Artificial intelligence, with its grandiose claims and sweeping promises to revolutionize drug discovery, may seem omnipresent in biopharma now. But Virginia Burger and RA Capital believe there are niches it has yet to touch.
After taking up residency at the star-studded Boston accelerator Petri, