August 12, 2025 | Diagonal Therapeutics founder and CEO Alexey Lugovskoy discusses the key lessons learned over his illustrious career, starting from his childhood in the Soviet Union to founding his own company, Diagonal Therapeutics. With host Tariq Ghayur, Lugovskoy shares insights gained from working with organizations of different sizes, taking on the obstacles of building a pipeline, and integrating AI/ML into discovery programs, as well as his most memorable achievement—and most memorable failure. He also offers advice to young scientists and entrepreneurs, emphasizing the importance of challenging problems and surrounding yourself with the right people.

GUEST BIO

Alexey Lugovskoy, PhD, Founder and CEO, Diagonal Therapeutics 
Alex is the CEO of Diagonal Therapeutics and an Entrepreneur-in-Residence at Atlas Venture. During his 25-year career in biotechnology, he served as COO of Dragonfly Therapeutics, CDO of Morphic Therapeutic, VP of Therapeutics at Merrimack Pharmaceuticals, and Associate Director of Drug Discovery at Biogen. Alex is an author of over 100 patents and manuscripts and an Associate Editor of the mAbs, a journal dedicated to the art and science of antibody R&D. He has received an Advanced Certificate for Executives in Management, Innovation and Technology from MIT Sloan School of Management, a Ph.D. in Biophysics from Harvard University, an M.Sc. in Molecular Biophysics, and a B.Sc. in Mathematics and Physics from the Moscow Institute of Physics and Technology.

MODERATOR BIO

Tariq Ghayur, PhD, Tariq Ghayur Consulting, LLC; Entrepreneur in Residence, FairJourney Biologics
Dr. Ghayur retired from AbbVie (July 2021) and works as an independent consultant. He has 30+ years’ experience leading multi-disciplinary and cross-therapeutic area Biologics discovery programs and developing novel Biologics platforms. Several biologics programs resulted in clinical development candidates. Dr. Ghayur led the team that pioneered the discovery and development of the Dual-variable-Domain-Ig (DVD-Ig) and other multi-specific platforms. Dr. Ghayur also led the team that defined the uptake, intracellular trafficking, and lysosomal degradation of anti-TNF mAbs/DVD-Ig, resulting in the concept of anti-TNF-ADC (next-Gen anti-TNF). In addition, Dr. Ghayur proposed and helped implement several corporate-wide (Abbott & AbbVie) initiatives to foster cross-functional/cross-TA collaborations to bring forward innovative concepts/programs.

TRANSCRIPT

Welcome to the Chain, the podcast exploring the lives, careers, research and discoveries of protein engineers, scientists and biotech professionals. We look at the impact their work is having on the field and where the industry is headed. Tune in to stay up to date on the newest advancements and to hear the stories that are impacting the world of biologics.

Hello everyone, my name is Tariq Ghayur and today I'm having a conversation with Dr Alexey-- Alex-- Lugovskoy, CEO of Diagonal Therapeutics. Alex has had a remarkable career. He is equally proficient in both small molecules and biologics research and development. As a matter of fact, he has worked on more small molecule approved drugs than biologics, and now he plans on shifting this balance to biologics through his current work. So today we will focus on his current work, but first let me talk briefly about Alex's work experience.

For the past two decades Alex has worked in various organizations in leadership roles. At Biogen, he led the physical biochemistry group that provided structural biology, computational chemistry, protein design and assay support to various project teams. Then at Merrimack, as VP Therapeutics, Alex built R&D capabilities for antibody-based drugs and led the development of ERB3 IGFR by specific antibody from concept to phase two. Then he moved to morphic therapeutics, where he served as chief development officer. Here he was responsible for building critical organizational functions, for example safety assessment, regulatory affairs, quality assurance, alliance management, it and data management systems. Here he also co-invented two small molecule integrin inhibitors, provided IND enabling packages and established collaboration with Big Pharma. He also played an active role in Series A and Series B fundraising and IPO. Then he moved to Dragonfly Therapeutics. Alex served as a chief operating officer here and led a group of over 100 strong R&D organizations. He played a critical role in establishing the internal pipeline to complement a network of externally partnered programs.

