AI, or Artificial Intelligence, is changing how scientists and researchers develop and discover new drugs. Usually, drug development takes many years, even decades, consisting of multiple clinical trials, hundreds of millions of dollars dumped into development, and thousands of chemical tests to determine side effects and success. Now though, with the help of AI, researchers are able to cut down on their time exponentially, and the advanced algorithms are able to run through thousands of different possibilities to see what humans might miss and in some cases, create new combinations we never would have thought of in the first place.

AI is able to aid and boost drug discovery in many ways. For starters, it helps with identifying the ideal target for the drug to be delivered. Through a plethora of collected data on genes, compounds, and biological reactions, AI can efficiently sift through medical data to figure out what works in the body and what doesn’t. AI then proceeds to simulate millions of different combinations to figure out which ones are the most effective with minimal side effects, as seen in Google’s AlphaFold and Microsoft’s Bio-Emu. An example of this is the “de novo drug design” where AI creates completely original molecular structures based on certain requirements of the compound such as toxicity levels and chemical binding properties..

A recent breakthrough from MIT highlights this potential: researchers used generative AI to design two new antibiotics that successfully killed superbugs such as gonorrhoea and MRSA in lab and animal tests. The AI explored over 36 million compounds, and assembled molecules atom-by-atom to meet specific antibacterial requirements, which resulted in powerful antibiotics that easily combatted these diseases. This marks a major step forward in combating superbugs, which are increasingly resistant to traditional antibiotics AI is also used to predict how a drug will behave in the body. It can model how the drug is absorbed, how long it stays active, and whether it might cause harmful side effects all through a simulation This allows scientists and researchers to cut out the weaker possibilities and focus on ones that show results. In clinical trials, AI assists in selecting patients who are more likely to respond to a treatment by tracking genetic data and medical history as well as monitoring data during the trial in real time to provide instant feedback. This makes trials faster, more efficient, and less costly.

This technology is already being used by major pharmaceutical companies, biotech startups, research labs, and hospitals. This aids both patients and developers, by enhancing medicine delivery and cutting down on research costs with larger success rates respectively. AI is not just used for the discovery of drugs, but also for the repurposing of them: it uses medical data to simulate existing drugs and figure out any weaknesses with them or any other existing possibilities where they might be more effective.

Although AI is proving to be very useful in the drug discovery field, some major challenges still exist. For example, the data required to successfully complete drug discovery tasks must be legitimate with applicable research content which isn’t always widely available. Another issue is that AI’s predictions are never fully correct, leaving many of the generated possibilities unlikely to be tested.

Also, with a majority of the simulating and development of molecules based through AI, there is always the possibility of hallucinations, which would further reduce the possibilities of new drugs. Despite this, AI is highly beneficial in helping researchers find better medicines, faster, and more affordably. With some more time put towards the AI side of development, the future of drug discovery is looking more and more prospective.

Sources:

Exploring the structural changes driving protein function with BioEmu-1 - Microsoft Research

Varda Space Industries says “low Earth orbit is now open for business” - Marketplace

How AI will accelerate biomedical research and discovery - Microsoft Research

Drug Discovery & Development - Pharma and Biotech Insights

Varda Space to make drugs that are ‘impossible’ to produce on Earth

AlphaFold Protein Structure Database

Drug Discovery - Microsoft Research

This Lab Robot Mixes Chemicals

ChatGPT