. The Patent Office, according to director Yan Q, they've recently doubled the number of examiner's that they have to manage cases in the AI space. And so who's doing it who's doing this innovation? Where's it happening? And it's probably not a surprise to many of you that, you know, the players in big tech are really key in this space. So that you know, so IBM, in fact, IBM, I think leads by a significant number of patent applications. But, you know, Microsoft, Google, Amazon, they're, they're all involved. In interestingly, and this has actually been been mirrored by some of our experience at Baylor College of Medicine, in terms of where AI related technologies are being developed, Israel has actually become a hotbed for startup activity and a hotbed for activity in the AI space. And in fact, in our institution, we've actually been, we've been approached by Israeli VC groups and Israeli and Israeli AI related startup companies that are interested in collaborating with and working with us. And again, just to give you just a snapshot of the countries where innovation is happening in the AI explosion, the US in terms of just published this is so this just covers obviously one dimension of the intellectual property matrix in this area, and then that in that dimension is published patent applications. So that, you know, the United States is in a leadership position, but China has been coming on very strong. And then and then we see again, other other nations in the world that you know, that are contributing to this, you know, contributing to this area, but again, an area where we're gonna see explosive growth, but really, if you will, sort of two main strong competitors, the US and China. So then, diving down into the the environment of the academic medical center where I reside, how are we relevant, and I think that this is going to be the next frontier in technology, commercial commercialization and licensing, and if it hasn't significantly impacted your practice it's going to, and so one of the areas where we are certainly relevant in this space, deals with deals with data. So academic medical centers, we have patient datasets. And as both Mary Lou and Steve have spoken about, we have those we have this data sets. And I think what makes us unique is that in academic medicine, we often have datasets that are specific to a unique group of patients. For instance, at Baylor College of Medicine, we have strength in we have strength in cancer diagnostics, we have strength, also in breast cancer treatment. And so we actually we've got datasets in those spaces. Additionally, we've long been a powerhouse in terms of dissecting genetic mechanisms associated with rare genetic disorders, particularly rare genetic disorders in the in the neurological disease space. And so we have some very unique assets in terms of patient datasets. So that you know, so that I think makes us an attractive partner for pay for companies that want to partner with us. So just to give you an example, we were recently approached by a company that was interested in partnering with us around the development, they had an AI tool that they had developed for interpreting mammography images. And but in order to improve that tool, or to make it better, they need access to additional teaching datasets. And so they approached us because they wanted access to our datasets. But we then entered into collaborative discussions with them, because our physicians and clinic clinicians felt that they could actually work with this company to make improvements to their tool. So they would they want it if you will, kind of more of an interdependent collaborative relationship, where we're making contributions to improve the tool. So it's not just about access to the data. It's also about an ongoing relationship. Next slide. So as I mentioned, data is the new oil. And we are seeing this just in our practice, certainly with our group. So many of the agreements that we are negotiating and executing now have a data have a data component to them. About what we also see as an academic medical center, as an academic medical research institution, in the middle of a large cluster of other academic medical research institutions where we have partnered relationships is that when we get approached by a commercial partner, and they want to and they want to actually develop a collaboration with us, and they want access to our datasets, it's not always a trivial proposition to make that happen. And in our particular case, AIX, because we have affiliated relationships with other institutions across the Texas Medical Center. Our data sets reside in different places on different servers collected under different electronic medical record systems. And in order to facilitate that relationship or that partnership, you first must aggregate them, you have to understand where the data are. You also you also have to understand if a third party partner wants to integrate a relationship with us, and they want access to those data do you have under the patient consent form that that allowed you to facilitate the collection of that data set? Do you actually have the clear, unfettered right to share those that data set with, you know, with a third party partner, that's not always trivial. Five years ago, we weren't dealing with these issues in our practice. Now we deal with them on a daily basis. And it's very interesting, the degree to which our practice has changed, and I expect yours has as well. So if you are approached, and if you're dealing with these issues, it's you've got to, you have to dive into and understand the datasets, you have to understand where they are who controls them, you have to understand the process by which the data were consented. And you know, and then again, you have to aggregate all this and bring it together in order to in order to facilitate a commercial relationship. And that's those things are those those aspects are not trivial. So what types of when we talk about datasets, there are a couple of data sets, that is you dive into this world of epic, you'll become familiar with this terminology. And for the most part, we've seen an explosion of agreements, dealing with what we what are called limited datasets. So limited data sets, they contain some pH i, and some which is protected health information. So they contain some protected health information that is necessary to facilitate the analysis and the interpretation of those data by the third party. But there are some but then there are some identifiers that can actually you could link the data to specific patients that are left out of a limited data set. And so we deal we deal extensively with agreements that involve limited datasets. And then and then we also to some degree deal with de identified data, but de identified data, it may it is data that is truly de identified and doesn't contain any of the elements that would otherwise comprise protected health information. But just over the past, I would say over the past year or so, just in terms of data use agreements that involve limited datasets, we've probably executed a couple 100 of them. And that and that, and that's a huge change for us. And five years ago, we were doing a handful of these. So this the the volume of work that we're doing in this space has exploded, and it's going to continue to grow.