With this broad experience, Alex joined Atlas Ventures as entrepreneur in residence. Here he helped life sciences entrepreneurs and investors to build breakthrough therapies and companies. And finally, with this broad background, Alex started a company of his own. So now he is the founder and CEO of Diagonal Therapeutics. Recently, Diagonal Therapeutics was recognized by Biospace as one of the 25 most promising upcoming biopharmac companies. So today we will discuss various aspects of Alex's illustrious career and ask him to share some lessons learned. But first, Alex, can you please tell us about your scientific journey, especially when did you decide you wanted to be a scientist?

Thank you, Tariq, for the kind introduction and the first question. Your question brings me many, many years back. Probably we're talking late 70s, early 80s. Growing up I was a voracious reader in a local children's library. I literally read every book from top left to bottom, right shelf sequences and kind of like put there. So I'm a type of a kid who spend more time reading books than chasing the ball and, of course, growing up in Soviet Union, a lot of kind of like this mix of scientific and engineering career was very, very strong.

Growing up I actually thought that I would become a physicist. I read the Feynman physics lectures in junior high. I enrolled in two physics colleges. The first was more focused on electrical engineering. The second one was more theoretically focused. It was more concerned with physics and technology and this desire to do physics persisted.

I think I remember like until my probably third year in college when I was sitting in a quantum physics lecture and there was a very distinguished professor talking about the quest to detect the last lepton, which was a tau neutrino, and it was like a warm May day, I mean like after a Russian winter, you have like the first birds are chirping the flowers. You can see the flowers from the window but the leaves are starting to become green and the professor is talking about like town tree. The estimated like times of the particle leaves is in femtoseconds I think it was like 200 and change leaves is in femtoseconds I think it was like 200 and change. And this point that occurred to me is that I'm significantly more passionate about green leaves and about our neutrinos. And you cannot smell, you cannot feel, you cannot touch, and that kind of prompted me to pivot towards biology and like a lot of young people do, I like dove headfirst into immunology.

So I went to the lab of Professor Nedaspasov studying TNF, ltl for signaling, but coming from a physics background that's an incredible amount of fun but a little bit unstructured. So that kind of coalescence of those two events moved me to biophysics. That kind of like coalescence of those two events moved me to biophysics. So I did my master's research in structural biology, nuclear magnetic resonance, and that brought me into the United States. But I always had this desire not only to observe nature, detect nature, catalog nature, but tinker with nature a little bit. And kind of this understanding of atoms in space, which is the kind of originating physics, ultimately took me to protein engineering oh, that's really exciting.

I didn't know all about that. You know that. Thanks for sharing. So, uh, you know it looks like you went from one field to another and, uh, thought process that went went into it. But based on your these early experiences, experiences, what were the key lessons you think you learned that you would, if you can share with young scientists? You know who are people, young folks who are thinking about their careers.

I mean, I think at some point there are things there are 20 things to do and there are things that kind of like make you go an extra mile right, like seeing, just like touch your unknown voice Because, like in early age, I had like very limited ability to conceptualize my interests, right. It's like in hindsight it appears very thoughtful and appears very logical, but it didn't seem that way day to day. So my advice to people would be just find these things that just kind of tickle their imagination right, and I think the odds of finding those things in the innovation area, rapidly developing areas, is just higher. But at the same time there are well-established techniques, well-established processes and a lot of people derive a lot of joy optimizing and industrializing those processes as well.

But kind of like listening to this inner voice, does this line of activity make me come on a weekend, read a book after the day is done on science as opposed to on science fiction, does it make me want to pursue continuous education? I think that's a very important voice because ultimately that will give probably a bit more perseverance in a fairly difficult field. Scientific field is a difficult field. It's innovation. Most things will never work out. There is less immediate gratification than probably in some other fields of human activity.

So, if I understand you correctly, it's about when you are young and thinking about all these things, to have kind of a broader view of reading different things and see how they integrate and how they connect and things like that.

