By Sheena Goel
Drug Discovery is a time and money consuming complex process. Bringing a new drug to market takes up to $4 Billion and 10-15 years.
There are 7 phases in a drug discovery:
1. Finding the cause of disease: It takes upto 2+ years to find target responsible for the
disease, These targets often consist of DNA mutations, misfolded proteins, and other
potential disease biomarkers. There are often so many targets that human beings find it very difficult to figure out each of them.
2. Discovering potential inhibitors: This phase also takes somewhere between 1-2+
years. During this phase all the compounds that can interfere with the targets are studied.
3. Medicinal Chemistry Testing: The inhibitor compounds are further tested in this phase, based on analysis of the testing result, compounds are further optimized for the targets. This phase also takes about 1-2+ years.
4. Pre-Clinical In-Vitro Studies: During this study, compounds are tested in the cell
system and again takes about 1-2+ years of time.
5. In-Vitro study, in the animals: In this stage, the compound is tested in the animals. But it is quite expensive and takes upto 1-2+ years of time. The drug failure rate is also very high in this stage.
6. Clinical Trials on humans: Taking upto 6+ years of time, clinical trials are conducted
on humans during this stage. Clinical trials are conducted to prove the efficacy and
safety of the drugs on human beings
7. FDA approvals and commercialization: Once the safety and efficacy of the drugs is
proven, It takes about 1 year of time to get necessary FDA approvals and the
production of drugs.
However as we noticed, the drug discovery and production lifecycle takes a very long time and is very expensive. As a result, once the drug comes out for public use it is very highly priced, making it unaffordable for the general public.
However, AI and ML has changed this scenario to a large extent, In many of the drug
discoveries it does not takes more than few weeks to get past through all these stages and launch a safe drug in the market. Thereby reducing the overall cost of drug discovery. Thus, ML can accelerate discovery of safe medicine by manifolds. It is possible to simulate the testing environment for these drugs, and then find out the impact of a newly discovered inhibitor on the targets as well as conducting simulated trials of these compounds on the human beings. Result of human trials can be measured accurately using the ML. This reduces the possibility of side-effects of drugs on human beings.
Animal trials can also be reduced by the use of ML. VAE (variational auto encoders) – a generative neural network technique can be used to generate molecules for drug discovery. Since, ML in biotechnology is still in nascent stage, not much break through has happened but it has huge potential to help find out much revolutionary treatments for various ailments at much lower cost, faster speed and with greater safety.
Refrences:
1. https://www.tessella.com/insights/how-is-machine-learning-accelerating-drug-
development#:~:text=%E2%80%9CBig%20data%20and%20machine%20learning,drug%20
effectiveness%2C%22%20she%20explains.
2. https://towardsdatascience.com/unlocking-drug-discovery-through-machine-learning-
part-1-8b2a64333e07
3. https://www.nature.com/articles/d41586-018-05267-x
4. https://data-flair.training/blogs/machine-learning-in-healthcare/
Comments