Speaker 4  48:00  
So back to the example then, because you know, so so this is the question that I think I think academic medical centers need to think about when they're approached by third parties who want access to patient or to clinical datasets. And so let's just assume we've that we've covered the basis that yes, the data is it's in a form that the third party can use, it's in a form that they can access, we've solved the for the consent issues, we know we can share the data set, we're in good shape there, then you need to think about okay, but how does this relationship advance the mission in the interests of the academic institution? What are you getting out of it? So you're in because we have been approached by you know, by third parties that will essentially they'll pay us for access to data? And so you can so you can certainly you can get revenue from it. But is that really does that it do does that revenue stream does that sufficiently advanced the interest in the missions of the academic institutions such that it makes it worthwhile for you to do it. We've also been really in I think, kind of our sweet spot is really coalescing around a model where we actually engage in a co development project with a third party partner. So yes, we're providing them with access to the data set. And this aligns really quite nicely with, with the living heart example that Steve just presented where, yes, it's access to data to a data set, but then we're there. It's also contribution. It also the relationship revolves around ongoing contributions of expertise to improve and further develop, let's say the algorithm that analyzes the data set to to produce improved patient outcomes. So that's, that's I think, where that's I think, what our focal point has revolved around when we're approached by third party partners or when we when we go out and we approach them. We're really looking for a relationship that at the end of the end of the end of that relationship, the outputs from that relationship, they actually align with our mission. And they enhance our research, research patient care or educational mission missions. And if that mission alignment piece is there, then we move forward with a relationship. And if it is not, then typically we don't we decline the relationship. And we search for partners that are mission aligned with what we're trying to do.

Speaker 4  50:24  
And so, I want to just comment for a moment on it. So both Steve and Mary Lou have commented on this topic of these aggregated datasets. And so I think, I think the real power for for precision medicine and for artificial intelligence to impact patient outcomes, which is as a biomedical research institution, that's what we're all about. We want to we want to, you know, we want to leverage technology, leverage, leverage science to improve health. And so to do that, I think from the from this universe of data, the power is going to come from these large aggregated data sets. And so we've actually been involved in, in data hubs. And it's, it's really interesting to see as well, that disease specific foundations are diving into this area. And they realize that in order to produce improved outcomes for our sometimes very specialized groups of patients, we've got to reach out to the to the nodes where those datasets and set it sets exist and pull them in. And if you will develop a hub of specialized data that can then be used to love you, too, can be leveraged to produce outcomes for a particular patient group. And it just as an example of this, and I think this is public information that it's out there that the Leukemia Lymphoma Society is actually doing this. And they they've been along a longtime sponsor of research at Baylor College of Medicine, we're working with them on this data hub, you know, the state of hub concept. And this is an area that again, I think we're going to see much, we're going to see much more work and we're going to see a lot of activity, because this is a route by which we can improve patient outcomes.

Speaker 4  52:09  
So just to provide some examples, let me back up one. Holly, can you back me up to slides please?

