Drug Discovery Using Artificial Intelligence
Is It Really Possible?
What makes AI capable of drug discovery?
Artificial Intelligence has carved itself an important place in the field of healthcare. It has been used to identify widespread diseases, the effectiveness of healthcare schemes, detecting cancer and designing personalized treatments. However, imagining AI being used for the purpose of drug discovery still seems like a feature of a utopian society. This belief, though, is quickly vanishing as artificial intelligence promises to find effective drugs in less than a year.
A typical pharmaceutical company will take around 4 to 5 years to come up with the new drug. The drug discovery process for a pharmaceutical company is not only lengthy and deliberate but also costly and error-ridden.
Around nine out of ten pharmaceutical drugs fail to get FDA approval.
By discovering the correct drugs efficiently and in a small amount of time, AI can speed up the process and make a much larger number of potentially life-saving drugs available on the market. Thus, AI has become a sizzling topic in drug discovery.
Many pharmaceutical companies are now taking to AI for their R&D. Among them is Pfizer, which is using IBM Watson in searching for new immuno-oncology drugs. IBM Watson is an AI-based computer created by IBM, for answering queries presented in NLP. If Jarvis from Iron Man had a twin brother with “slightly” less computational abilities and a much more sober UI, that brother would be Watson. It has defeated human intelligence in some areas like winning the quiz show Jeopardy against its former champions. More importantly, it has successfully diagnosed a woman with leukemia. Many of the big pharma-corporations have now teamed up with similar programs.
Basic algorithms related to AI used in drug discovery
Think of drug discovery as trying to fit a key in a lock. The lock is analogous to the target ligand with which we hope for the drug molecule (the key) to bind to produce a reaction. The AI algorithm tests thousands and thousands of keys to find the one that fits perfectly. The only difference is that AI does this in seconds, not months.
There are two important aspects to drug finding — data mining and deep learning. Data mining uses the abundance of available research data, small molecule libraries on the 3D structure and binding properties of molecules to narrow down and identify the target ligand with which we want our drug to bind. Once the target has been identified, deep learning technologies, such as convolutional neural networks are used to create extremely accurate binding profiles. It then selects the best molecule considering a herculean amount of parameters, and voila! New drug found.
Network-driven drug discovery is another novel approach where we explore a potential drug’s ability to influence disease networks, rather than specific targets. This is done using large-scale, exclusive databases and special computational tools. These have shown effectiveness in at least 12 biological pathways.
Many startups and R&D centers are modifying and fine-tuning these algorithms as we speak, and it is not an unrealistic expectation to see AI covering strides in drug discovery within this decade itself.
Some examples of current research studies/trials where AI is actually being used for drug discovery
- AtomNet, a deep convolutional neural network by the company Atomwise, screens more than 100 million compounds each day. Coupled with transfer learning, AtomNet has a huge potential in helping startups to find essential drugs on a range of diseases.
- British based startup Exscientia made headlines recently for creating the first medicine using Machine Learning to be used on humans. The drug — known as DSP-1181 — will be used to treat Obsessive Compulsive Disorder (OCD). They are also actively working towards finding potential drugs to treat cancer and cardiovascular disease.
- Another company, Insilico Medicine has made a breakthrough in rapid drug development via its smart algorithm, in just 46 days. The potential drug is now being tested in mice.
Some interesting companies in this domain
Be it career opportunities or you are just curious to know what the future holds, these are some companies you can follow to stay up to date with new technology and products.
- Atomwise, which we talked about just now, has found evidence for potential drug candidates to treat Ebola using their cutting edge DL technologies. Such a discovery would have taken months or even years without AI.
- Antidote is a company that tries to match patients and researchers in clinical trials. Their focus is on facilitating better clinical trials for new drugs.
- Turbine.AI designs personalized cancer treatments for patients.
- Cambridge Cancer Genomics focuses on the precision oncology solutions to transform the way cancer patients are treated by using a data-driven approach.
- Deep Genomics is a startup that aims to create what they call “Genetic Medicines”. They have interesting ongoing projects like Project Saturn and the drug discovery platform.
Impact on the healthcare industry and future trends. What can we expect?
AI will surely transform every stage in the process of drug discovery and development. However, expecting it to be a complete replacement for certain processes is far fetched. For example, processes like chemical synthesis, clinical trials, and regulatory approvals.
The impact of AI will be seen in faster discoveries, reduce is research costs, risk prediction, and efficient clinical trials.
Here are some more resources to explore further-
Additional reading and references
- Fleming, N. (2018, May 30). How artificial intelligence is changing drug discovery. Retrieved from https://www.nature.com/articles/d41586-018-05267-x
- Agrawal, P. (2018). Artificial Intelligence in Drug Discovery and Development. Journal of Pharmacovigilance, 06(02). doi: 10.4172/2329–6887.1000e173
- Top Companies Using A.I. In Drug Discovery And Development. (2019, September 20). Retrieved from https://medicalfuturist.com/top-companies-using-a-i-in-drug-discovery-and-development/
- Smith, S. (n.d.). 230 Startups Using Artificial Intelligence in Drug Discovery. Retrieved from https://blog.benchsci.com/startups-using-artificial-intelligence-in-drug-discovery#design_drugs
Editorial Note-
This article was conceptualized by Aditya Vivek Thota and written by Arshika Lalan of The Research Nest.
Follow our publication for diverse research trends and insights from across the world in science and technology, with a prime focus on artificial intelligence!