Exactly right If it seems interesting, right and passionately interesting. So it's probably a good avenue to explore, right, and then, as you explore more, you would decide for yourself what is a good avenue to pursue. Yeah, yeah. And then, of course, the peer group. Like Tariq, as you know, peer group is incredibly important right. Every now and then we're all fortunate just to be in a situation where we're surrounded by high-performing individuals, and that has also a very strong foundational effect on learning.

Absolutely so, Alex. When I described your background and the work, you have worked in and interacted with so many organizations of different sizes, I mean size is really important, including now, including helping new companies, start new companies and now that you have your own company. So one thing can you share with us some insights into? How do you? What have you learned in terms of managing seamless transitions across various functions within an organization? And a related question, I guess, is that how does the size of an organization impact these transitions? You know, the drug discovery is such a long process with so many different specialties, so the transitions are critical.

I mean, tariq, this is a million-dollar question, right, and a lot of management books are kind of like how to manage organizations. But I think that fundamentally, particularly for scientifically inclined people, for people it's very important to understand the purpose, right, it's important to understand the connection to what they do, what organization needs to do, to the needs within the ecosystem. And then drug discovery is invariably patient needs, right? Ultimately, we're developing new medicines and delivering them to the patients who are suffering right now. It's easier to maintain this link in small organizations because all spectrum of corporate activity is just more transparent to people, right? Every piece of data, when your company of three, four or five becomes a communal piece of data, everyone from finance to lab operations go to the same meeting and, as a result, everybody speaks the same language. So I think, from this perspective, in early companies just spending enough of time together discussing things that matter to the company, I think that creates kind of like a common, if you will, currency of exchange. Right, and I think that's why small organizations by definition more cohesive.

As organizations grow, and for the right reasons, they become functionalized. Right, and different functions have different objectives, different priorities. Right, and different functions have different objective, different priorities and different objective priorities introduce structural friction. That is kind of unavoidable. So I think it's important to embrace different functional units. They operate slightly by different rules and how they fit within the organization is not the same. I mean research to development friction is very, very real. Research in this kind of idealistic form is ideation function right, and you ideate, you produce a prototype, and so the fortified prototype is data. Development delivers it in a shape that is acceptable to regulators. Right, which means quality procedures control. You cannot change things very, very rapidly, very different Monday. So I think it's important to understand that some of this friction just isn't going to go away. Right, it's just purely structural and you functionalize to increase your own fidelity on both sides of the spectrum. And I think some way to mitigate it is to kind of like encourage cross-cultural exchange right Again, joint meetings, clear information exchange and again, connecting activities to what organization needs to do to deliver medicine to the patients.

Right, I think by the time you go to multinational companies, the largest company I worked in was Biogen, which I think research was. I mean, the company was like a thousand people. When I joined it was not super big, but I think in the big companies that becomes remarkably difficult because the company is also public right. There are certain things that you cannot disclose to employees before you disclose to shareholders and then communication becomes kind of like canned and that sometimes breeds a rumor meal. And I don't think this is a problem that is quite solvable intrinsically. But I think FaceTime with management is very important even in those settings. But my experience suggests that generally in a smaller company the groups operate in more cohesive fashion.

So in your opinion, is there an optimum size of a company? You know like, as the companies grow, let's say, once they reach 100 people or 200 people, then you spin it out, or something.

I think it's very difficult to maintain a worldwide commercial presence if you're a company of 200 people. It just isn't going to happen. But if you're talking about pure drug discovery and development to proof of concept, we actually kind of see it in the ecosystem. I mean, there's an increased number of drugs that are discovered in biotechs, which are smaller companies that optimize themselves just from idea to first clinical proof of concept, and then pharma picks it up, because you can just build a simple organization to accomplish this task. But I think right now the ratio is like 50-50. 50% of drugs originate in big pharma, 50% of drugs originate in bio. But I think this example speaks to the points that you raised.

Yeah, so, okay. So let's let me ask a few questions about Diagonal, because it's quite an intriguing company that you are building. So first, when did you decide to embark on this journey I mean, start Diagonal and what led you to that decision?

So it's interesting, right? So when you think about antibodies, it's very bipolar in a sense, right, and every time we want to inhibit something or deliver a cargo, we go like, okay, let's use an antibody. Every time we are trying to activate something, we engineer a natural hormone. There is a lot of successful examples like insulin, gf1s, gaps, glps. But there is also a lot of attempts of people to engineer interleukin-2, interleukin-12. I was involved in some of those.

Unfortunately, we didn't quite deliver on the expectations of the patients. Right, there's a life-saving transformative drugs and I always remember that early, and you remember probably that too that early in the years when I entered the industry, which is like early 2000s but probably late 90s, was the debate where antibodies would ever be drugs. Right, the first antibody was credible immune-enchaning. People thought you only would give them once. And then we were saying like, okay, we're going to use receptor traps. Right, receptor traps is going to be absolute best modality. And then Embryol versus Humira happened. Right, and commercially Humira was significantly more successful because it was more convenient to patients. And then the pendulum swung to antibodies. So all this felt that for activators, biologic activators, like clustering antibodies, the same thesis could be brought forward and we actually worked in Biogen on like many people on organisms of CNF receptor superfamily, trlr2, ltbr, you name it. Those molecules were invariably super potent, not very well tolerated, and then the restigenary CD28 happened and an entire field was just like sent back a decade.

But the need hasn't disappeared, right? If you believe that you can engineer antibody but it's a receptor trap, you can probably believe that you can engineer antibody to it's a receptor trap. You can probably believe that you can engineer antibody to cluster receptor compounds that activate the signaling. So this desire persisted for a very long time and I understood for a long time that the task has a significant combinatorial complexity because when you think about activator, your signaling domains which is inside the membrane, are decoupled from where the binding happens and we actually don't have a lot of understanding of how this coupling happens. Maybe there's an exception for integrins, right, or maybe GPCRs to a certain degree.

So always understood that the way to go about this problem is to just build as many possible kind of like agonistic antibody crosslinkers, and if you have homodimetric complex like TNF receptor, that's fairly easy. Normal antibody format allows you to do that. But if it's a multimetric, heteromeric receptor complex, you end up in a world where there is a lot of possibilities. So there is a couple of things that happened towards probably end of 2010. One of them is a lot of IP on how to robustly make bispecific antibodies.

Hence went off patent cleave Knob-center holes with stabilizing disulfide A lot of mutations that will help you to modulate the factor function, because if you're building an activator, you don't want to rely on secondary cross-linking because it makes the activation context dependent. A lot of these mutations became broadly available, so there's an opportunity to aggregate protein engineering tools and the other things that happened. Computational techniques has caught up and, as a kind of like an aid to help you to walk through this universe of possibilities, of agonists in a logical sense. So I think both of the things combined kind of like, give me an inkling that maybe right now is a good time to try this aerial antibody research again.

So when did you start Diagonal?

So Diagonal was incorporated in January of 2022. I started working on this idea probably in late 20, late 2021, but it was percolating in my head for quite some time. Like a lot of these things you know, you're an innovator yourself I mean you just keep on thinking of them, keep on thinking of them and then like, eventually, something crystallizes right.

No, but what was the final event that you decided okay, let's start a company Diagonal to do this? I mean, was it because you were at Atlas?

No, no, no. I actually joined Atlas already with the story in hand, Like I joined Atlas to do Diagonal, and Atlas community was incredibly receptive, for they really liked the idea and they were very supportive all the way through. So this idea predates me joining.

Okay, so why did you choose the name Diagonal? What is the significance of the name?

Well, you start the company and then you start looking for a name and then you realize that all good names are taken, right, all zodiac signs are taken, all constellations are taken. So before you know, you're scraping kind of the bottom of the bottle and I wanted, particularly, given my charming Slavic accent, to give people names that people can actually pronounce and spell. And remarkably there was not a company but a company called Diagro. So and I made it an acronym that stands for Digital Agonistic Antibody Ligands and originally, when we started this company, we thought like well, if you don't like the name, we'll change it. But now, with three years on counting, everybody likes the name. So it probably was like one weekend of brainstorming, coming up with a lot of different names, looking them up, realizing somebody claims this name already, until we come to diagonal, which I think actually describes what we do. Because antibodies bind to the receptor, typically not in plane. You can draw a lot of diagonal lines through the pictorial representation of this complex.

