Limited data isn’t an issue when it comes to how to best use computational tools to better prevent disease and treat patients—but how to more efficiently analyze existing information is.
Industry veterans with experience in life sciences and data science floated their ideas about how the pharmaceutical industry could better plug in to the “big data” ecosystem during a BIO-Europe Spring panel last month. The annual conference, put on by EBD Group (which is owned by the same parent company as Xconomy), was held virtually last month given coronavirus concerns.
Industry and governments looking to tap into the potential of data science tools for improving healthcare need to reconsider their “risk tolerance, appetite for collaboration, and the talent pool,” said Shelley Epstein, vice president of corporate and public affairs at Imagia, a healthcare AI startup in Montreal, Canada.
The healthcare industry has “valuable but dormant insights” in its data, and to extract them the information needs to be analyzed collectively, she said.
“Typically, this industry is very competitive, and rightly so,” she said. “However, if we are going to be able to extract insights and work together and really start sharing learnings across institutions, across partners, with governments, with academia, and so forth, I think there’s this level of trust that we’re going to have to start instilling in one another so that we can work together and have these meaningful outcomes.”
That collaborative attitude should extend to building relationships with other sectors, which could prove a rich source of diverse talent, new perspectives, and best practices, she added.
Guna Rajagopal, global head of computational sciences at Janssen’s Discovery Sciences, which focuses on small molecule drug discovery, said the organization agrees, and has identified academic and industry partners with which to work toward such outcomes.
His team within Janssen, the pharmaceutical arm of Johnson & Johnson (NYSE: JNJ), for example, has teamed up with UK Biobank as part of a larger consortium to get access to data—including whole genome sequences—from 500,000 participants who agreed to have their health followed.
“[That data] coupled with health data, provider data, and all the data they collected from the clinical samples that they got… technologies are available, such as AI and machine learning, to translate from data into insights that will guide our search for safe, effective therapies,” he said.
Yann Gaston-Mathé, co-founder and CEO of Iktos, a French startup using AI tools to design new drugs, said the company is cognizant of the fact that optimizing drug discovery is a huge task, with many facets to be tackled before patient outcomes are improved.
“We are focused on how to explore the chemical space, leveraging the data which has been generated in the past by our collaborators or which are accessible in the public domain, in order to come up with the optimal structures as fast as possible with as little trial and error as possible,” Gaston-Mathé said. “We hope to be able to make a difference in this very specific area of drug discovery and development, but we are very conscious that, you know, it’s much bigger than that.”
Collaboration is a way to bring together the skills, expertise, and resources needed to improve that big picture, he said.
“Most single institutions and industry players actually do not have enough data on their own to make these big discoveries,” Epstein agreed.
Especially in tackling seemingly insurmountable challenges, data science tools could make a significant difference, she said.
“Imagine the insights that could be derived from failed [Alzheimer’s disease] trials… I think this is we really have to start thinking outside of the box,” she said. “This is where you have industry sort of taking a step back because they are so risk-averse, for the most part, and don’t feel comfortable engaging in these new types of technologies, but this is where the wins are going to happen.”