Scaling the production of biotherapeutics is challenging and requires more than just multiplying the ingredients in a recipe. That’s true for gene therapies as well where simply changing the nucleic acid of the payload can have dramatic effects on the product. Form Bio is seeking to address the challenges of making genetic medicines by leveraging AI to accelerate the development of cell and gene therapies. We spoke to Claire Aldridge, chief strategy officer for Form Bio, about the challenges of developing gene therapies, how Form’s AI platform can address that, and how it works grows out of tools that were developed to bring the woolly mammoth back from extinction.
Daniel Levine:
Claire, thanks for joining us.
Claire Aldrige:
Thanks so much for having me. I’m excited to be here.
Daniel Levine:
We’re going to talk about AI form Bio and how the company is leveraging its platform technology to improve the manufacturability of cell and gene therapies. Before we do that though, perhaps we should start with the wooly mammoth. Can you explain how Form Bio came about?
Claire Aldrige:
It’s funny, everyone always wants to start with the wooly mammoth. I don’t understand. So yes, the Wooly Mammoth Project, colossal Biosciences, that was a company I first got introduced to the C E O Ben Lamb a couple of years ago, and of course, who hasn’t heard of George Church if they’re in the genomic space. And Ben is a Texas-based entrepreneur, and I have been working in North Texas around biotech for the last 25 or so years. And he came to me and said they had this idea to extinct the wooly mammoth. They wanted to build that company here in North Texas. And that in addition to the science they were bringing in from Harvard to start the work of doing multiplex CRISPR editing on the Asian elephant genome to basically recreate the wooly mammoth genome from the over 50 genomes that we have because we’ve been able to get D N A from those frozen samples that we find every so often in the permafrost, that they would also be building a software platform to facilitate that.
The amount of computational work that needs to be done in order to understand what variants are important, what are the edits that need to be made to really change the fundamental nature of an animal taking for example, an Asian elephant and turning it into a cold adapted furry wooly mammoth that looks and feels like wooly mammoth. That is an enormous computational task. So that was something that really interested me as someone who had been in this space for a long time, I knew that having really sophisticated computational tools that were easy for biologists to use could really accelerate discovery. So they built that tool to again facilitate this editing and genomic analysis to work on this biodiversity and de-extinction project. And then there was a recognition that this platform that they built and these tools that they built were broadly applicable across molecular biology. And in order for that to make sense as an independent business unit, it needed to come out from under colossal stand up as its own company form bio and start to look for other problems that the tools that had been built solve. And really just through some great relationships that we had and some understanding was a recognition that that manufacturing process for these very complex therapies like cell and gene therapies was a great place to kind of direct our attention where we could capture a lot of value for these types of companies.
Daniel Levine:
I’m wondering if you could expand on that a bit. How did you come to recognize that those tools were applicable in this way?
Claire Aldrige:
At the time, we were working very closely with a number of gene therapy companies. We’re here at North Texas, UT Southwestern Medical Center has a gene therapy center of excellence. And so really just through a lot of the work that we’d been doing in North Texas around cell and gene therapy and seeing the problems, have also been a biotech investor for a number of years, seeing the problems of trying to translate these kind of therapies from the academic lab into the commercial realm. When you think about small molecules and how we develop those, when we start working on a small molecule and we think we have something that might be a drug, we do an exercise where we see how would we synthesize it, what would the yield of that synthesis be? Is it the right solubility? All those kind of things. With these advanced complex things like cell and gene therapy, that process hasn’t really evolved yet.
So you really just take what was done in the academic lab and say let’s commercialize it. And so because of that, we knew that there was a lot of inefficiency in that process. And so that was a place where if we directed our tension and we started to think about how do we do that better? How do we say what’s the right promoter to use? For example, how do we optimize this sequence for manufacturability? How do we ask some of these questions that are kind of analogous to small molecule scale up in this space? And so it was really just through a lot of the expertise that we have in north Texas around gene therapy that we knew the right people to talk to, understand those problems
Daniel Levine:
Well, what are the problems gene therapy developers face that form Bio is positioned to address?
Claire Aldrige:
I think probably the biggest one is that when you put something into the bioreactor to make your drug, your commercial, your clinical trial batch lot, and you’re translating the process that is a small scale up process, taking it from say, a 250 milliliter flask all the way to a 500 liter bioreactor. That scale up isn’t just kind of multiplied. You multiply it, the ingredients and you get the exact same output. It’s kind of like when you make a cake at home with the ingredients and then you try to make the largest cake ever for some Guinness book of record and you just multiply that ingredient list and you don’t end up with a cake. That’s just the same because there’s so many other variables. And one of the things I like to joke is these kind of things. These are living factories, bioreactors are cells that make these drugs.
