Standing on the Shoulders of Giants: The Art & Science of Machine Learning & AI for Drug Discovery & Repurposing

Contributed by

face
main_img

DataArt’s Daniel Piekarz looks at how machine learning and artificial intelligence can play into the important task of drug repurposing, potentially providing solutions to the most pressing healthcare concerns of today, including the COVID-19 pandemic.

 

“Standing on the shoulders of giants” is a metaphor, made famous by Isaac Newton, which implies that progress is made based on previous findings and knowledge of those who have come before.

 

Optimizing the efficiency and cost-effectiveness of drug discovery has never been more critical, with the median cost of advancing a single drug from concept to market estimated at nearly USD 3 billion, according to a study in the Journal of Health Economics. And with a 12-year approval process and odds of success at only 1 in 5,000, that cost becomes even more significant.

 

Drug repurposing, also known as drug reprofiling, reusing, repositioning, and rediscovery, builds on available data and has become an increasingly popular alternate development strategy, offering substantial benefits in both efficiency and cost reduction. Who can deny the success of Botox, North America’s leading multi-billion dollar cosmetic procedure? After starting out as a treatment for painful eye problems, it is now approved as a mass-market treatment for muscle stiffness, migraines, wrinkles, overactive bladder, excessive sweating, and more. Of course, one of the key benefits is the tried and tested proven safety profile of the medication, which is why it is being recognized in all areas of medicine.

 

However, traditional methods used for drug repurposing research are insufficient – often too computationally expensive, and not scalable enough to meet modern challenges. Pharmaceutical companies are still sending validated drug-like structures to organic chemists to synthesize.

 

So where do we go from here?

 

We’ve seen a growing potential for technologies such as artificial intelligence (AI) and machine learning (ML), which have been introduced to automate segments of the process involved in drug repurposing, to reduce costs, minimize errors as well as speed up the results. The application of AI and ML enables large-scale screening of compound libraries to identify potential therapeutic opportunities with less adverse reactions.

 

The implications are significant given that machine learning in fact learns about the domain and is well-positioned to handle the complexities, sieving through billions of molecules to identify patterns and chemical structures. With the enormous impact of subtle changes to drug molecules, AI can be used to digitize such changes and see the difference driven by a human scientist.

 

Existing incentives to support investment in the development of repurposing drugs include three years of market exclusivity in the USA and ten years in Europe.

 

The ideal AI and ML solution fuses mathematical and biological knowledge to build the consensus estimates of a compound in seconds, something that would take human teams hours, days, or even years. All calculations are done using datasets augmented with the basis data which is a genuine source of pharmacological knowledge. With the growing availability of big data through electronic capture and a plethora of biomedical data, pharmaceutical companies with access and clinical development capabilities are well-positioned to take advantage of the benefits of AI for effective and data‐driven repurposing. And the good news is that organizations do not have to rely on building out AI solutions in-house or from the ground up.

 

Ultimately, the goal is to leverage technology to accelerate innovation through drug repurposing that complements de novo medicine development – not replace it. Today when we talk about the acceleration of new drug research, the word “acceleration” is not just an empty marketing slogan. The pandemic of SARS-CoV-2 taught us that we must be ready to react fast. The standard approaches have capitulated in the face of a new virus, though AI-based methods have not yet provided impressive insights relating to SARS-CoV-2. AI-augmented drug repurposing must be developed further to become the front-line weapon in a war with the emergent hazards. There are a plethora of AI approaches to drug repurposing, and the one used in our research is a prospective candidate for finding key drugs in our fight against SARS-CoV-2.

 

To explore the results of a recent implementation of AI technology to generate bona fide leads, contact me via email or LinkedIn to request a copy of the full report titled, Antithyroid Drugs Repurposing: The Data Driven Approach.

Add Your Comment


You must be logged in to post a comment.

Related Content

Latest Report