DAVOS, Switzerland — Developing new medications is a notoriously expensive, slow, and failure-ridden process.
Using artificial intelligence could change all that.
AI is the “next great tool” in drug development, Novartis executive Jay Bradner told Business Insider, and the $220 billion Swiss drug giant is “extremely organized” around deploying it. As president of the Novartis Institutes for BioMedical Research, or NIBR, Bradner leads the company’s drug research and discovery.
But, speaking at the World Economic Forum in Davos, Switzerland, last week, Bradner also acknowledged that the technology isn’t quite there yet, noting that “examples of AI-based drugs, at present, are few and far between.”
“Amazon knows just what ad to put at the bottom of the page when I order my squash racket. But the organic chemist in the fume hood at NIBR doesn’t have access to a massive repository of insight when choosing the next site on the molecule to fluorinate,” Bradner said, referring to a common process intended to influence a drug’s properties.
Using data science to hunt for new treatments
Data science is already an entrenched part of Novartis’s drug-development work. NIBR has invested in the space, employing about 400 data scientists alongside 6,000 drug hunters.
It’s so key to the process that a drug-hunting team typically consists of a chemist, biologist, clinician, and data scientist, Bradner said.
Of pharma companies, Novartis is most explicit about its shift to being a data company, Peter Lee, the corporate vice president of Microsoft Healthcare, told Business Insider in another conversation at Davos.
But all drugmakers are now thinking that way, Lee said, and trying to use data to improve everything from drug development and clinical trials to pricing and selling medicines.
Novartis is placing bets especially on the promise of AI. It employs a dozen or so in-house machine-learning experts, Bradner said, but also works strategically with external experts in that area, “of which there are many, many more.”
A twist on a strategy once used for Viagra
The company has identified about 12 places where AI could make drug development faster and better, according to Bradner. AI could help researchers go from a lack of clarity on a disease to having a target for medicines, he said, and transition from the starting point of a computer simulation to a promising chemical compound.
Right now, though, finding new uses for drugs is where AI can be most helpful, Bradner said. That’s called “drug repositioning,”
The strategy has long been used by pharmaceutical companies. It even created the profitable erectile-dysfunction medication Viagra, which was originally developed for chest pain.
Technology, though, has allowed for more firepower. AI can process large amounts of data from clinical trials to see if, say, a drug for heart health could also benefit patients with a rheumatic disease — “drug repositioning on steroids,” as Bradner put it.
Novartis has used this approach both to open and close doors while developing drugs, he said.
Most exciting to Bradner, though, is an even more specific prospect: Using AI to crunch huge amounts of data, including DNA, data from images, and measurements in clinics, to make connections between potential new drugs and subpopulations of patients with a disease.
Knowing exactly what kinds of patients drugs could work best in could make the costly process of drug development much more productive, and perhaps faster, helping them get to patients quicker but also giving companies like Novartis a leg up over their competitors.
from Trendy News Today http://bit.ly/2G82n0o
via IFTTT
0 Comments