As I was looking at your LinkedIn profile, there you mentioned the focus of diagonal and I'll just quote you, which is that diagonal is developing innovative antibody therapies that target the root causes of diseases by reactivating pathologically impaired signaling pathways inside the cell. So you explained earlier some of the efforts that were ongoing in terms of finding agonists to activate these pathways, but can you elaborate on and you also mentioned that you have these clustering antibodies you know that you are using to activate these pathways, so can you elaborate on this? This is a very interesting concept.

So I'll start with your second point, the reason why it's a classic cluster of antibodies. What we build is the molecules, is that force assembly of the signaling receptor cortex right. Some of these molecules will drive the signaling cascade we're going for, so it will be agonistic antibodies. Some of these antibodies are just not going to drive the signaling cascade right and the job of us as a company to select, select the ones that encapsulate the native signaling in the best way possible. But the idea of using antibodies as a genetic medicine is actually very interesting, particularly for the targets that express is only vascul or in highly vascularized organs.

There is a lot of human diseases and we're working on a couple of those with our lead program that are driven by either haploid insufficiency in the receptor or loss of function in the ligand. And if you can with your antibody in a circumstance like that, and if you can, with your antibody in a circumstance like that, assemble the deficient receptor complex or bypass upstream defect into the ligand, with this clustering antibody you in essence provide any full restoration of the phenotype Because you're reinstating the deficient signaling pathways that link to this pathology exactly where this pathway has been lost. Right and we are hoping, and our early preclinical data supports it that with this approach we will be able to achieve deep disease-modifying effects. For example, our lead program is an is an activator of ALK1 receptor complex. It's a receptor complex of six receptor subunits, two endoglins, two ALK1s, two BMPR2s, and there are two diseases, important diseases, where this complex is impaired.

In a bleeding disorder called hereditary hemorrhagic telocentasia, in a bleeding disorder called hereditary hemorrhagic telocentasia, you have deficiency in one of the copies of ALT1 or one of the copies of the endoglin and you lose the ability to adequately respond to natural BMP9. Our antibody binds to this complex and reinstates signaling in this deficient background and as a result we have strong disease-modifying effects in preclinical models. And another disease where another component of this complex is mutated or downregulated is pulmonary arterial hypertension. And we have shown that with this approach you can achieve the same. So you kind of get into almost like restoration of a normal hormonal action. So that's what we mean by using this new kind of like engineered antibody modality to correct the root cause of the disease. You get antibody to restore the complex that got genetically inactivated and you calibrate the importance of your signal to what has been lost in the disease.

Are you enjoying the conversation? We'd love to hear from you. Please subscribe to the podcast and give us a rating. It helps other people find and join the conversation. If you've got speaker or topic ideas, we'd love to hear those too. You can send them in a podcast review.

So this is quite a it's a lot of understanding of the target, biology and the signaling events downstream and all that. So, in your opinion, what do you think are the current challenges in building a pipeline that we're, you know, having several of these clustering antibodies in your pipeline. So what are the challenges of identifying targets and, whether it's the biology that is limiting or is it the technology that is limiting, how do you look at it?

I think it's fundamentally technology. I think we solved right and we can talk about technology a little bit later. With our technology we feel we can build those clustering molecules in a highly reproducible and robust fashion. It's fundamentally, it's always a biology right robust fashion. It's fundamentally, it's always a biology right. It's a link of your post-in vitro and in vivo pharmacology to the human disease. That's why tackling genetically defined diseases makes it easy because you know what your target patient population and you can do a lot of high fidelity translational research, genetically engineered models, patient derived cells. So I think being in this genetic medicine space helps us to increase this translation fidelity from preclinical to the clinic where we hope to be in about one year time.

