I recently authored an article in Life Science Leader discussing my own misperceptions and preconceived notions about applying AI to an extremely challenging and complex process, drug discovery. Like many, my 30-year career in pharmaceuticals was rooted in the tried-and-true methodology of the traditional R&D process, with wins few and far between. But my dissatisfaction for the lengthy and often unsuccessful process to bring new molecules to market reached a heightened sense of urgency in 2017 when I started consulting for Aria Pharmaceuticals and realized there might be a better way. I wanted to take a moment here and share what exactly drove me to the conviction I now have.
During my first few months as a consultant for Aria it was apparent that their approach to early-stage research was unlike anything I had ever seen before. They had completely reimagined the entire drug discovery process from the ground up.
Aria’s approach provides unprecedented efficiency and predictability at the very beginning of the R&D process, where we build an in silico model of the disease using vast amounts of genomic, phenotypic, and chemistry data. Unlike the traditional R&D approach, our computational methods and algorithms identify disease features that can often be missed by discovery methods that rely on directed exploration of a single predefined hypothesis. Then we cross reference the disease features with our library of drug and drug-like molecules, screen out candidates mostly likely to fail and produce a rank-ordered list of hits with predicted efficacy signals.
All of which is accomplished in just a matter of weeks — a stark contrast to the years I was used to seeing with traditional methods.
In fact, Aria’s unique approach allows us to finish in vivo testing of multiple novel hypotheses within 6 months, while delivering a 30% success rates for these efficacy predictions — well above the dismal but widely accepted one to two percent success rate of traditional methods.
But what exactly makes this approach so much more successful than the linear methodology of typical R&D? The concept is surprisingly simple.
Through our approach, we are able to better understand the complexities of biology and disease, because we can process massive amounts of data efficiently and provide scientists and researchers with a rationale for each prediction in a very short period of time. During which, we are able to identify not just one potential target but find nine or ten different targets in the very first step of the discovery process.
So, with all this in mind I decided to join Aria full time as Senior Vice President of R&D.
Let’s take this one step further and break down our lupus program for example. Lupus is a disease that is highly complex and heterogenous with a high unmet medical need and in this case, we started our program with an initial analysis of more than 2.5 billion pieces of data. From here, scientists were able to identify molecules with predicted efficacy that represented targets and mechanisms which had never been examined in lupus patients before. Using standard animal models, we began in vivo testing on nine selected molecules, each representing a new mechanism of action, only four weeks after starting our program.
Early results for two of the nine candidates showed significant improvements in organ function and significant decreases in inflammation, with a strong indication of good tolerability. Currently, Aria is looking to advance both candidates, possibly providing a new chemical entity for lupus treatment.
And this is just one example of potential success. Our research shows us that this approach is applicable for more than 1,000 diseases. Aria has already modeled nearly 300 of them and have more than 18 programs in development.
Now when I look at the traditional R&D process, I see endless opportunities where we can speed things up and find greater efficiency. But here at Aria we made the conscious decision to focus on the very beginning of the research process in order to reimagine the entire drug discovery ecosystem from the ground up.
If you too are a former skeptic who believes more can be accomplished in the R&D space through innovation and collaboration, I encourage you to reach out to me directly. We’d love to build with you.