And so it’s a game of collisions. It’s a game of making sure that you have all the right ingredients in the right ratios, and then also that they can bump into each other. And so that doesn’t just scale up easily. And so when you think about it from that perspective and you realize that we need to analyze this with the computational power that we have, because these are kind of collisions and reactions and things that are so many different parameters, the human brain can’t really think of all those parameters at once, but a computer can if we tell it to. So that’s the problem is how do we set something into a bioreactor with the knowledge that what comes out at the other end is going to be the best and safest drug that we could make. And I think when you look at many of the bioreactor runs that we know of through all of these different indications that we’re going after now, if you hold everything constant, you hold the entire recipe constant, the cells, the media, the time, the capsid, all of those things, and all you change is the payload, the nucleic acid that goes inside the gene therapy.
You can get wildly different outputs on the other side, wildly different yield, wildly different number of full viral genomes in the capsid versus filled with backbone or host D n a or other parts that are not therapeutic. And so how do we start to ask the question, why is that? What is it about this payload that makes it more manufacturable? And that was the question that we ultimately decided to try to go after. And then once we understood that a little bit more, how can we then be more forward thinking around the design and so that we can kind of edit out design flaws that we might have, design flaws that we might have in those constructs so that again, what comes out at the other end is as much therapeutic drug as possible.
Daniel Levine:
It’s interesting to think that the payload itself could have such a dramatic effect. It strikes me that for this to work well, you need a lot of data around what’s been done.
Claire Aldrige:
I always like to think of it as hungry, hungry hippos.
Daniel Levine:
The problem though is I would imagine there’s very limited data relative to other modalities on what happens in a bioreactor with a gene or cell therapy. How are you addressing the data and do you have access to data sets?
Claire Aldrige:
Absolutely we do, but these are very data hungry applications. One, we get data sets from our customers, so that’s been an incredibly helpful thing to allow us to train. We also access public data sets that are out there in the public domain, but I think most importantly in one of our differentiators is we are generating our own proprietary data sets. We have some relationships with some academic partners who have freezers full of these kinds of manufacturing runs, the output of them. So we sequence those ourselves. And so we’ve added almost 50 different manufacturing runs and each one of those manufacturing runs has thousands of data points within it. So we are generating our own proprietary data that has a ton of breadth to it. And so we’re looking at different capsids, different promoters, different genes of interest to make sure that we’re not incorporating bias into our data. Again, I think when you think about this space, that’s something you have to be cognizant of from the very beginning is keeping your data sets incredibly diverse so that you don’t inadvertently bring bias into your machine learning and other analysis.
Daniel Levine:
How can people think about the difficulty and complexity of manufacturing these products and what makes them so difficult?
Claire Aldrige:
What makes them so difficult is that the factories that you’re making them in are alive. And so biology is a game of trends, not binary decisions, which makes it harder to model than some other things. But if you think about a small molecule, for example, say it’s got a 16 step synthesis and you try to run each step to completion. So what goes into the next step is as pure product as can go into that next step with these biological factories, you don’t have any control of that. You put in the ingredients and then the cell does the work and the cell moves on to the next step without there being an opportunity to purify or concentrate or anything like that. And so these go on for days, and if there’s an error in the design, like we talked about in days, that error is going to be amplified because you don’t have the opportunity to correct it along the way. So these biological factories that are cells that are making this gene therapy, we’re basically hijacking the way viruses work to use them as, I like to think of them as a FedEx envelope, a FedEx envelope where we put our payload inside.
We are asking the cell to do all of that, and we’re looking at the end. We’re not able to look at each step of the way, which we usually can do in small molecule synthesis. So that really gives us pause because there’s so fewer places where we can intervene. And so you really have to put the right ingredients in at the beginning and the right ratios and have the right temperature so that the likelihood of things getting amplified along the way is reduced as much as it can be. It’s never going to be zero, and so you’re just trying to nudge it in the direction you want it to go in so that that’s what gets amplified. And if you have something that’s got, for example, we know of some drug products where very early in the construct there’s a truncation where the way the D n A, the sequence of the D N A has a particular signature that tells the machinery to fall off, and now you have a truncated version of this. If that is what happens 66% of the time, that’s going to get amplified every time you go into your replication step. And so we’re trying to avoid those things, minimize them as best we can so that what comes out the other end has the best likelihood of being the full viral genome that’s therapeutic.