Speaker 4  52:28  
Perfect, thank you. So for most of us that are there in technology transfer roles in academic institutions in the biomedical space, the AI, if you will, the AI in the AI landscape represents something new for us. You know, most of our activity has traditionally revolved around patenting therapeutics, patenting diagnostics, and and now what we're seeing is that we're diving into this new world, where in we have researchers that are developing other developing algorithms, and they're developing tools, in in many cases with what we're seeing in the AI world is that the researchers lead developers in this space, they tend to make their the outputs of their work available via Open Source Formatting. And so this, you know, from from their standpoint, they want to get the tool out there, get it out there, let other researchers build upon it and enhance it. And they view that obviously, as a tool to enhance their own reputations, but they also use it as a recruitment tool to identify talent that commit that they can bring into their programs. And then, as I mentioned previously, you know, a lot of the big tech companies are pursuing patent portfolios in this space and primarily pursuing them for for defensive reasons. And then what you see is that it becomes if you will, kind of almost kind of a land grab, where were the different where it were, if you will, you have kind of this concept of Mutually Assured Destruction if we don't patent everything, the other guys will. So we have to so that we can defend our relative positions. As academic institutions, we typically we aren't competing at that level, we don't have the resources. So most of the patent applications that we filed in the AI space have really revolved around if you will, kind of niche algorithms for a specific purpose.

Speaker 4  54:18  
So just to give you a couple of examples, of what I mean by that, we can go Can we go back a couple slides, Holly, two example one. Fantastic, thank you so much. So a team at our institution. And this is a great example of kind of a real real world example that we're working on at our institution. We have a research team that's actually engaged in developing an AI neural network that is trying to actually mimic the way that the brain actually responds to and processes information. And so what they're actually they're what they're doing is they're actually using animal models that have a reporter gene. such that they can actually follow the pattern of neuronal firing in the actual mammalian brain as the brain responds to either a visual stimulus or an auditory stimulus. So they can then expose the model to to a stimulus, and then actually physically observe and collect data from the pattern of neuronal firing that actually occurs in the brain. They then take that dataset and try to design algorithms around it that more closely mimic how the brain actually processes information. So what from an IP standpoint, what might come out of that? So several things, one would be obviously potentially patentable subject matter if you know novel algorithms are can arise from the project that if you will process information in a different way. And if also, if we've got sufficient we've got sufficient written description enablement to pursue a patent application, certainly, we will see I think patentable subject matter arise from this, we will also see I think, through the development of novel software, some potentially copyright, copyrightable output. And then we might get into this, we might get into this realm of trade secrets. Well, what this is, again, for an academic institution where we are typically devoted to the dissemination of information, the trade secret space is a bit different for us. This is an important difference from from many of our industry counterparts, that deal with trade secrets on a on a frequent basis, we don't so much, right, we're typically when we talk about intellectual property that we manage and license, it's typically patented intellectual property, or it may be non patented research tools. It may be copyrighted, copyrightable software. We've also in recent years increasingly, jumped into licensing very specialized know how that isn't, you know, isn't patent protected, per se, it's not really it's not really a trade secret. But yet, it's very specialized knowledge. But we built a company around this in the microbiome space, where the foundation of that company was really built around very specialized sets of know how that one needed to leverage information associated with the microbiome. But going back to this example, I think what we're going to see arise from this is going to be it will be a combination of different types of intellectual property, certainly patentable, copyrightable, potentially trade secret area, potentially trade secrets, as well. And so you have to be sure, and this is, I think, also a time to actually go back and examine your institutions intellectual property policy, does your institutions policy adequately capture all the rights that might arise from a project like this, you know, many of us are dealing with with intellectual property policies that were drafted maybe 10 or 20 years ago? Are those policies updated? Are they current? And are you adequately capturing everything that your research researchers might, your researchers might be able to may wish to leverage. And then, so the team that's actually working on this project they want to do, they want to form a startup company. And so another challenge that academic institution institutions face in the AI space is that it is often difficult to recruit talent to come in and work in the academic environment. And again, the big players in tech, if you if you are a programmer, and you have you have, obviously, if you've got specialty, you've got specialized skill sets. You're a software engineer in this AI space. It's difficult just because from salary constraint standpoint, for a for many academic institutions to actually effectively compete. So the the academic researchers at Baylor that are working on this project to develop if you will, kind of a neural network approach that more closely mimics how the brain works. They want to form a startup and hire engineers into the company, and then collaborate with them in order to to advance the project, because they know a barrier for for you for executing the project is actually is actually being able to pay the software engineers actually recruited in the academic environment. So let's form a company, recruit them there, and then we collaborate with them and CO develop the software. So from our standpoint, if we end up licensing the rights that are developed at Baylor into this startup, what should we be thinking about? And in this, this project is so early, it's so investigation that you know that one, there may be a variety of downstream applications that might arise from this. But it's very difficult to envision that you know it from where we stand today. And so, I tend to think that if this project is successful, and if the project actually yields the development of, you know, proprietary algorithms and software that might more closely mimic how the brain processes information. From my standpoint, I'm probably thinking about taking equity in that company. Because regardless of the outcome that the that the work produces, it's likely that if they are successful, they're probably going to get acquired by a bigger player in the field. And if that happens, and if I've got an equity stake in that company, then my institution is going to benefit from that you will benefit from that downstream success. We also think about that field the thinking as well about potential fields of use. This is a small city of a very small startup company. That is yet it's being developed in the biomedical space, yet the software that they develop, or the approaches that they develop, might have applications outside biomedicine. And those applications haven't been completely thought through. So what you're one of the areas that we're thinking about, as we contemplate a licensing relationship revolves around will revolve around fields of use and what what rights should we license initially, what rights should we potentially hold on to, perhaps for later transactions, and then it's, it just add a little more complexity. The software now is being it's being developed in the academic setting. And it's an asset as asset of the academic institution. But then we're also sub contracting with this company vehicle to continue to develop, obviously derivative works and improvements on what has been developed within the academic within the walls, the academic institution. So we've got to make sure that we marry both of those and bring them together.