The other challenge that you're fundamentally working on novel targets. The other thing that our approach does it unlocks universal antibody targets. Right, we probably targets that you can block are probably very well mined right now, but also targets that can be activated. So while we're going and activating the targets where there is a natural line and posting on cascades that existed, you also need to be very careful that you trigger the pharmacology to the required amount right. So I think this is kind of like this is a typical drug discovery challenge. It's not necessarily unique to the modalityality, but it kind of like brings your antibody research to the 90s, where we're dealing with significantly more novel antibody targets and an average biotech does these days yeah, so.

So, uh, do you think one of the challenges may be the duration of treatment, or you think that may that may not be a critical challenge? You know, like antibody half-lives and things like that, so you can modulate the half-lives, but how long to activate the receptor and and to get the biological effect?

I. I think a lot of the things will have to be figured out in the clinic, right? We utterly do not know. We can study how natural ligand signals We've been very diligent figuring out the difference between all cluster antibodies and the natural ligands in terms of signaling but a lot of these questions about what is an optimal regimen from the patient perspective will have to be clinically determined. I think, as you said, right now we're sitting on a broad toolbox of different engineering solutions that allow you to modulate half-life for antibodies. We also know how to develop antibodies for low-viscose subcutaneous administration, which is more convenient for the patient. There are a lot of tools at our disposal.

When you go for diseases with high medical need, your therapeutic effect has to be primary, objective, right, and it has to be an acceptable arrangement for patients. But it's a slightly different world If you go for best-in-class next-generation anti-TNF I'm sorry to pick up on anti-TNF you probably want to focus on, like twice a year sustained release self-injector, and it becomes a very important component of TPP early on. If you're working on a disease where there is not an approved standard of care, such as hemorrhagic telogenation, you want to deliver effective medicine to the patient first and then think about an indemn optimization of what type of delivery later, once you've solved the first task.

Yeah. So, Alex, I have a couple of questions, because what you're describing is really fascinating, that you know if this approach works. That would be quite, quite amazing and it would be really exciting. So, but you know you were describing earlier on that you started with physics and you have a lot of background in computational approaches and you have also employed AI and machine learning approaches. So, from developing clustering antibody point of view, how is diagonal taking advantage of AI and machine learning? I mean, how does that integrate in your discovery program? That is the first question. And second question is that which is related, that you have also I think you have also developed some really high throughput methods of selecting the appropriate antibodies to convert them into a clustering network. So can you talk about that, because that is really relevant to all the activities that are happening these days?

Yeah, and Tarek, thank you for this question. Before I dive on how we use computation, I just want to kind of like say something that everybody knows but very few people say the use of computation in antibody engineering is like not new, right. If you go to veryiquin, antibody humanization, I think it was like 1989, it was a structure-based method, right. And then you can go to Zencore, Stephen Mayo, the method for FC engineering, that's early 2000s, that was electrostatics based in nature, computational affinity maturation's early 2000s, that was electrostatics based in nature, computational affinity maturation early 2000s. That was happening. So I think the use of computation to allow researchers to be more focused on, say, decision-making, this is not new, right. Can you use computation to streamline engineering? That you can do.

I think what is very different right now, up to very recent, the techniques that we use in computational things. With deterministic techniques, right, you kind of you have an outcome of a computational technique and you kind of can figure out from the first principles why it came to be. It's wonder wells, it's proximity electrostatics, by stacking, you can kind of wink-wink on it and you can see what's happening. Objectives are more statistical in nature, right. They provide a lot of interesting solutions, a lot of those you go like, wow, there is no way it's going to work, and some of them do. So I think that creates kind of like a little bit fairytale environment solution. A lot of those you go like, wow, there is no way it's going to work, and some of them do Right. So I think that creates kind of like a little bit fairytale environment. I think that drives an interest. But fundamentally, I think how we use computation as a field aren't going to shift. It's a powerful productivity enhancement tool.

