- AI tools are creating a real opportunity for pharma but a real danger if you don’t have people with the skills to do the analysis.
- Data is an integral component to using AI in pharma. Recently the industry has seen a surge of data, which created a huge opportunity for the use of AI.
Like almost any field in the healthcare industry pharma wants to get in on the latest technology trends. Recently big pharma has been looking to artificial intelligence as another tool to help facilitate drug research and help the company progress.
At the World Medical Innovation Forum in Boston on April 24, a panel of pharma leaders discussed the future of AI in the industry.
“It’s quite top of mind for us at Novartis as we are reimagining life science companies like ours as medicines companies powered by data and digital and also because the radical advances, perhaps driven by consumer applications of this critical vector of computer science, have overt, immediate, and profound downstream relevance,” Dr. Jay Bradner, president of Novartis Institutes for Biomedical Research, said at Tuesday’s event.
That doesn’t mean companies are ready to replace their researchers with robots anytime soon. Instead Bradner explained these tools are expected to help scientists.
“I think it will be some time before we displace the discovery chemist. Rather, we imagine an augmented reality chemist, where empowered by a legacy of data, that a new chemist at Novartis wouldn’t immediately download or have access to, inferences hard won into molecular recognition, into the biological behaviors of molecules … enable fundamental decision-making during the artisanal process of drug discovery and lead optimization,” Brander said.
Data is an integral component to using AI in pharma. Recently the industry has seen a surge of data, which created a huge opportunity for the use of AI, according to Jean-Francois Formela, partner at Atlas Venture and moderator of the panel. However, there is a danger of putting bad data into the computer system or what Formela calls “garbage in and garbage out.”
Yet many pharma companies are investing in getting more and better data.
In fact in February, Swiss pharma giant Roche acquired oncology tech company Flatiron Health for just this reason. At the time Roche CEO Daniel O’Day said the New York company was “best positioned to provide the technology and data analytics infrastructure needed not only for Roche, but for oncology research and development efforts across the industry.”
The acquisition gave Roche access to more data in the field.
“Several years ago we recognized the value of real world evidence and that treatment patterns actually were very different than what we were running in our traditional clinical trials. We were seeking multiple partners at that point,” Lee Lehman-Becker, senior director in Roche’s digital and personalized health care partnering unit, said at the panel. “Something that Dr. Abernethy (Flatiron Health chief medical officer and chief scientific officer) spoke about yesterday, that we are very excited about, is real-world control arms. So actually having a pairing of our clinical trials of something that is happening within the Flatiron network and the ability to understand against true standard of care the way it is actually being practiced in medicine in a real setting versus breakthrough therapies.”
While Roche is turning to data and the skills of AI to help look at drugs in the field, Mark Murcko, chief scientific officer of Relay Therapeutics said he is interested in using the power of AI and data to look at research on the molecular level.
“I think one of the big challenges we have in being able to get better at drug discovery is to think through how to represent the interactions the drug molecules are making with their protein targets to understand that in a deeper way,” Murcko said. “That requires generating more data but also more diverse kinds of data and thinking very carefully about how all the data is put together just to tease out just a little bit of that insight. That helps a drug discovery team move a little faster. …If you could use a more diverse data stream and couple that with simulation and then apply AI, that will enable you to gain that insight which then helps that drug discovery team to crack at the problem.”
But the technology isn’t just used in research and development. In fact, at Johnson & Johnson it is being integrated across the company in everything from human relations to drug discoveries.
“I started my career in Bell Labs when we were developing models and theories for AI. But we didn’t have the data or computational power to prove it. But the time has come now I think we are at an inflection point,” Georgia Papathomas, global head of Data Sciences at Johnson & Johnson, said on the panel. “I am trying to set up data scientists for all of J&J from strategy to process and governance as well as trading data as an asset and apply it in every single function across J&J from finance and HR.”
However, Papathomas warns that in the future AI technology unleashed without knowledgable employees at the helm can cause issues. She has hired people with both health and computing backgrounds to remedy this.
“AI tools are creating a real opportunity for pharma but a real danger if you don’t have people with the skills to do the analysis.” Papathomas said.
Date: Apr 25, 2018