Niven Narain, CEO, BERG Heatlh
0:00 How does your platform compare to IBM’s Watson?
5:32 De novo discovery
10:23 Building patient maps before they take any drugs
18:41 Looking at 15 trillion data points on a single sample
23:03 What can we expect in the next two years?
The promise of rational drug design has driven pharma companies for years. The history of the industry has been one of trial and error, or “guess and check”, as scientists often say. Companies have screened thousands and thousands of compounds looking for one that might work—the proverbial needle in the haystack. With the arrival of molecular profiling and an explosion in understanding of basic biology, many in our field have hoped to make drug development more predictable: to start with the biology, and with the aid of new computing power, design a drug to work in a specific biological setting. But instead we've been making the haystack bigger. When will we see a tipping point where more drugs than not are derived through logic rather than luck? Though there have been great examples—Merck’s Truspot being the first and Novartis’ Gleevec perhaps the best known—rational drug design still remains a mostly unfulfilled promise.
BERG Heatlh, a Boston based pharma company, is working to change that with their AI platform and a new “back to biology” approach. Today we talk with BERG CEO, Niven Narain, who also is a co-founder of the company.
“We’re flipping the entire [drug development] model upside down,” Niven says. "We’re going to ask the patient biology what has gone wrong and generate as much of the omics data as we can from that patient. We then correlate that to their histories and health records in a population based way and compare it to healthy individuals. Out of that analysis we derive potential drug candidates and biomarkers.”
BERG has secured a considerable amount of press with the attention grabbing headline, “company using AI to cure cancer.” Backed by billionaire Carl Berg, the company can’t be faulted for thinking small. But are they doing anything different than other pharma companies? In today’s interview, Niven says they are looking at 15 trillion data points on a single sample. And the scientific approach is perhaps more robust that we’ve seen in the past. Rather than relying on the latest biomarker studies, the company takes a non-biased approach by letting the patient’s data create it’s own picture of the biology.