So now coming back to what we do in diagonal, so if you think about the task of bringing two or more receptor subunits together with an antibody, in the way that the triggers are signaling downstream, with the caveat that you don't know how the coupling across membrane happens A comprehensive or best way to do that would be to cross-link every single epitope with every single epitope on those receptors and come up with this minimal but exhaustive set of cross-linkers. And you can vary leakers a little bit and vary FCs a little bit, but if you're in a situation where there is 20 epitopes on one side and 20 epitopes on the other side, if you can come up with 400 crosslinkers, if there is an antibody that drives the signaling exists. You probably will find it right. So in this exemplification of the tasks that Diagonal has set up for itself, we decided that we will use computation to help us to find this minimal antibody repertoire that would recognize those 20 epitopes on one side and 20 epitopes on the other side. And so we use computation to bridge very deep immune repertoires that we derive immunizing animals display, invariably coupled with next-gen sequencing to get a lot of sequences to this polyepidotic output. And we do this kind of like reduction in complexity of immune repertoire in two steps.

At step one we have, in collaboration with Dima Kazakov, who is a professor in Boston University in Stony Brook, developed a method to very rapidly go from sequence to structure. That's an ML-based method. We kind of know what an average antibody looks like and you can do it right now very, very quickly on a scale of hundreds of thousands of sequences. Then, once we have a spatial representation, we can compare CDR shapes using different, geometrically invariant fingerprinted techniques and the idea is that antibodies with distinct CDR shapes will bind to different parts of the receptors and antibodies with the same shapes probably will bind to different parts of the receptors, and antibodies with the same shapes they'll probably will bind to the same epidote.

So we only take the presumed repertoires of the polyepidotic to the second step, where we in silica would try to match these antibody shapes to the shape of the receptor right. And you can do it with the current techniques on the level of the receptor right and you can do it with the current techniques on the level of the model right. You capture enough of topological kind of distinct surfaces to be able to do that and again you only would pick the modules to go forward that are predicted to bind to unique spots right. So in essence, if you think about it, we use computational techniques in two steps for fully the topic, enrichment of very diverse repertoire. And conceptually this is not very different from how we for years used computational chemistry techniques to screen small molecule libraries, where you can have ligand-based techniques, you can compare molecules based on small strings and then you can dock them and select your kind of like a smaller privileged libraries and actually synthesize and test.

So, for your lead program, how many molecules you identified, based on computational approaches, and then tested them, actually making them?

So we typically had hundreds of molecules. I think it was like if I were to look at the five programs where we have successfully applied our method to date. We're like at 400 to 1100 range because it's very routine that you will find 15 plus epitopes and you want to explore some of the sequence variation within one epitope space and also some like linker FC functionality diversity there as well. So once you build this set of cross-linkers you have to test out experimentally because you don't know which one of them would signal and which one of them would not, and for that we build high-throughput screening assays that allows us to score receptor complementation using high-throughort plate-based assays.

What I was trying to ask you, Alex, is that, as you go through this process of computational identification to biological identification, you are learning, so in a way that the data sets that you're collecting right, they will help you in your future programs. So all that information right? So that's what I was trying to get at.

Yeah, yeah, I think you're exactly right. So within the same receptor topology there is clearly a lot of learning. You very quickly find epitope pairs that drive a stronger signal. What's interesting, depending upon how you cross-sign the complex, you actually could drive somewhat different signaling cascades. But it's probably very difficult to talk about it without actual data. But you're exactly right, it's the same receptor complex.

Once you identify this productive epitope pair, you can go back to your original repertoire and you can recover sequences that we predicted to bind to this epitope pair, and we also will give you a signaling event. And so we put together in the clustering antibody. In our experience to date, like probably one out of six antibodies that you build drive some kind of signal. But the type of signal that you drive through the cascade is very dependent on the exact molecule. We've seen some significant differences in downstream signaling based on what molecular entity we take for.