Speaker 4  1:01:32  
And just to give you another example, just for some additional flavor of the types of technologies that we're dealing with, we have another research team at Baylor College of Medicine. And this, this fits right into the younger precision medicine sphere, where they're actually developing a proteomics and microscale proteomics assay system for dissecting if you will, kind of the proteomic data set associated with a patient's tumor. And the goal behind this project is to essentially to take that proteomic information and to do to develop new software tools and systems for analyzing that information, to guide the patient to the appropriate treatment the first time, so that the tool to effectively take the if you will the guesswork out of cancer treatment. And so so this so when you look at this, when you look at a technology like this, what do you have? So one area that we're so one area that where we do where we know we have technology revolves around patentable subject matter. And so the patentable subject matter revolves around a couple of key areas one has been, the research team has been focused on micro scaling, if you will, kind of the the proteomic analysis pathway. And so there's proprietary technology for potentially proprietary process related technology associated with if you will, the scaling of the the scaling of of this, of this technology, because what the what, what the team wants to do is essentially be able to work with a very small microscale biopsy sample from a patient, you know, as small as a couple cells, and run a proteomic analysis on that sample, interpret and process that information, and then guide the patient to the appropriate treatment decision. So that's also that will also potentially yield algorithms associated with new software that you know, that will arise from this effort, potentially, we will, we'll be diving into into copyrightable matter through the development of new code, trade secrets may come into this as well. And so one of the questions that we're addressing is, there are there so there are already industry players in this space. This is a hotbed. It's a hot area, we have it we have expertise here, but others do as well. And so one of the questions that that we're addressing from the academic research institutions point of view is, does it make sense to try to put our own stake in the ground and build a startup? Do we have enough critical mass in this area to build a viable startup around this approach? Or do we approach one of the bigger players in the field that's already established in the in this area? And if you want and develop a collaborative relationship with them, or potentially license our IP rights to this bigger player, so this is just an area where we're having ongoing discussions at our institute? We haven't resolved it yet. But I, you know, I do think we will. And this is, you know, this is an area where we're just seeing, again, an explosion of new technology of new work, and it ties in. And I think the key point, the reason that I put this example, into this talk is it's not just about AI. Yes, there's an AI component to this to this technology platform. But it's it's AI, but it's also if you will, it's also it's also an analytical part halfway involving the other physical manipulation of samples, so it's, if you will, it's kind of a complex mixture. And I think you know, for many of us in the tech transfer world, this is what we're going to be diving into is it will be this kind of complex mixtures of technologies that involve an AI component, but involve other components as well.

