Can you introduce Absci and give an overview of how the company came to life? 

Absci did not start as a generative AI drug creation company, but as a cell-line development company aiming to change how antibodies were made. The initial idea was to use E. coli to produce antibodies. Ultimately, this would lower costs and decrease the amount of time to get drugs into the clinic. I graduated a year early from school, moved back in with my parents, and got a basement lab up and running. Many sleepless nights and a lot of failed experiments later, I was finally able to engineer an E. coli microbial cell to — for the first time ever —produce an antibody. What I did not realize then is that E. coli was going to be the hero of Absci’s story for a different reason: producing the biological data needed to apply generative AI to drug discovery. 

Our company was able to create our E. coli platform and a proprietary screening assay, ACE. Combined together, these technologies allow you to look at the protein-protein interaction and the functionality of billions of antibodies in a short period of time. Our model is very simple, it is data to train, AI to create, and wet lab to validate. This platform is able to identify billions of protein-protein interactions in a week for machine learning training, our AI then creates novel designs for antibodies, and our wet lab can then validate over 2.8 million unique AI designed antibody candidates in a given week. It is typically a 6 week cycle from data to train to wet-lab validated AI antibody designs. 

That’s an industry milestone: the first technology of its kind and with this cycle time. This throughput has led to huge advancements. Today, Absci is at the forefront of merging biology and AI to meet our vision of designing biologics and antibodies at the click of a button. This is the future of where the industry is headed, moving from a paradigm of drug discovery to drug creation - instead of finding the needle in the haystack, we are actually creating the needle. 

It takes approximately five years to go from an idea to a physical drug into the clinic, and the success rate for R&D is historically less than 5%. This has not changed significantly over the past 30 years. We need to strive for significantly better success rates in drug discovery and development, and being able to apply generative AI is the disruption which will lead to the solution. 

Being able to go directly from target to antibody at the click of a button will decrease the amount of time to get into the clinic, and increase success rates. We are changing the paradigm and are striving to get better drugs to patients faster. 

What is the general attitude in the sector when it comes to modern technologies like AI?

The industry is continuing to wrap their head around AI and how it can be applied to drug discovery and development. Pharma is recognising the importance of AI and is realizing the need for the  integration of AI into their processes. We are collaborating with pharma to pursue bigger and bolder visions and to make use of AI in drug discovery and development a reality. 

What companies tend to show more openness to integrating these types of AI solutions?

Absci has partnered with Merck, a company that is science-first. Our values and vision align with that of Merck, and we have sufficiently demonstrated the success of our technologies. This is why they decided to partner with us. 

Looking at recent deals in the industry, we are seeing M&A becoming more AI focused, and companies are starting to invest heavily into this space. However, there are still many pharma companies lagging behind and still figuring out what their strategy will be. I believe that those who have the strategy now and have been implementing it for some time already will be best-placed to become the leaders of the future.

What differentiates Absci from competitors in the AI space?

Today, there are many emerging AI drug discovery companies, but most are focused on small molecules. The reason for this is access to data. Many companies can take a million member small molecule library, screen it, get the data, and train their models to develop a drug in pill form to go to market. The same is not true for biologics and developing antibodies. This is because there is a data scalability problem: every antibody you want to screen must be produced within a living organism. 

If you want to test and screen different antibodies, you need a different cell to make each antibody. Currently, you can scale that to approximately 10,000 antibodies a week, but that is just not enough data to ultimately feed and train your models with. 

Prior to getting into AI, Absci spent 10 years developing wet lab technologies and our ACE assay. This was necessary to spawn enough data to feed into our models. The beautiful part about working with a microbial host like E.coli is that you can take a test to an engineered E.coli, encode a billion different antibody sequences on DNA, and have every single E.coli make a different antibody that you can then screen for their functionality. 

For screening these large quantities of antibodies, we developed our proprietary ACE technology, where we are able to interrogate every cell and look at the protein-protein interaction data to measure features such as binding affinity and Naturalness score, both of which are key factors for determining how an antibody might fair in clinical trials. We can then feed this data into our models. With our innovations, we have been unlocking the scalability of biologic data. This has not  been done previously by anyone working on applications of generative AI. 

Today, Absci is the only publicly traded company that is focused on generative AI drug discovery for biologics. Not only do we get data at high throughput, we also have very rapid cycle time; billions of protein-protein interaction data points go into the model and then validated designs come out. Our innovations allow us to validate 2.8 million AI designed antibodies in a given week. These hard numbers are testament to our success and are what differentiates us from the rest of the industry. 

I understand Absci’s has a vision to develop its own drug creation pipeline?

We recently brought Andreas Busch on board. Andreas was previously the head of R&D at Shire, and prior to that at Bayer. He is one of the most prolific R&D executives out there and has had over 10 drugs approved under his leadership. He is now our chief innovation officer, and is focused on building out our own portfolio and assets. We aim to become the Google search of biologic drug discovery, but we also want to develop our own pipeline as we strongly believe in the technology we have developed.

Our own pipeline is already in development and we plan to have an IND submitted in 2024.

We believe we are going to be the first company to be able to design an entire antibody on a computer from scratch using zero-shot generative AI, and bring it to the clinic. If and when we achieve this, we will be at the nexus of a pivotal point in the history of the biopharma industry. 

What are Absci’s main objectives for the near future? 

It was a huge breakthrough for Absci to demonstrate that we can design De Novo antibodies using zero-shot generative AI. We are using this work to continue to drive partnerships with large pharma. We will continue developing our pipeline, and we will also continue to build out the team as we move forward.