Daniel Levine:
At the same time, monoclonal antibodies are made in living cells. Is there something fundamentally different about manufacturing a gene therapy in terms of complexity than say a monoclonal antibody?
Claire Aldrige:
It is. And when you follow the trajectory, if you look at monoclonal antibodies, the early days of monoclonal antibodies had these same problems and it was a matter of optimizing the manufacturing process, making some tweaks to the sequence of the FC portion of the monoclonal antibody to improve half-life and all these different kinds of things. So we are in that early stage, but when you think about a monoclonal antibody, you’re asking a cell to make a protein and then you harvest that protein. When you think about a gene therapy, you’re asking a cell to make the D N A that goes in it to make the enzymes that allow that D n a to be packaged appropriately and to make the protein that makes up the capsid. So you have all three of those biological steps that have to happen correctly for you to get a therapeutic molecule at the end. When you compare that to a monoclonal antibody where it’s just one of those three steps, it’s three times as likely to have a place where it can go wrong.
Daniel Levine:
At what point do gene therapy developers generally consider the manufacturer ability of a potential therapeutic?
Claire Aldrige:
Usually a little later in the process than we would prefer because so much of this work is done in discovery labs, like at academic medical centers, they’re looking for efficacy, they’re looking for an effect. They’re looking to develop something that works in an animal model of the disease, usually a mouse model or some cell models, especially cells taken from actual patients with the disease and they don’t have access to the breadth of the tools. For example, a lot of times they’re using this particular promoter because that’s the promoter they’ve got in their lab. That’s what they used on the last one. It’s got the right tissue specificity, the right strength. So they use that promoter, they use this I T R, which are the ends that allow it to get packaged inside the viral capsid because they’ve got it in the lab, so they’re looking to see that it works.
Then that gets transferred to a commercial entity that’s going to see is this something we can develop into a drug for a patient? And if they even consider manufacturing at that point in time, they don’t want to have to repeat all of that work for efficacy. So as a general rule, they might not do any work thinking about manufacturing until they’re actually manufacturing for, and this would be a smaller scale manufacturing than a clinical lot, but manufacturing for your I N D enabling work, and you don’t often know because until that point in time, you’ve always done these small scale bioreactors, benchtop, bioreactors, and you don’t know what’s going to happen when you take those volumes much larger and is there a design flaw that then gets amplified, for example? So you don’t know that until you do it, and those are very expensive experiments to do.
One of the things form has developed is a way to model that. So we have a simulate machine learning process that will actually take your sequence and give you an idea if there are any design flaws in this that would impact your ability to manufacture enough drug to make it work from a cost of goods perspective and safety perspective. And so that’s kind of where they think about manufacturing. The more we learn about this space, I think the more that’s going to go upstream. But this is such a young field, we’ve really just been asking how do we make a drug that works and now we need to ask the question, how do we make a drug that works that we can manufacture and is the safest one we could put in patients?
Daniel Levine:
Perhaps we should take a step back. Can you walk us through the manufacturing process for a gene therapy?
Claire Aldrige:
So just the manufacturing process? Yeah.
Daniel Levine:
How are these made? Yeah,
Claire Aldrige:
So basically you would either have bioreactors in your own manufacturing plant or you would work with a contract manufacturer that have their own bioreactors and perhaps even process. And then what you do is you provide them the D N a, that is your construct, so your genetic payload that has your therapeutic gene on it, and depending on who you’re working with, whether or not you’re doing it in-house, you then have a couple of other plasmids that you would put in that same bioreactor and those plasmids code for the capsid protein and those enzymes that enable the packaging and all of those will go into these large bioreactors. Think of those vats kind of like that you brew beer in. It’s a similar process into some cells. Typically what’s used is heck, 2 93 cells, these are human embryonic kidney cells, and you ask those cells to be your factory.