Speaker 4  1:05:27  
And with that, Mary Lou, I think I'll hand it back to you just to go ahead and summarize our talk for this morning.

Speaker 2  1:05:33  
Thanks, Michael. And thanks, Steve, I tried to distill everything down into just easy, put in your back pocket, take home tips out of all of this. So I think these were great illustrations of how to use a combination of IP, don't just get tunnel vision down to the same old invention, disclosure provisional application, and then don't go anywhere. There really is this area of AI and precision medicine, ripe for that whole combination across copyrights, trade secrets, trademarks, and patents. This, the second take home tip is to really think about how your institution or company is set up to handle that training data area. There may be some policy issues also kind of giving you guidance there. But don't excuse that as far as a non active area there certainly are uses out there that could be bought and sold. So for a revenue generation purpose, as well as for promoting further project development. And the the last point, I think we also made in our first time through this presentation a couple of years ago that this truly is a growing area, as Michael put up the stats, and there's an explosion here. And so you should just keep an eye and ear out for the those additional players that keep coming onto the scene. Sometimes it's very technology specific area in the medical field, it may be disease specific area of research that these AI and precision medicine projects seem to come about. So keep your eye open for those special consortium groups and special interest groups. That's it that I had as far as our formal slide presentation. So I think we can oops, sorry, open this up for question and answers. Let me just check to see if we received any.

Unknown Speaker  1:08:00  
Holly, do we have any questions?

Speaker 1  1:08:03  
Yeah, we do. Let's see. And just letting everybody know, we've got we've got about 1520 minutes for questions. So go ahead and put them in the q&a. And we'll get to those. But let me start with what I have here. In the last couple of years. Have you seen trends or changes towards standardizing or sharing of patient data for research purposes?

Speaker 2  1:08:28  
I'll let Michael or Steve handle that is in the medical industry more aimed at patient data and the lack of standardization if you've seen some strides there.

Speaker 4  1:08:39  
So sure. Thank you. Thank you, Marie. I just I yeah, I'd be happy to briefly comment on this. So at Baylor College of Medicine, we are in the middle of in terms of patient care volume, the world's largest medical center. So we are we are we are centered right with with so Baylor MD Anderson, University of Texas Health Science Center, Rice University, Memorial Hermann system met him at Houston Methodist. So there's this huge cluster of academic of academic medical institutions that that were all collecting huge volumes of specialized patient data. And we are all part of an infrastructure organization called the Texas Medical Center, and the Texas Medical Center, they are indeed trying to drive efforts to if you will kind of build a TMC specific data, your patient data warehouse, and this is the other so there are ongoing efforts in this area. Because I mean, if you can imagine if you could, we are the largest and just in terms of patient care volume, we were just the largest medical center in the world. If you can leverage that entire data set rather than the datasets owned by the individual institutions. You can certainly increase the power of what you can offer and so so yes, we are seeing much activity in this area.

Speaker 3  1:10:05  
And I can add? Well, certainly, the demand, as Michael pointed out, is phenomenal. And the need I think, is being heavily discussed in many communities. In the US, I also sit on a board for a European community who are addressing this exact problem. And I think in the European community, they're even more concerned with privacy. Interestingly, because of nationalized health care, the government has all your data. And of course, people are least comfortable trusting the government. So they're even more concerned in Europe than here. But but there is no doubt that the demand is driving the boundaries to be pushed more and more so research data is becoming available. And I think there are is definitely, I think, as, as Michael pointed out, there was very clear rules about what what needs to go in to make data de identified. And more and more there are machines being put together that immediately, for example, we we have a machine that will de identify data, automatically, it will even go in and look at digital images, and any will reprocess the image will identify if there's a name on that image and eliminate it, those character recognition. And so you can actually get reliable the identification machines becoming available. And I think once we, once we crossed that boundary where people are comfortable, if you process it with a machine, you can use it. And then that will open up, I think a lot of research availability.

