The Role of Artificial Intelligence in Accelerating the Pharma Clock: Revolutionizing Drug Discovery and Development

Article written by?Doug Nissinoff

[email protected]

For more insights on the intersection of AI and healthcare,?visit our website ?and?subscribe ?for regular updates.

Revolutionizing the Convoluted Drug-Discovery Process

If you've ever walked around a university campus, you've probably walked past a few laboratories conducting basic research, many of which are actively working on discovering the next generation of pharmaceutical drugs, with the hopes of making healthcare better.

Let's say that one of these labs are successful... how do you move it from the lab to the patient? Well, as it turns out, it is a rather convoluted process...

Why are we so bullish on the application of artificial intelligence to drug discovery? Simply put, AI can be applied throughout the entire process to truly expedite the lengthy 10 year (on average) process:

Basic Research:

• Duration: Variable, can take several years.
• What it is: Scientists study diseases and identify promising targets for potential new drugs.
• How many drugs: Thousands of possibilities.

How can AI make this better?

• Data Mining: AI can quickly sift through vast scientific literature, helping researchers identify relevant disease patterns and molecular targets faster.
• Predictive Modeling: AI can model complex biological systems, potentially identifying new areas of interest before traditional methods would.
• Potential time saved: up to 1 year

Drug Discovery:

• Duration: About 2-3 years.
• What it is: From the many possibilities, scientists choose the best drug candidates.
• How many drugs: Out of thousands, they narrow it down to a few hundred.

How can AI make this better?

• Molecule Design: AI can design new potential drug molecules that might be more effective or safer.
• Enhanced Screening: Using AI, scientists can more rapidly screen which compounds might be effective treatments.
• Potential time saved: up to 1 year

Pre-Clinical Research:

• Duration: 1-2 years.
• What it is: Before testing in humans, the drug is tested in the lab and sometimes in animals to make sure it's safe.
• How many drugs: Of the few hundred, only about 10 make it to the next step.

How can AI make this better?

• Simulation & Modeling: Before real-world testing, AI can simulate how a drug will work, potentially predicting its success rate and safety.
• Data Analysis: AI can analyze pre-clinical results faster, identifying potential concerns or avenues for optimization.
• Potential time saved: up to 6 months

Clinical Trials: Time to see if it works in people!

Phase I (Safety):

• Duration: About 1-2 years.
• What it is: Tests the drug on a small group to check safety and proper dosage.
• How many drugs: Out of 10, about 7 move on.

Phase II (Efficacy):

• Duration: About 2 years.
• What it is: Tests the drug on a larger group to see if it works and is still safe.
• How many drugs: Out of 7, roughly 3 move on.

Phase III (Validation):

• Duration: 2-3 years.
• What it is: Even more people are tested to confirm it's effective, monitor side effects, and compare it to existing treatments.
• How many drugs: Out of 3, usually 1 might make it to the next step.

How can AI make this better across ALL phases?

• Patient Recruitment: AI can analyze patient data to identify ideal candidates for trials, speeding up recruitment.
• Real-time Monitoring: AI can continuously monitor and analyze patient data during trials, quickly spotting potential issues or promising results.
• Data Synthesis: AI can integrate data from various sources, improving the accuracy and speed of results interpretation.
• Potential time saved: up to 2 years

FDA Review:

• Duration: About 1 year.
• What it is: The FDA looks closely at all the data to decide if the drug is safe and can be sold.
• How many drugs: Not all drugs that reach this stage get approved, but those that do can be prescribed by doctors.

How can AI make this better?

• Data Presentation: AI can help in presenting the clinical trial data in a more organized and insightful manner, aiding the FDA review process.
• Safety Monitoring Post-Approval: After a drug is approved, AI can assist in monitoring any adverse effects in the wider population, ensuring continued safety.
• Potential time saved: up to 3 months

In total, from starting research to getting a drug approved, it can take roughly 10-15 years, and out of thousands of potential drugs, only 1 might make it all the way. According to our estimations above, AI could shave up to 4.75 years off of this process!

Why are pharma/biotech companies so interested in AI?

When a company develops a new drug, it's granted a patent. This patent gives the company exclusive rights to sell the drug in the U.S. for 20 years from the date they file it. Think of it as a 20-year countdown clock. However, the journey to turn a potential drug into an approved medication can eat up a significant portion of those years. If it takes 15 years to research, test, and get approval for the drug, the company only has 5 years left of exclusive sales. This exclusivity is crucial because it's when the company can price the drug to not only recover its investment but also make a profit. Now, with the integration of AI, if we can reduce the development time by several years, the company gets a longer window of exclusivity, translating to more revenue. In simple terms: the more time saved in development, the more profitable a drug can potentially be for its creators.