And so those cells take up those plasmids, start producing the proteins and the nucleic acids as coated by those plasmids. And then inside the cell, the viral capsid assembles gets loaded with your therapeutic payload. And then at some point in time, after some amount of manufacturing, the process is stopped, the cells are lysed and you purify it to get your therapeutic capsid out. You go through in the purification steps, you’re often trying to separate empty from full because some of the capsids will be assemble without anything in them. And then in the ones that are full, you try to figure out how do I separate the ones that are filled with only my therapeutic payload and not anything like contaminants, whether those are host genomes or backbone or chimeric pieces or the rep cap plasmids, which those are the ones that are the helper plasmids, but sometimes that’s very hard to do because they’re all the same size, and so it’s hard to separate things by something other than size when they’re that small. And then once you have that purified product, you do some assays to understand what’s in your product, how much full viral genome do you have, and then you go through a process to get your virus checked off so that it’s ready to be used in clinical trials.
Daniel Levine:
Two concepts in manufacturing a gene therapy that might be useful to explain are replication and packaging. Can you explain what these are and what happens when they don’t occur as efficiently as desirable?
Claire Aldrige:
Yeah. So replication is the xeroxing, for lack of a better word, of your therapeutic construct, your nucleic acid that codes for your gene of interest, the protein you’re trying to correct in a patient, and you have to make tons and tons and tons and tons of copies of that. So you have that to go into all of your capsid packaging is where that therapeutic payload is then put inside your viral capsid to make a whole therapeutic molecule, and that entire drug product is made up of your therapeutic payload covered in this protein coat that allows it to get into a cell.
Daniel Levine:
Another element of these therapies is what’s known as a promoter. Can you explain what a promoter is and what it does?
Claire Aldrige:
Yes, of course. The promoter is the start. It is a sequence in D n A, any D n A that has a protein that’s being expressed needs a promoter to tell the machinery inside your cell bind here and make what comes after. And so promoters are some number of nucleic acids that not only tell the machinery inside the cell to make the sequence afterwards, turn it into mRNA, which then the mRNA gets turned into a protein transcription and then translation or those two steps. It also tells the cell where to express that protein, when to express that protein and how much of that protein to express. So it’s a place in gene therapy actually where there’s a lot of opportunity for advancement and where there’s some nuances in a lot of the drugs that are being developed because you want your protein to express in the right cells.
So for example, if you’re trying to express a protein in the nervous system, you might want to use a promoter that is specific to the nervous system so that if that promoter got into a kidney cell or got into a liver cell, it wouldn’t make the protein because it’s not. The address of that protein is not for those cells. You also might want to use promoters often have regulatory elements associated with it, and that’s where you might say there might be a feedback loop that tells the cell, you have enough of this protein, so stop making it for a little while and then we’ll tell you when we run out because we’ve got this feedback loop will tell you when we run out and when you should start making some more. So that’s another key component to getting these therapeutic genes expressed that you want to be careful of.
And then how much, what you want to be cognizant of is for different diseases, how much drug do you need? For some diseases like cystic fibrosis for example, you only need about 10% of normal in order to correct that flaw in those patients. For other diseases like Rett syndrome, you need to be almost exactly on the physiologic concentration of that protein because if you have too much that is bad for the cells, if you don’t have enough, that’s bad for the cells. So again, the strength of that promoter, and that’s something else you look at in the laboratory before you take this into manufacturing, that’s part of the work that happens usually in the discovery lab. You want to pick the promoter that’s the right promoter, the right strength, the right cell type, and the right timing wise, making sure you get it at the right time when the cell needs it and you’re not making a ton of it when the cell doesn’t. So
Daniel Levine:
For any gene therapy construct, how much variability can there be with regards to the choice of a promoter and is this unique to the cells and the actual gene you’re using?
Claire Aldrige:
They are often more to the cell you’re targeting than the gene you’re using. Usually what you want to do is you want to try to find a promoter that has some sort of what we call tissue specificity. You want a promoter that expresses in neural tissue. If your disease is a disease of neural tissue, you want it to express in heart if that’s what you’re trying to target because that gives you an extra level of control over that gene therapy so that if it gets into other cells, it won’t make too much of that protein and be deleterious or bad for those cells. So there is a lot of options for different promoters and usually what you’re trying to do is just try to make sure that you’re getting it into the tissue that you want to get it into and getting the right amount of protein made. But there’s an enormous amount of flexibility there. And again, that’s something form looks at a lot. We work with customers where they say, these are the eight promoters that we have in our toolbox that we know have the right tissue specificity and strength. Which of these eight promoters is the most manufacturable for this particular gene of interest?
