The use of Artificial Intelligence (AI) and Machine Learning (ML) in the field of drug discovery has exploded in recent years, and earlier this week a group of researchers reported using a proprietary deep generative machine learning model to create a lead drug candidate against a therapeutic target found to play a role in fibrosis and other diseases.
Using this ML model, called Generative Tensorial Reinforcement Learning (GENTRL), the researchers analyzed a vast library of compounds for blocking the therapeutic target, DDR1 kinase, and narrowed down their search to just four compounds which were found to be active against the target in biochemical assays. Two compounds were also validated against the target in cell-based assays. The researchers narrowed it down to one drug candidate and tested it in a preclinical mouse model which showed encouraging pharmacokinetic results.
Notably, the whole process from starting the drug candidate search to finally finding one that showed positive activity against the target in preclinical studies was done in just 21 days. The news made headlines all over the world, since discovering therapeutic targets against the desired target is traditionally a time and labor-intensive process. It involves trial and error testing in the lab, which can take many years before a suitable drug candidate is discovered. While this research is not the first effort to use AI/ML in drug discovery, the shortened 21-day time period from start to finish was remarkable enough to get it media attention globally.
Image showing how a GANS (Generative Adversarial Network) functions.
While most large-cap biotech companies like AbbVie, Celgene, Bristol-Myers Squibb, AstraZeneca, GSK, Janssen, etc. are all using AI/ML in developing new drugs, some examples of developmental stage small-cap biotech companies using AI/ML in drug target discovery are given below:
Gritstone Oncology (NASDAQ: GRTS): The company is using a proprietary EDGE(™) platform using AI/ML to accurately identify the mutations that form true cancer neoantigens. EDGE showed 9x higher predictive value for neoantigens and 20x higher predictive value for HLA Class II predictions than publicly available algorithms like NetMHC. The initial therapeutic targets are cancers with a high frequency of KRAS mutations like lung, pancreatic, colorectal, etc.
Neon Therapeutics (NASDAQ: NTGN): It is using a proprietary RECON Bioinformatics Engine which is powered by deep learning neural networks and trained on large, proprietary, high-quality data-sets. RECON processes the tumor biopsy data from a cancer patient and predicts the most therapeutically relevant cancer neoantigen targets. The initial therapeutic targets are melanoma, non-small cell lung cancer, and bladder cancer.
Wave Life Sciences (NASDAQ: WVE): Wave has a collaboration with Deep Genomics, Inc. to use its machine-learning biomedical platform to develop novel therapies against genetic neuromuscular disorders. Wave’s first target is developing oligonucleotide-based therapies against Duchenne Muscular Dystrophy (DMD).
The most notable company in the private sector that has made significant advances in using AI/ML in drug discovery is InSilico Medicine. The company has collaborations with over 150 academic and industry collaborators worldwide. They also published a research paper using a deep neural network Adversarial Threshold Neural Computer (ATNC) to produce drug candidates successfully.
InSilico Medicine’s AI Pharma Pipeline
This article represents my own opinion and is not a substitute for professional investment advice. It does not represent a solicitation to buy or sell any security. Investors should do their own research and consult their financial advisor before making any investment. Investing in equities, especially biotech stocks has the risk of significant losses and may not be suitable for all investors. While the sources of information and data in this article have been checked, their accuracy cannot be completely guaranteed.
I/we hold a long position in GRTS, WVE.