Speaker 4  1:11:57  
Actually, Steven, it's interesting that you brought up that you brought up Europe and that you brought up the Rio the privacy violations that that exists there, because they in fact, are more stringent than those that we have here in the United States. And we have observed this, in some partnerships that we've been attempting to negotiate with partners who are also dealing with European data sets such that they want us as a US institution to actually represent and warrant that our privacy protection, you know, that we are effectively GDPR compliant. We are as a US based institution, we are not. And so we can't give them so that has actually created some complication, you know, for some of the partnering relationships that we've been engaged in, so it's interesting that you brought that up. Yeah. Yeah. Okay, okay.

Speaker 1  1:12:54  
All right. We have a couple more questions. One of them actually are the question asker wants to ask it live. So I'm gonna go ahead and unmute you, DJ. It's related to HIPAA compliance, and the nature of data. So, DJ, I'm gonna unmute you, you'll see a pop up. And then you should be able to speak.

Speaker 5  1:13:17  
Great. Can you hear me? Yes. Great. First of all, what a phenomenal webinar and what a timely topic. So So Thanks to all three of you and autumn. I'm BJ nog, I'm the Chief partnership officer Advantech solutions, and we manage probably one of the largest datasets for the Center for Medicare, Medicaid services about 60 petabytes of data. And obviously, compliance and governance is is a big topic that we deal with every day in our 500 person organization. My question, Michael, you were talking about some of the research datasets and the enterprise data warehouses. And we see a big challenge and you know, getting that data from universities and research. And and I guess my real question is going to be you mentioned about the research data, which is, it's not simple, it's never simple. But how do you deal with ensuring that the informed consent, the HIPAA form that patients have filled out are actually complying and you can actually license that data? And the second part of that is most of your data, the ICD nine ICD 10 codes and all of that information resides in your EMR? How are you handling that side of data? Would love to get get especially your response, but would like to also understand if Steve, you are dealing with both kinds of data or just kind of research type of data? Thank you.

Speaker 4  1:14:47  
Sure, DJ, you've actually posed, I think, a couple of great questions. And so what I said with regard to the first question, we actually work with other groups at bay Learn that have expertise outside of our own. I mean, we have traditionally we focus on licensing, intellectual property management. But when we start diving into, you know, into to being sure that you know that let's say that a patient consent form a lot, you know, might allow us to share data set with a third party partner. There's we also have a compliance team at Baylor, and we will reach out we will engage with them. In many in many cases, if we need to, we will also reach out to a to a specific IRB. But so there so it the the answer DJ is it varies from dataset dataset. But that if you will, that that expertise around around HIPAA compliance and consent, that's that's not an EMS not expertise that resides in depth within my team, per se, that's not, if you will, our key mandate. So we reach out to other groups at the college, obviously, you know, the veteran and do that are involved in the management of compliance and conflict of interest to go to interface with them to be sure that, you know, before we contemplate entering into a partnership, we really do have the the necessary, a that we can we've appropriately protected the information and then you know that we are in a position to be able to share it with a third party partner.

Speaker 3  1:16:19  
Thank you. Yeah. Well, I would say the least from from my experience, at least, the focus that we've had, fortunately, has been on smaller datasets, at least a bit. Well, we really have two initiatives, one that I talked about today, which tends to focus on smaller datasets, in which case, we can be a little more hands on in terms of ensuring that we have compliance with our larger datasets, where we manage clinical trial data, etc. You know, as I said, I think we're, we're taking responsibility for ensuring that that that compliance is handled, and then we're licensing that out. And so we are we are taking on that responsibility. And so that, and that's, of course, massive value to the people we license or make that data available to. And so it's an interesting question in terms of the value, how much is how much of the value is in the data and how much is enough the management of the data. And I think over time, you're going to find out that it's the management of the data, that's really the critical piece. Because that's the only thing you can rely on from building a business.

Unknown Speaker  1:17:48  
Great points. Thank you, Steve.

Speaker 1  1:17:49  
Yeah. Thank you both very much. Let's see, our next question is have you done any trade secret licenses? How do you handle breach terms? And once the secret is communicated? What recourse Do you really have to hold the licensee accountable?