Daniel Levine:
These are costly therapies to produce in part because of the manufacturing yields. What’s known about the ability to use your platform to both accelerate the development process and improve yields? Is there any way to quantify that?
Claire Aldrige:
There are ways to quantify it. It’s very bespoke though, so it’s kind of each different therapy has a different opportunity for optimization, but the first is making sure that the way you’ve designed it is really designed to be the most optimal from an efficiency of manufacturing perspective. And that really comes into play. The more full viral genomes you can package, the more doses you get out of a run. And it’s important to remember that these manufacturing runs cost millions of dollars and they cost millions of dollars, whether you get two doses out of it or 50 doses out of it, and you want to kind of be thinking about that so that what comes out the end has the greatest tighter or potency that you can have. But then the other way is because we have so many sophisticated tools for modeling a lot of this beforehand, we can save gene therapy companies a lot of time. Instead of taking 20 potential constructs into some validation work, we can look at those 20 and tell you which are the best three or you can give us your gene of interest and we can run millions or billions of different variations using different promoters, different ITRs, different regulatory elements, and using other AI powered tools to give you an optimal construct for this process so that you could take, again, a smaller number and a greater likelihood of success when you actually start to do that very expensive wet lab work.
Daniel Levine:
Is it known what makes one gene therapy easier to manufacture than another? Is it the vector, the payload, the combination of them?
Claire Aldrige:
What we have discovered, and this was really interesting, what we knew is there was something about the nucleic acid sequence and what we were able to discover is that there are kind of three dimensional signatures when we think about D n A or r n a nucleic acids in general, we think of them as a string, a string of letters as T C’s and G’S uses if we’re talking about R N a, but in reality, it’s a dynamic three-dimensional that is not static, and so it is changing, its secondary and tertiary confirmation all the time based on what’s happening around it in the cell just based on the fact that things are bumping into it. There’s energy being passed around the system. So what we know is that there are specific nucleic acid sequences that lead to secondary and tertiary structures that are signatures of truncation events.
So they really lead you to understand where is it most likely that the machinery, that replication machinery we were talking about earlier, where is it most likely that that machinery will come up on a snag or a snarl, a traffic jam, if you will, and then fall off. And so we are able to understand where those areas are, and so we are able to then help you design around them or minimize their impact. The other thing we know is that there are also secondary and tertiary structures that are a little bit protective that seem to reduce the likelihood of truncation. Again, biology is never plus or minus or very rarely, plus or minus. It’s all about how do we nudge this towards being the most manufacturable? And so that’s how we analyze this as we’re really looking at that nucleic acid sequence as if it were not just a string of letters, but really a three dimensional structure that we can help improve the design of so that it makes it more replication friendly, which I think is a phrase I just made up.
Daniel Levine:
What would you say form BIO does? How does it make these more manufacturable?
Claire Aldrige:
The basic thing is we help you model it. We help you understand where are those potential truncation events, what are the problems that might be there before you even go into manufacturing. Then we help you swap out swappable elements like promoters, like poly tails or introns, things like that to allow you to take the most well-designed construct into the clinical development path. And then the other thing is then the next level of that is using AI to really generate a new construct sequence that takes into account GC content and synonymous code on substitution to make something that is even more manufacturable and is potentially new IP for your company. So it’s kind of a two-step process. The first is just understanding what can you swap in and out to make it more manufactur to remove any design flaws that might be there for the swappable elements. And then the next layer is now let’s use AI to make the same protein, to make the same therapeutic at the end of the day, but to improve that manufacturing step and weeded out any problems that might be there from triggering the innate immunity or having coons that lead to that signature that lead to truncation.
Daniel Levine:
What are the inputs to your system and what is the output? Do you give them a recipe at the end of the day?
Claire Aldrige:
Yeah, the very basic input is what is your construct sequence, what is your gene of interest and what are the other things you’ve surrounded it with that you think are going to make a therapeutic gene product at the end of the day? Now, we can also take in additional data that’s kind of the very bare minimum. What is your D N A that you’re hoping to put into this process or R n a, depending on what kind of virus you’re trying to make, but we can also take your long read genetic data. We can take other things from your experimental data to pull into that as well to really fine tune the model. But our model is good enough now that if you just gave us your D N A sequence, we could give you a really good idea of what the manufacturability of it is and how you might change it in order to improve that manufacturability
Daniel Levine:
And what’s been done to validate the system. And is there any way to quantify the cost benefits or the yields or the time savings?
