The AI Promise: Examining the Potential Impact of Artificial Intelligence in Life Science

 

Artificial Intelligence (AI) is taking off in a variety of industries, with nations taking steps to stimulate greater utilization of AI, such as the large investments into AI by China and Europe. The situation is similar in the US, with the White House making AI funding a priority, and the US FDA approving the very first AI-based medical device for diabetic retinopathy detection in April. With the vast promise that AI holds to optimize workflows, numerous organizations worldwide are developing methods to further advance the technology, including in the life sciences.

One organization working to educate the life science industry on AI and help drive its adoption is the Pistoia Alliance, a global, nonprofit coalition of life science companies, vendors, publishers and academic research groups that focus on accelerating research and innovation in the industry. Companies represented on the Board of Directors include DNANexus, Dotmatics, Elsevier, Merck, PerkinElmer and Roche.

IBO had a chance to speak with Dr. Nick Lynch, consultant at the Pistoia Alliance, about adoption, applications and the future of AI, as well as efforts undertaken by the Pistoia Alliance’s Centre of Excellence for AI in Life Sciences, which fosters collaborations between stakeholders to drive AI utilization in the life science industry.

 

Surveying the Life Science AI Landscape

Earlier this month, the Pistoia Alliance released the results of a survey of a group of professionals in the life sciences to assess the adoption of and attitudes towards AI, finding that 72% of respondents believed the life science sector is “lagging behind” compared to other industries in regards to AI adoption.

The survey examined responses from 229 life science professionals regarding the use of AI. According to Dr. Lynch, the definition of AI in the survey was covered the many areas for which AI can be used. “The Pistoia Alliance has adopted a broad definition to include other processes that fall under the umbrella of AI, such as machine learning, deep learning, neural nets and chatbots,” he explained. “We decided on this scope to cater for the significant amount of interest in all areas of AI from our members.” He also clarified that, as per the survey, a life science professional was defined as “a business owner, or employee of a company, vendor, publisher, academia or government body that contributes in some form to life sciences discovery.”

 

“AI is only as good as the quality of the data used to inform and build an algorithm.”

 

The survey found that almost 70% of respondents utilize AI in some way, indicating a high adoption rate. The survey results are complementary to research conducted by the Pistoia Alliance last fall. “As of September 2017, the majority (46%) of AI projects were currently taking place in early discovery or preclinical research phases,” said Dr. Lynch. “Natural Language Processing (NLP) is also employed by just under a third (30%) during early phase research.” AI usage in development and clinical applications represented 15%, while imaging analysis accounted for 8%, Dr. Lynch expounded. “When it came to machine learning, more than a fifth (23%) of respondents were using it for target prediction and repositioning, followed by biomarker discovery (13%) and patient stratification (5%).”

However, this data from the Pistoia Alliance’s research indicated that although AI adoption was substantial, it was not universal, as Dr. Lynch noted. “[A] notable number of respondents [were] not yet using AI (11%), NLP (27%), or ML (30%) at all,” he said. “[A] further 8% of respondents [admitted] they knew ‘next to nothing’ about AI and deep learning, highlighting the need for greater education and knowledge sharing.”

 

Hype or Helpful?

The benefits of AI have been touted as being transformative to research, and, as Dr. Lynch stated, AI can have an advantageous impact on safety in the lab and general operations. “Take, for instance, a pair of smart glasses—augmented-reality lenses that a chemist or scientist can wear in the lab,”explained Dr. Lynch. “Perhaps the chemist is about to perform an experiment for which she or he is inappropriately trained or at risk of incorrectly handling a hazardous substance. Maybe the laboratory isn’t set up for the job; maybe it lacks a category 3 biosafety rating required for examining a biological sample. In these instances, the glasses can alert the scientist and prevent an accident.”

 

“If the data isn’t standardized, it means it is impossible to collect, aggregate, harmonize and analyze data on which to build AI algorithms in the first place.”

 

The potential advantages of AI are endless, but ultimately, research is a discipline that requires a human element that AI will not be able to replace, as Dr. Lynch noted. “It has been suggested that algorithms are capable of replacing a human scientist, but this is not a realistic near-term proposition, particularly when we look at the complexity of the problems addressed in the field of science and discovery,” he said. “Curing cancer, finding effective treatments for Alzheimer’s disease, seeking new methods of carbon capture and pioneering alternatives to lithium ion batteries involve research efforts that differ fundamentally from the kinds of tasks with discrete sets of outcomes or solutions that are ripe for automation. They require interdisciplinary expertise and judgment that a machine cannot provide.”

That notion of judgment is an important one, considering that for all the promise AI holds, the technology will depend on the quality of data that is input in it. Due to a lack of cohesive data standards, this poses a significant challenge. “AI is only as good as the quality of the data used to inform and build an algorithm—this means that if the data used is inaccurate or not varied enough, so will the outcome be,”said Dr. Lynch. “This is one of the main obstacles we’re trying to overcome at the Center of Excellence, as the industry is regularly hampered by a lack of data standards. If the data isn’t standardized, it means it is impossible to collect, aggregate, harmonize and analyze data on which to build AI algorithms in the first place.”

Other projects at the Pistoia Alliance’s Center of Excellence are also geared towards promoting and organizing the facets of AI in the life sciences. “The Pistoia Alliance is currently working to establish the steps needed to turn the widespread adoption of AI into a reality,” stated Dr. Lynch. “We want to see how the greater community can help the industry overcome many of the obstacles it is currently facing.” To accomplish this, Dr. Lynch noted that the Center of Excellence organizes events, webinars and hackathons covering numerous challenges with the purpose of initializing AI projects or prototype AIs. “This will help us to work out where the main challenges are, and to establish an adoption strategy that serves these needs,” he explained. “The Pistoia Alliance’s community also provides a virtual and physical co-working space, enabling life science professionals to share best practice and learn how best to apply AI, machine and deep learning to R&D—wherever they are in the world.”