Warning: mkdir(): No such file or directory in /home/innovspa/public_html/gradvalley.in/wp-content/plugins/photo-gallery/framework/BWGOptions.php on line 355
Use of AI in Drug Discovery - Hype or Overblown?

Drugs that come out of AI-driven technologies will be history and revolution. Being positive and optimistic is the art of research.

Developing a new treatment is an expensive affair and it’s estimated that the industry and government agencies spend approximately US$2.6-billion. Most of the efforts go waste because it includes the 90% of the money that is spent on drug candidates that fail somewhere between phase I through III trials and the regulatory approval.

The industry is beginning to adapt to new technologies such as Artificial Intelligence (AI) to identify new drug candidates or repurposing the existing drug for new indications, thus making the drug discovery faster and cheaper.

“However, many in the field doubt the need to do things differently as they feel the use of AI in drug discovery is either hype or overblown.”

The survey from 330 scientists who work in drug discovery conducted by BenchSci in partnership with The Science Advisory Board in December 2017 shows that 40% of the scientists are unfamiliar with AI in drug discovery, despite the extensive media coverage for AI in the past years. The respondents shared that it’s increased speed of discovery that they perceive as a benefit (61%) followed by the increased comprehensiveness of research (46%), increased opportunities for existing compounds (45%), and increased novelty of targets and compounds (45%).

To explore a disease area, for example, let us consider Alzheimer’s disease, the most common form of dementia. Some experts believe this could bankrupt medicare and insurance companies if a therapy is not approved at least in a decade. According to the Alzheimer’s Association, $259 billion cost of Alzheimer’s care in the U.S. may reach an “unsupportable” $1 trillion annually by 2050, a scary scenario.

For many years, many drugs for Alzheimer’s Disease fail in development, while the cancer drugs are reaching a 20% success rate. There may be many factors such as inaccuracy, imprecision, bias, failures to follow or lack of operational protocols for applying Clinical Trial methods, inter-site variance, and lack of homogeneous sampling using disorder criteria, dementia experts have questioned the narrow focus from biotech companies on amyloid protein (Abeta) and some on tau. Although, the recent statistically significant results of Biogen drug that targets the amyloid protein shows a promise, many predict there is only a 50% chance for approval.

“How could this be solved? How the new targets and new drug candidates can be identified, if not curable but at least to slow down the progression of the disease?”

Machine Learning-based Virtual Screening (VS) is one way of identifying new drug leads. A recent study published on June 2018, in Current Pharmaceutical design, aimed to review ML-based methods used for VS and applications to Alzheimer’s disease (AD) drug discovery. The authors concluded that all techniques had found success in VS, but the future of VS is likely to lean more heavily toward the use of neural networks – and more specifically, Convolutional Neural Networks (CNN), which are a subset of ANN that utilises convolution. They additionally conceptualised a workflow for conducting ML-based VS for potential therapeutics for the AD.

The above is one of the approaches using AI to detect the right targets or drug candidates. However, as the technology is in the nascent stage, the predictions come out of AI, with overconfidence and expectations on such techniques to disrupt the drug discovery might fail and turn out to be a bubble.

“As there are no drugs that came out of AI technologies and may take few years to prove, all we can do is keep our fingers crossed and invest in such research, embrace success or failures.”

In the BenchSci survey, 71% responded to the survey that education about the technology would best help overcome the barriers to adopting new technologies, including AI. Whether it is self-driven learning or taking instructor-led courses on concepts, technology or the tools, the key takeaway from the survey is “while organisations are adopting the tech, there’s significant untapped potential for those willing to be more aggressive. However, the industry—and society—will only realise the potential for education and relevant success stories”.

“In spite of the controversies and debate, to keep up with the changing trends in the technology and be on the game, taking steps to become informed, updated and agile is unavoidable.”

If any of the drugs that come out of AI-driven technologies will be history and revolution. Being positive and optimistic is the art of experimental science. Right?

What’s your thought on this?

Leave a Reply