
2022-Korea-Showcase-Aria-Pharmaceuticals

-
Interactive transcript
[NON-ENGLISH SPEECH]
Aria Pharmaceuticals is a pre-clinical stage pharmaceutical company. And what we have is a proprietary artificial intelligence platform that allows us to decode very complex biology in ways that others can't. So compared to our last speaker Roy, his deep expertise is in the chemistry of creating new medications. And our work is in the biology of creating new medications.
My history-- I studied bioinformatics at Stanford University, which is where the technology behind the company got started and co-founded the company with a gentleman of a very similar name who got his MBA at Sloan. Green button. I want to go back.
So the problem, if you will, in understanding very complex disease pathology is that these data sources that we utilize to better understand how these diseases operate are very limited. And I say that because there is no single source of biomedical data that tells you everything you need to know about disease.
And so, as a result, what we have is lots of very different and siloed data sources that each give us some small picture about the overall disease pathology. Now analyzing these data sources is critical to understanding how these diseases operate. But it's also very challenging because current artificial intelligence methodologies do not process these heterogeneous and orthogonal data sources all in one go.
And so what our technology is able to do is process all of these very different data sources about disease pathology as a single unit. And so what other people typically do when they're looking at collecting data and using that to better understand disease biology is they look at one single data source and the hypothesis being that, if they believe that they can collect data about a particular disease and that single data source can reveal some new information about that disease, they can ultimately make a new discovery that can lead to a new medicine.
Now I'm not saying these approaches don't work. They most certainly do. But they are limited. And they're limited in that you have to make sure that the data you're collecting actually has the solution to the problem that you're trying to solve. And in very complex disease areas, we may not even know what is the right biomedical data to go after to better understand that disease.
And so what our capabilities are they involve, first of all going, after as many orthogonal and unrelated disease data sources as we can, typically dozens of these. And what our technology does is, first of all, identify, if you will, puzzle pieces that are represented in these different data sets and then, ultimately, put them together and get a better understanding, a more complete picture of disease biology.
So to put this in maybe very common terms, you could imagine having a puzzle piece for a puzzle. And you've only got one puzzle piece, but it's a very crisp and clear view of the picture of a face of a rabbit. So you don't need the rest of the puzzle pieces to recognize the creature that you're looking at. But what happens if that very crisp view is of the sky or the ground? That doesn't really tell you much about the creature that you're dealing with.
In our approach, we don't necessarily need a very crisp view of the face of the rabbit. But if we have the outline of a rabbit in these puzzle pieces, we've got a very good understanding of what that disease pathology might be. We're able to do this because we have solved a computer science problem. And the computer science problem that we have solved is how to incorporate heterogeneous data directly into deep learning models.
So existing AI algorithms today-- this is a typical deep learning model-- they can only incorporate one type of data up front into the input layer. And in our system, we actually inject all of these very different data sources right into the middle of the deep learning models. And why that's important is because that means we can process all of this data as a single unit.
Other people might say something like, oh, we collect lots of different data too and process it. But the reality is they're processing these things one at a time, and then they're looking for overlapping evidence. So we've been able to use this technology to build a very extensive pipeline of therapeutics. Our lead programs in lupus, IPF, and CKD are approaching IND-enabling studies.
We have done also additional partnerships with folks here in Korea with both SK and with 1st Bio. We've got a number of different disease areas that we're working on that are in the earlier stages. And ultimately, I can't necessarily in seven minutes show you preclinical results from all of our programs. If you're a pharmaceutical scientist, this is a very exciting slide. If you're not, don't worry. I'll explain to you what we're looking at.
So this is some of the preclinical results in our lupus program. And very briefly, lupus is a disease where immune system basically attacks your organs. And it can be obviously very problematic because, first of all, lupus has no cure. But the best drug we have to combat these severe flare-ups is a drug called cyclophosphamide. Now it is highly effective at treating lupus, but it has some very toxic side effects.
So as a lupus patient, you can't take this drug very long. So what we're looking at here is this is kidney function. And what we're looking at is this reference compound cyclophosphamide. Can we actually produce similar results in terms of the efficacy but without the toxic side effects?
And so we've been able to do that with a couple of our compounds-- 711 and 712. 711 is basically the same statistical significance as cyclophosphamide but without the anti-proliferative properties that make cyclophosphamide so toxic. And so this new experimental medication has the potential to become the new standard of therapy for this disease. And we, of course, work with several disease experts and luminaries who are very excited about this invention.
So in terms of what we're here to talk about with folks in Korea, we are in the business of outlicensing these molecules as they move forward. These are NCEs that we wholly own, that we filed patents on in each of these disease areas. We have several more diseases that I'd be happy to tell you more about in person.
And we're very interested in doing regional licensing rights, specifically in Korea or Korea, Japan, and China something like that and happy to chat with you more later today about those opportunities. Nice to speak with you.