AI-Discovered Drugs: Now in Human Trials

The landscape of drug discovery is undergoing a tectonic shift with the entry of artificial intelligence, and Insilico Medicine is at the forefront of this revolution. Their drug, INS018_055, has now etched its name in history as the world's first generative AI-discovered and designed drug to enter Phase II trials. This anti-fibrotic small molecule inhibitor, fully envisioned and crafted by generative AI, has made promising strides in its early-stage trials. With its initiation in the Phase II clinical studies, which are randomized, double-blind, placebo-controlled, it aims to gauge the safety, tolerability, and preliminary effectiveness of this groundbreaking treatment.

What We are Looking for When Investing in AI-Enabled Drug Discovery Startups: Be a Specialist, Not a "Master of None"

In the AI-driven drug discovery paradigm, the volume, quality, and specificity of data reign supreme. For AI models to attain their highest predictive accuracy and utility, vast amounts of high-quality data are indispensable. In this context, a specialized niche focus offers a distinct advantage. When a startup hones in on a specific area, such as oligonucleotides, it can curate and concentrate on amassing rich, relevant datasets specific to that domain. Such datasets not only improve model accuracy but also streamline the often convoluted training process. Conversely, companies that spread their net wide, claiming expertise from oligonucleotides to gene therapies (i.e. "Masters of None"), face the Herculean task of sourcing and managing an exponentially larger and diverse data pool. This could jeopardize the quality of insights derived from their AI models.

Furthermore, we speculate that the future trajectory of AI-enabled drug discovery will mirror the traditional pharma model. Here, nimble, specialized startups perfect a particular niche and then become prime acquisition targets for larger pharmaceutical and/or AI giants seeking to bolster their portfolios and capabilities. This prospective acquisition trend amplifies the attractiveness of specialized startups from a venture capital perspective. Their clear focus not only promises breakthroughs in drug discovery but also positions them as coveted assets in the eyes of pharma behemoths.

Some of the Key Players in AI-Enabled Drug Discovery

Several key players in the AI-enabled drug discovery realm loom large as the not only the current champions in the space, but also as potential acquirers of specialized AI-enabled drug discovery startups:

Atomwise : This company uses AI for drug design, especially focusing on predicting how different molecules will interact in the body. Their technology is used for both discovering new drugs and repurposing existing ones.

DeepMind (Alphabet's subsidiary): Originally known for their work in creating AI models that mastered the game of Go, they have since applied their expertise in deep learning to solve biological problems, notably protein folding.

BenevolentAI : This company focuses on using AI for drug discovery and development. Their platform processes scientific literature, clinical trial data, and other datasets to identify potential drug targets and candidates.

Exscientia : Based in the UK, Exscientia is among the front-runners in applying AI to drug discovery. They focus on automating drug design to expedite the drug development process.

Recursion Pharmaceuticals : Recursion combines experimental biology with AI in an iterative process to identify promising drug candidates. They've built a massive dataset of cellular images and use AI to analyze these images for insights into disease and potential treatments.

Schrodinger : While it's been around for longer than the recent AI boom (founded in the 1990s), Schrodinger's tools are built around computational simulations of molecular interactions, and they have integrated AI methods into their workflows.

Tempus : Founded by Eric Lefkofsky, Tempus uses AI to analyze clinical data to derive insights for drug discovery and development.

More about Intelligence Ventures

We are an emerging venture capital firm dedicated to cultivating innovation at the intersection of artificial intelligence and healthcare within the United States. Our commitment lies in the strategic investment and nurturing of pre-seed, seed, and Series A companies, fueling their growth and fostering the next generation of industry leaders.

Our first fund, AI Health Fund I, will be focused on companies that use artificial intelligence to increase efficiencies and/or solve computationally intractable problems that place a ceiling on our ability to develop new drug s, advance them through clinical trials, and ultimately diagnose and treat patients. We are industry vertical agnostic and believe that generative AI and more specific ML models can be used to accelerate innovation in biotech, pharma, medtech, and diagnostics.

For more information, visit our website at www.intelligencevc.com or reach out to [email protected] for any inquiries. Be sure to follow us on LinkedIn and Twitter , and subscribe for further installments of The Intelligence Report .