Posted on Dec 31, 2019, 1 p.m.
In recent years artificial intelligence has been receiving much attention, while there has been some disappointments in the case of drug discovery AI technology has had significant effects and we are on the cusp of AI developed drugs being a reality.
The Human Genome Project was hoped to usher in a new generation of precision medicine, but this proved to be more of a complex and nuanced challenge than was anticipated; of the close to 25,000 human genes only 2,418 have been associated with specific diseases which only explains a small portion of human pathologies.
In the coming year(s) we may begin to utilize the abilities of artificial intelligence to develop/create new medicines. In recent years there have been breakthroughs that helped us to gain better understandings about the complexity of diseases such as biology arising not only from DNA but also from metabolites, proteins, different cell types and the interactions between them. When something goes wrong in the complex biological system of reactions disease occurs, and if we were able to understand them all a new approach may be needed to investigate the system as a whole, and we may need AI to accomplish the task.
The volume of data arising from gene mapping experiments is continuously increasing and becoming intractable, as Hal Barron of GSK warned, “It's definitely overwhelming for any human to think about how to deal with this.” Recent studies have shown that machine deep learning is emerging as an accurate approach to find patterns in this data.
GSK has partnered with Exscientia which is working to develop artificial intelligence deep learning algorithms that can design new molecules based on pharmacological data; and early in 2019 they revealed the discovery of a new molecule that may be able to be used to treat chronic obstructive pulmonary disease.
In the industry, other big pharma giants are also racing to be the first to discover new drugs by collaborating with artificial intelligence biotechnology startups. Merck has partnered with the startup up Atomwise which applies convolutional neural networks to model chemistry interactions using software to simulate more than 10 million compounds a day to analyze molecular reactions to predict how they may act within the body.
The startup OccamzRazor is using natural language processing to combat Parkinson’s disease, their algorithm developed in partnership with Stanford’s AI Research Laboratory reads scientific reports and data to extract information and continuously map the biological network of Parkinson’s. Recently they mapped all of the information science has learned about the disease, and scientists are beginning to identify cellular dysfunctions that lead to it. In the near future they hope to be partnering with pharmaceutical companies to develop treatments.
In this coming year it is possible that we may see the first clinical evidence demonstrating the efficacy of drugs that have been developed by artificial intelligence, this may include those such as BPM31510 which was discovered by the startup Berg for treatment of pancreatic tumours and carcinomas; or BenevolentAI’s trial for a drug to treat sleepiness in Parkinson’s disease patients. Drugs such as these may be the first of many to come that have been discovered and designed by artificial deep learning machine intelligence.
Just as what happened with The Human Genome Project there will be much attention around artificial intelligence with some disappointments, but when it comes to drug discovery this technology is already beginning to show that it can have significant effect, and over the oncoming years AI may prove to be important to medical breakthroughs.
Materials provided by:
Note: Content may be edited for style and length.
This article is not intended to provide medical diagnosis, advice, treatment, or endorsement.