[APPLAUSE]
-
Interactive transcript
[NON-ENGLISH SPEECH]
Aria Pharmaceuticals is a pre-clinical stage pharmaceutical company. And what we have is a proprietary artificial intelligence platform that allows us to decode very complex biology in ways that others can't. So compared to our last speaker Roy, his deep expertise is in the chemistry of creating new medications. And our work is in the biology of creating new medications.
My history-- I studied bioinformatics at Stanford University, which is where the technology behind the company got started and co-founded the company with a gentleman of a very similar name who got his MBA at Sloan. Green button. I want to go back.
So the problem, if you will, in understanding very complex disease pathology is that these data sources that we utilize to better understand how these diseases operate are very limited. And I say that because there is no single source of biomedical data that tells you everything you need to know about disease.
And so, as a result, what we have is lots of very different and siloed data sources that each give us some small picture about the overall disease pathology. Now analyzing these data sources is critical to understanding how these diseases operate. But it's also very challenging because current artificial intelligence methodologies do not process these heterogeneous and orthogonal data sources all in one go.
And so what our technology is able to do is process all of these very different data sources about disease pathology as a single unit. And so what other people typically do when they're looking at collecting data and using that to better understand disease biology is they look at one single data source and the hypothesis being that, if they believe that they can collect data about a particular disease and that single data source can reveal some new information about that disease, they can ultimately make a new discovery that can lead to a new medicine.
Now I'm not saying these approaches don't work. They most certainly do. But they are limited. And they're limited in that you have to make sure that the data you're collecting actually has the solution to the problem that you're trying to solve. And in very complex disease areas, we may not even know what is the right biomedical data to go after to better understand that disease.
And so what our capabilities are they involve, first of all going, after as many orthogonal and unrelated disease data sources as we can, typically dozens of these. And what our technology does is, first of all, identify, if you will, puzzle pieces that are represented in these different data sets and then, ultimately, put them together and get a better understanding, a more complete picture of disease biology.
So to put this in maybe very common terms, you could imagine having a puzzle piece for a puzzle. And you've only got one puzzle piece, but it's a very crisp and clear view of the picture of a face of a rabbit. So you don't need the rest of the puzzle pieces to recognize the creature that you're looking at. But what happens if that very crisp view is of the sky or the ground? That doesn't really tell you much about the creature that you're dealing with.
In our approach, we don't necessarily need a very crisp view of the face of the rabbit. But if we have the outline of a rabbit in these puzzle pieces, we've got a very good understanding of what that disease pathology might be. We're able to do this because we have solved a computer science problem. And the computer science problem that we have solved is how to incorporate heterogeneous data directly into deep learning models.
So existing AI algorithms today-- this is a typical deep learning model-- they can only incorporate one type of data up front into the input layer. And in our system, we actually inject all of these very different data sources right into the middle of the deep learning models. And why that's important is because that means we can process all of this data as a single unit.
Other people might say something like, oh, we collect lots of different data too and process it. But the reality is they're processing these things one at a time, and then they're looking for overlapping evidence. So we've been able to use this technology to build a very extensive pipeline of therapeutics. Our lead programs in lupus, IPF, and CKD are approaching IND-enabling studies.
We have done also additional partnerships with folks here in Korea with both SK and with 1st Bio. We've got a number of different disease areas that we're working on that are in the earlier stages. And ultimately, I can't necessarily in seven minutes show you preclinical results from all of our programs. If you're a pharmaceutical scientist, this is a very exciting slide. If you're not, don't worry. I'll explain to you what we're looking at.
So this is some of the preclinical results in our lupus program. And very briefly, lupus is a disease where immune system basically attacks your organs. And it can be obviously very problematic because, first of all, lupus has no cure. But the best drug we have to combat these severe flare-ups is a drug called cyclophosphamide. Now it is highly effective at treating lupus, but it has some very toxic side effects.
So as a lupus patient, you can't take this drug very long. So what we're looking at here is this is kidney function. And what we're looking at is this reference compound cyclophosphamide. Can we actually produce similar results in terms of the efficacy but without the toxic side effects?
And so we've been able to do that with a couple of our compounds-- 711 and 712. 711 is basically the same statistical significance as cyclophosphamide but without the anti-proliferative properties that make cyclophosphamide so toxic. And so this new experimental medication has the potential to become the new standard of therapy for this disease. And we, of course, work with several disease experts and luminaries who are very excited about this invention.
So in terms of what we're here to talk about with folks in Korea, we are in the business of outlicensing these molecules as they move forward. These are NCEs that we wholly own, that we filed patents on in each of these disease areas. We have several more diseases that I'd be happy to tell you more about in person.
And we're very interested in doing regional licensing rights, specifically in Korea or Korea, Japan, and China something like that and happy to chat with you more later today about those opportunities. Nice to speak with you.
[APPLAUSE]