Palo Alto startup harnesses generative AI for drug discovery


Palo Alto startup harnesses generative AI for drug discovery

Drug discovery is a long and high-risk process, with the development and approval process for each new drug costing up to $2 billion on average. Biostate AI, a start-up based in Palo Alto and Houston, is seeking to revolutionize this process with the launch of a new data-sharing platform for researchers.

Founded by Ashwin Gopinath and David Zhang in Texas, Biostate AI aims to use artificial intelligence (AI) to predict how certain drugs will affect human health. The company is currently developing AI models to predict a living organism's health outcomes based on its DNA, RNA and protein biomarkers. The data-sharing platform, Omics Web, is the first step in this direction, according to CEO Zhang.

Omics Web, an online repository of biological data, is designed to be user-friendly by allowing users to filter data by various metadata tags such as organism, drug, organ and date.

"Omics Web is like a library with lots and lots of books in it, and we're constantly adding in more books [or] additional biological experiments that we do and data that we collect," Zhang said. "But inside there are also librarians, which happen to be AI, who will read all the books and tell you what the highlights are."

He said the "AI librarian" within the database is an important tool for researchers since it can otherwise be extremely hard to comprehend the complexities of the experimental data that Omics Web contains.

The company uses the Omics Web database to "smarten" their AI machines and models to predict human health outcomes, Zhang said. Researchers, professors and graduate students are highly encouraged to share their data with the platform and start new collaborations.

"I think the future of Omics Web is that we're basically building the great library of Alexandria, except for biological books," he added.

Omics Web contains data from biological experiments that Biostate AI has run themselves along with experiments that the company has run for other researchers. According to CTO Gopinath, one of Biostate AI's main goals is to make research accessible by cutting the cost of RNA sequencing.

For example, Biostate AI does much of its research through RNA sequencing and has found ways to reduce the cost of the process to $40. By keeping costs low, Gopinath said the company aims to extract more data for themselves and those who work with them.

Biostate AI "is cheaper than going with somebody else," Gopinath said. "If you go with us, you'll be able to do five to ten times more experiments, which means five to ten times more data, which means you have five to ten times more probability of getting more insights."

The data Biostate AI finds for others is then uploaded to Omics Web, if the researcher that they are working with allows it.

Despite the promises of Omics Web advertised by Biostate AI, some researchers -- such as pediatrics and computer science professor Gill Bejerano -- remain skeptical.

"When a researcher like me looks at a resource like that, we would say, 'Dude, I got 10, 15 resources like this vying for my attention,'" Bejerano said. "'Why should I even spend... a month or two trying to figure this out?'"

Bejerano said that sharing research data through platforms like Omics Web can be difficult, especially when similar government-owned or other company platforms exist. However, these alternative data-sharing platforms are not without their faults. While government platforms are more accessible and affordable, they take a long time to access and sometimes are unorganized.

According to Bejerano, Omics Web could potentially set itself apart from other available platforms based on its cost and its ability to effectively "curate" biological data for researchers to easily utilize.

The most promising aspect of tools like Omics Web is their ability to open up new possibilities for researchers, Bejerano said, comparing the platform to looking at menus at different restaurants.

"As a researcher, you try to read as many menus as you can, because the menus allow you to imagine which questions you could be asking," he said. "And then you go and say, 'You know what? If I go to that restaurant and I buy these items from them, I can ask that question that nobody's asked before me, maybe I'll make a cool discovery.'"

In the future, Biostate AI hopes to use information from Omics Web to help pharmaceutical companies eliminate drugs that the AI accurately predicts are unsafe, which, Gopinath said, can save billions of dollars and help drugs reach patients in a more rapid manner.

"If we have better healthcare [and are] providing the right drugs to the right people at the right time, we can get everybody up to near the 90 years that people can naturally, biologically live to," Zhang said.

Previous articleNext article

POPULAR CATEGORY

corporate

10189

tech

11464

entertainment

12517

research

5669

misc

13262

wellness

10079

athletics

13241