So I think this experimental testing is very important and the role of computation is just to make this process feasible in a lab, because if you, let's say, recover 100,000 of sequences to two receptors or NGS, if you want to permute all of them, you're going to end up in 10 to the 10th space. It's not feasible experimentally. So the role of computation is to shrink it to manageable sets that you can build and analyze. And the role of computation number two to make sure that if you go through this process and you merge empty-handed, you know that you cannot build an antibody activator.

It's like your process is exhaustive and comprehensive.

You're, in essence, linearizing drug discovery, which is very important in any so let me ask you this question so, with all the work that you have done and you are doing, with all your ideas that you have, what is your vision for Diagonal? I mean, where would you like to see Diagonal in three to five years?

Three to five years. I hope that we are a clinical stage company and we will get delivering our medicine to patients with scleroderma, hemorrhagic telogenization and pulmonary arterial hypertension, and we would very much like to see patients getting better right and getting help with disease. Beyond that, we've been pragmatic about how we build an early pipeline. We're thinking about new program every year or every other year. We have other ideas for clustering antibodies that we can actually use not only for loss of function disease, for gain of function as well. I would like to preserve the small size of organization we currently have 23 people for the reasons that we discussed before. I think that just gives you better cohesion around organizational goals and better familiarity with each other, which just creates a lot of good, positive social interactions and just makes an office brighter.

Yeah, Okay. So I'm going to ask you two or three quick questions and just give me a brief answer. So number one you know, in research we have both successes and failures. Yeah, so what would you consider your most memorable success in your career?

It's a difficult question to ask because and to answer because, first of all, I don't think my career is over yet, so I'm always like so far, so far. What we do in Dino is incredibly exciting, right. These molecules are extorting when you pass through a genetic loss and disease. They're so incredibly potent, right, and we are very excited to see what they're actually going to do in patient population as we progress them to the clinic.

One other thing that I'm very proud of is a molecule P-RAF inhibitor called tovarafenib. I worked on it in Biogen, co-invented it in '06 or '07 time frame, and for a while it was like traveled from one company to another, from one shelf to another. And then Samuel Blackman in day one pharmaceuticals taken it off Takeda's shelf and developed as a drug for kids with pediatric gliomas. So this molecule got approved last April. So that was incredibly rewarding and I'm very grateful to the entire Day One team for making that happen. That happen. You know things that I find incredibly rewarding and seeing people whom you work with to do very well in kind of like the community so that's another thing of being in the community and contributing to community and seeing great people do great things that's also something that just gives me an extra jolt of energy when I need it.

All right. So on the opposite side, what is the most memorable failure that still haunts you?

I don't think it haunts me, but it's the target that alpha-beta-6 is one of the strongest in the internet and one of the strongest drivers of fibrosis. I worked on this target three times, two modalities in two companies Biogenes stromatics, alpha-beta-6 antibody called 3G9. Unfortunately the molecule didn't find some margins in the clinic. Then we worked on it in Morphic with two generations of molecules. That was a partnership with RV and again we couldn't conquer the target. So there is a disconnect between the promise of the target right and an ability to deliver on the promise and it's easy to convince yourself that you've done just enough. But there is this like inkling still right that there is a harvestable, useful biology for treating human disease that just for whatever reason, you miss. So I think this feeling still persists. I'm not saying I'm going to work on alpha-bipedal six and die and I'll save it for a later company, but this is like one of the biology which I would say remains uncomfortable, right and frustrating for yourself.

So, Alex, quickly you know. So you know, one of the goals or objectives of these chain podcasts is to also provide guidance to our lessons learned, to young scientists and entrepreneurs. So, with your such a diverse background, do you want to share any final advice or anything? Or let's share something lessons learned for younger, younger scientists or entrepreneurs my advice is simple, but it's difficult to implement.

Right work on challenging problems of strong pharmaceutical relevance, surrounded by a good, good group of people. You learn better that way okay.

So, Alex, uh, with that, I think I would really like to thank you very much for sharing all your ideas and your wisdom and your experiences, and we wish you and your family all the best and all the success, and I'll be looking forward to Diagonal's success and the molecules that you're moving forward. So, thank you very much.

Thank you so much, Tariq, for a wonderful conversation.