Cancer drugs can take over a decade before they are available in the market. This situation is owed to the lengthy processes involved in making the drugs including human trials and animal testing. A regulatory review is also conducted through which less than 7 percent of the proposed medicines are approved. Therefore, 33 percent of four out of 10 American people have less time to access effective treatment, as estimated by the National Cancer Institute.
In a study published in the medical journal Oncotarget, a team spearheaded by Insilico Medicine aims at reducing the time taken in research using artificial intelligence (AI). Insilico Medicine is a biotech research company based in Baltimore. The researchers came up with two computer networks known as generative adversarial networks. One network proposes new molecules that contain cancer-fighting properties while the other disregards the suggestions by known treatments.
Once the networks are done testing each other, they can be utilized to evaluate compounds in terms of their cancer-fighting abilities. Similarly, a team from Insilico Medicine vetted 72 million chemicals sourced from a public database. The networks correctly identified the drugs by selecting 60 patented cancer medicines.
Researchers can narrow down the number of new cancer treatments in a month to just 100 most potential leads. Mainly, this is because the in silico (computer-tested) technique is faster compared to the in vitro experimentation. For failed drugs, an extra 1.6 billion is added to the cost of every successful medicine. Hence, a reduced number of potential leads means millions or even billions would be saved in terms of labor and resources. This approach is not only cost-effective but also promotes faster development of cancer treatment drugs.
However, the application of in silico faces critics from a few individuals like Mamoshina who asserts that numerous cancer researchers are unaccustomed with artificial intelligence, which can yield doubt. Olexandr Isayev, an assistant professor at UNC, also ascertains the existence of a wave of excitement regarding the cutting-edge technology that is yet to produce material results. Additionally, he pointed out that the predictions could be wrong since they are computational.
Instead of licensing the technology in a software-as-a-service model, Insilico Medicine is intensifying the research into those molecules identified by the networks to contain cancer-fighting potential. After the compounds are passed through the traditional in vitro testing, they will then be licensed to pharmaceutical companies for additional regulatory review and marketing if everything works out. Recently, sources announced that Insilico Medicine has partnered with pharmaceutical giant GlaxoSmithKline in a bid to apply the new research techniques.
Judging by its decision to license the discovered drugs as opposed to the tools of discovery, Insilico Medicine’s firm belief in the new research method is evident. However, for the company to clear any doubt that AI can eradicate the guesswork associated with early drug discovery; it requires providing more proof from the lab.
Original source Newsweek