Speaker 4  1:18:12  
So we, I mean, I can take somewhat of a stab at that. So so the trade secret area is in for us as a new area. We so we really don't have much depth of experience with actual Yeah, with actually executing agreements that that have a trade secret component. We have done quite a bit in the space of a really a licensing, they're really specialized know how. And in that case, what we've done is we've really taken a lot of care to actually build out a very detailed description of what we mean by the know how asset that you know, that we're licensing, and that we're conferring rights to to the third party company to develop. And I mean, like in one case, where we formed a company in the microbiome space, there. So there were assets, you know, associated with we're with sequencing techniques with bio vo with and then with bioinformatics analytic techniques. So there were some copyright components there, but also a lot of spare literally, the schedule of the appendix of the agreement, the schedule of specialized know how that we were licensing into the company was several pages long. So we actually really took a lot of time to build out a very detailed description of exactly what the parameters were around the the asset that we were licensing, and I would expect that we would, we would handle a trade secret in a similar fashion. But we really don't have much depth of experience in that area. As of as of this moment.

Speaker 2  1:19:44  
How Campion the times that I've used trade secret clauses in a licensing arrangement, are often combined with the other areas of IP patents in particular, but specific to breach then one As a trade secret is out there, there's no calling it back. And so we wrote in clauses that today's for breach of specific to this this secrecy the confidentiality of the trade secret and surrounding information, a particular dollar amount. And as far as needs to be repaid, it's often very difficult to try to project what the value in the market would have been if the secret had been kept for longer time. But sometimes I get a counter offer, then to those clauses, because trying to put a particular dollar value is gets very difficult if you're trying to cover a number of years into the future. So mediation and arbitration is often also thrown in there, as far as the the breach happens, then we need to arbitrate over the dollar value that need for repayment, and how to then move forward. Because that certainly would terminate the overall License Agreement arrangement. And that's been my experience on how to handle that aspect.

Speaker 3  1:21:22  
Yeah, mine has been fairly similar. We have done cases, you know, we have much of what we're doing, as I described, we maintain this trade secrets, we try to, we try not to patent the knowledge or the application. And so we keep the knowledge, as much a trade secret. But we do share it, and we license it for people who need it. And we handle it more or less than a Mary Lou, you describe that we, you know, there are financial clauses, which are limited, and often heavily negotiated. And fortunately, we've not run into a problem yet. And there have been cases where organizations, particularly when you're working with academic, for example, organizations that might be mentoring a startup or cultivating the liability. You know, essentially, the organization doesn't want to take on the liability for the individual, right, which I completely understand that. So that becomes a stumbling block for us. And then we have to decide on a case by case basis. Fortunately, we haven't run into that problem yet.

Speaker 2  1:22:40  
But you brought up a good point as far as academics, when that's a party to the License Agreement, or those who are around the table trying to negotiate. Because many universities, that's one of their priorities is to publish, rather than keep a trade secret. So we do at length and some of our license agreements have this sense of this, the sequence of steps that need to happen in order to be approved to publish under this license agreement. And that also gives time, an amount of time as well as the process and puts the other party on notice then, if that's about to so called publish a portion that would impact the trade secret that goes into the thinking. So that's a kind of a side area to also help with the overall stepping stones and plan within the overall license agreement. Make sure one part fits with the other. In other words, and Michael, maybe from a university standpoint, you have run into that publish area and how to handle that in the license agreements.

Speaker 4  1:23:59  
Yes, we have and in license agreements and collaborative research agreements as well. Where we do we do. We will accept, obviously there, there's typically a publication review period. So the Clelia, the party that we're collaborating with, they'll want maybe a 30 or 60 day review period before you before we submit a publication. And so we work with the party on that, obviously, as an academic institution, we want to publish and disseminate information and so that's so we don't want what we what we don't want what we won't accept as an absolute bargains publication. But we will work with a partner on a review period and then we do understand the sensitivity partner around confidential information that is specific to them. And to just as a quick corollary, corollary, I would also bring up one of the topics that we've dealt with in this area in collaborative relationships has been when a third party company wants to share A trade secret with us as an academic institution. And now we have liability, potentially liability for protecting a third party trade secret. We're academics we talk, we talk, we publish, we present, that's what we're all about. And so it can be a real challenge for us to be confronted with a relationship where we're now going to be exposed to a proprietary trade secret held by a for profit company. Because as Mary Lou and Steven both both pointed out, well, what happens if a breach occurs, you now have a liability issue. And as a nonprofit, academic research institution, we're very wary of accepting of entering into into, if you will, kind of provisions in an agreement that expose the institution to that liability, you know, and might damage our endowment or our mission.