Claire Aldrige:
Absolutely. We are validating it in academic labs and with C D M O partners as well as with our customers. So I think I do like to say that as scientists, we are skeptical by nature and we want to see the proof, we want to see the data, and so that’s been an important part of our process, which also then continues to train our models. The more data we get from doing these validation experiments allow us to continue to feed our models and further refine them so that now we actually are really confident in our error bars. Our confidence interval has really improved over the last 18 months, so we are validating this in a number of different ways. We’re also validating it with kind of like checking against the answer key, taking things where we or our customer or our partner knows what the manufacturability of it was and giving it to us and let us give them the score that we came up with, and we’ve done really well at being able to identify those problems theoretically, that we were then able to match up with data that had already been generated.
Daniel Levine:
And what’s the business model? How does form work with its customers?
Claire Aldrige:
It is a business model with a number of different ways that we work with our customers, really depending on where they are in their pipeline. If you’ve got a drug that you are getting ready to take to the F D A to do an I N D opening up your I N D meeting, then we can help you characterize that drug product so that we can take your sequencing D N A sequencing information and tell you how many therapeutic full viral genomes do you have in the ones that are not full viral genomes. What are they made of? Are there any problematic open reading frames with promoters there so that there might be a potential of making a truncated protein that might make a weird either immune response or some sort of constitutively on or off protein. So we can do that, and that is really a basic bioinformatics process that we can help you with today.
If you have a product that you are already committed to that sequence because it is far enough along in its development path and that has a charge associated with it, we can then take you through that entire process of swapping these elements of modeling thousands of different elements so that we can give you the best one with the different promoters and different intros and ITRs and things like that. Or then layering on top of that AI to come up with brand new IP that has been optimized for manufacturing as well as efficacy. All of those, depending on who we’re working with and where they are. Gene therapy, as you know, is a little depressed right now, and so some of our companies don’t have as much capital and so happy to work with them for profit sharing and having a way to work with them from that perspective.
Others were doing upfront payments, others were doing upfront payments plus success milestones. If you think about it, the way a term sheet that you might do a strategic partnership with somebody, there are a bunch of different levers you can pull. There are upfront payments, there are milestones, and there are royalties at the backend. We can do any combination of those that works with the company depending on what their situation is and what they’re comfortable with. We know from some of the large pharmas that we’re talking with, they don’t want royalties. They don’t want things like that. They would rather pay more as an upfront payment and less downstream. Other companies are more price sensitive and they want to pay more downstream, so it’s just a matter of how do we do a combination of those three levers in order to get something that works for us? Both.
Daniel Levine:
AI is becoming ubiquitous within the biopharmaceutical industry. What do you think the ultimate benefits of its use will be with regards to drug development?
Claire Aldrige:
I’ve been working in this space for a long time and we’ve seen computational solutions kind of come and go in and out of favor. I think this time is different because we’ve gotten so much more power computing power behind it, and we’re also, I think as an industry, we’re a little better about understanding what are the right questions to ask, and so I think what we’re going to see is we’re going to see a real acceleration of discovery that we’ll be able to do more work in silico before we do that cost and time intensive work of the wet lab. I think one of the other areas that I’m particularly excited about is the ability to start to mine data in a really sensible way. I sometimes think that there are probably a lot of solutions already out there hiding in the data that we just haven’t been able to tease out yet because we haven’t had the computational power, we haven’t had the way to analyze these thousands and maybe millions of different parameters all at once.
So I think we’re going to see some advances that come out of some of our data banks that we’ve never been able to get those answers out of. I do think that this will always just be a tool for a creative, clever scientist to be able to accelerate or leapfrog where we are as an industry. I think it’s important for us to not forget that this is a tool. It’s not going to replace a scientist. What it’s going to do is allow us to ask slightly different questions and really bring the creativity of the scientist to the forefront so that they can look at a different subset of the information that they’ve gathered and draw different conclusions, maybe some kind of tangential conclusions that then informs that next experiment or informs the next step in the pathway. That’s my hope, my sincere hope, and honestly where I think we’re going, I think that we’ve finally got a comfort with the content and the power, and we just need to make sure that we’re using it to accelerate the creativity and cleverness of our scientists.
Daniel Levine:
Claire Aldridge, chief Scientific Officer for Foreign Bio. Claire, thanks so much for your time today.
Claire Aldrige:
Thank you.
Stay Connected
Sign up for updates straight to your inbox.