How Pharma Companies are Using AI in Drug Discovery

From identifying disease mechanisms, finding potential drug candidates, testing drugs, performing clinical trials, and getting approval, AI is helping everywhere. But what about the success rate and how big companies are using it? We have covered all of your questions. 

Artificial Intelligence in Drug Discovery: 3 Key Areas in the Spotlight

The drug development pipeline has been traditional for years, and it will not change at least the prior steps: identifying the target for drug interaction, designing a molecule to understand how the target works, creating and testing the molecule, which includes multiple refinements, and then conducting a human trial. 

Most of this is repetitive and manual. The result is mostly a failure because a drug that works in the lab might fail when tested on real bodies. This is because of biological differences between labs created and real environments. 

This is only for developing one drug; to identify the best fit, 20 such drugs are developed. It takes cost, time, and effort to repeat the steps for every drug. However, computational technologies have been improving the process, but they are also slow and lack stimulating real-world conditions. 

AI automation is helping in all three areas; let's see how. 

1. Finding The Right Target 

Pharmaceutical companies are using AI technology and natural language processing for data mining. This process screens data from thousands of research papers, academic papers, and publications on biological data, previous clinical trials, and patient medical records. 

All the learning is then encoded in knowledge graphs to feed the ML models. These models identify the most promising target for treating the disease. With this targeted approach, scientists save time on the trial-and-error method in early drug discovery phases. 

2. Predicting Drug-Target Interaction 

The human body is complex, and everyone is built differently. Therefore, the real challenge is knowing how a particular drug will interact with patients' different compositions. 

So far, scientists have lived up to the results of lab tests to understand this interaction. However, most of the innovation in healthcare is targeted at this phase of drug development, helping scientists design molecules with the best success rate. 

Healthcare companies are using Generative AI in drug discovery to design new proteins and understand the interaction between drugs and those proteins. This saves time and cost on physical testing while reducing the side effects of new drugs. 

3. AI-Driven Clinical Trials

The last stage of the drug development pipeline is the riskiest and costliest of all; after all, it directly impacts patients' lives. Traditionally, it takes 10-20 years for a single drug to reach this stage, and most of them fail here. 

Imagine all the hard work and money just to experience failure. The pain is even more when it doesn't reach the patients who are suffering in the hopes of getting cured one day. 

Pharmaceutical companies use AI to make the process less risky and cost-effective by finding the right candidate for the trial. This means companies don't have to continue testing drugs that might not perform. 

Sometimes, even a drug that failed multiple times on certain groups of patients might 

work for others. AI can help identify these possibilities by matching the drug with the right patients. The key to achieving this is involving patient data to train the AI model early in drug development. 

What is the Success Rate of AI in Drug Discovery?

Although AI in drug development is new and early, the success rates are promising. Research published in the magazine Drug Discovery Today says that AI-discovered drugs showed an 80-90% success rate in the first phase of clinical trials. In the second phase, the success rate matched the industry average, i.e., 40-60%. 

Leading Pharma Companies Using Artificial Intelligence in Drug Discovery

Seeing the potential and cost-benefit of AI in healthcare, several companies in the US are using this technology. Here are the ones leading the charge. 

1. Pfizer

Pfizer has used AI even before it came into the limelight for other industries. It is using AI technologies like machine learning, deep learning, NLP, and more to:

• Understand disease biology and its link with the symptoms.
• Know how patients from different testing groups respond to diseases and treatments.
• Generate clinical documents for faster drug approval. 
• Screen millions of drug compounds. 

2. Exscientia

Exscientia, a Biotech company in the UK, uses AI technology to develop effective medicines. The company is reducing drug design time by 70% and cost by 80%. 

Here's how. They are using Generative AI in their design-make-test-learn solution (DMTL) and AWS cloud for 

• Designing drug compounds in the cloud, which are then made by lab robots. This minimizes expensive lab experiments.
• Designing potential drug candidates that match with target product profiles. 
• Optimizing the upcoming design cycles by analyzing data from previous design cycles and experiment results. 

3. BenevolentAI

BenevolentAI, a biotechnology company, uses artificial intelligence to select the best drug targets. Their AI systems analyze data from scientific literature, genomic data, and clinical trial results to train the models on real information.

The company collaborates with AstraZeneca, Novartis, and GKS to identify targets for treating idiopathic pulmonary fibrosis and chronic kidney disease. 

Final Thoughts

AI has already made its way into the healthcare industry, from personalized medicine to automating administrative tasks. However, drug discovery is the initial stage, which promises to make the most impact. 

The trial-and-error method is now experiencing AI-powered automation. However, it will speed up only a few process parts, like drug screening. Around 40-50% of the time can be saved with AI, whereas costs can be greatly reduced. 

Several leading companies have already entered the world of AI-powered drug discovery, showing the path for small and medium ones. The future is bright, but complications like how pharmaceutical companies use patient data in their processes will always exist. FDA approval will also change as more companies start using AI. 

To successfully overcome these challenges, pharmaceutical companies must hire experienced AI solutions providers who understand both AI and healthcare.