Speaker 1  1:25:58  
Great, thank you all those great insights. We've got time for just maybe a couple more questions here. So let's see COVID-19 has generated an unprecedented amount of research that's freely available, what are your thoughts on the IP implications of this data being available, and might it set precedent for the future?

Speaker 4  1:26:23  
So I'm happy to happy to jump in and comment on it when the COVID 19 pandemic hit, and instead, we all transition to this new method of working. We, as a researcher, as an academic research institution, decided that we were going to immediately prioritize anything related to, to to our work, that could have a potential impact in positive impact with COVID-19. So that meant that any operative project, any data use, or data sharing related to COVID 19 data sets, we would prioritize, we would prioritize those agreements, those relationships, and those went to the top of the pile, and we move them forward, and we executed them first. And I can tell you that we've executed a flood of agreements related to so we've been as an as an academic research institution, we're very active in developing COVID new and developing COVID-19 diagnostics for the community. We also have vaccine development programs we've had, we've had multiple partnerships around the development of vaccines. And interestingly, we have also done a lot in terms of actually developing wastewater testing to detect the COVID-19 viral particle. So that COVID is inserted. So with all those, if you will, with all those priorities, so essentially, we've prioritized COVID-19 related collaborative projects, move them to the top of the pile, and we've gotten them done.

Speaker 2  1:27:56  
That's great. That's terrific. I think, from the client base that I work with, certainly, we're very sensitive to any of their requests, whether it's a research angle versus a business angle, and those who have and I'll say NIH, grants are funds at stake. So we tend to take those invention disclosures that were specific to COVID-19 and place them into whatever time track that the client needed. Certainly, those were pushed along as, as the clients timing dictated as far as the data coming out of the COVID-19. Era. And all of the the collection there I we fall back on the particular HIPAA compliance other policies that that say was this information given with permission to share or to help keep confidential within that organization? So we're very cognizant of that. And if some of the information came from overseas from Europe, yes, we needed to understand their GDPR compliance and regulations. So that's always interesting when it's across international realms.

Speaker 1  1:29:38  
Thank you. So it looks like we are at the end of our time together. Does anyone do either of you have any parting comments before I close this out?

Speaker 2  1:29:48  
This has been wonderful. I always learn a lot from Michael and Steve. So thanks to the panel for pulling together once again. Thanks, Holly. Thanks, all Yes,

Speaker 4  1:30:00  
I hope this was useful to your My autumn colleagues. Like, again, just just an excellent topic and, and one that I would say as a practitioner, we're just starting to ride the crest of a wave. And I think we still don't really understand the degree to which AI is going to impact our practice impact what we do. And I expect this is a topic that we will be revisiting for years to come.

Speaker 1  1:30:27  
Great, thank you. On behalf of autumn I want to thank Mary Lou, Michael and Steve for this great discussion. I hope everybody listening got a lot of really good information out of it. And thank you for everybody who attended today. Just a brief reminder, recording of this webinar will be available for viewing within just a few days of this event. Access to the recording is included in your registration fee. Um, you can visit the autumn website to view the recording or purchase a past webinar you might have missed. If you have any questions about CLE availability for this webinar, please contact me directly. It's brand new that we're offering CLE for webinars. So if you have any questions, I'll work it through with you and we'll make it work. When you close out of this evaluation will pop up please complete that that really helps us plan for future webinars. And with that I will conclude our program for the day. Thanks for joining us and have a really great afternoon.

Unknown Speaker  1:31:23  
Thanks all. Thank you.

Unknown Speaker  1:31:24  
Thanks, everyone.

Transcribed by https://otter.ai