In order to discover new drugs, Insitro starts with biology. Using cells from humans, the company induces them into pluripotent stem cells which can turn into almost any cell in the body. Researchers then create models of genetic diseases, and use machine learning to figure out what the difference is between healthy cells and sick cells. These subtle differences are “often lost on a person because of the sheer volume of data,” Koller says, “and that’s where machine learning really shines.
Insitro hasn’t discovered any new medications yet, but last year it inked a deal with Gilead to work on a drug for liver disease with $15 million up front and the potential for $1 billion down the road. The deal allows Insitro to use data from Gliead’s clinical trials to train its machine learning platform. The company also focuses heavily on neuroscience, and hopes that using human cells instead of animal models will help find new drugs for neurological diseases.
Some of the new capital from the funding round will go towards building up the company’s capabilities to develop drugs, including eventually hiring regulatory experts and other staff that have experience with drug development. Seeing as how the company hasn’t yet found its first drug, however, those particular hires are still a long ways off. Money will also go towards scaling up the company in other ways, like continuing to develop its liver disease research.
The field of using machine learning to aid drug development is becoming increasingly crowded. Other startups including Recursion Pharma and Verge Genomics are also using machine learning to speed up drug development, and major pharmaceutical companies like Novartis and Merck have partnered with companies to improve machine learning capabilities.
One thing that makes Insitro different from the rest is that it produces massive amounts of new data to train its machine learning platform. Many other companies have taken the shortcut of relying on existing datasets to teach their machine learning, Koller says, but some of those data sets are messy and will result in poorly functioning platforms. “My experience is that machine learning is really only as good as the data you feed it,” she says.
Despite Inistro’s promising avenues as a project, most are faithful because of the company’s founder: Daphne Koller.
United States Latest News, United States Headlines
Similar News:You can also read news stories similar to this one that we have collected from other news sources.
Source: BusinessInsider - 🏆 729. / 51 Read more »
Source: ELLE Magazine (US) - 🏆 472. / 51 Read more »
Source: Forbes - 🏆 394. / 53 Read more »
Source: Women 2.0 - 🏆 149. / 63 Read more »
Source: people - 🏆 712. / 51 Read more »
Source: CNBC - 🏆 12. / 72 Read more »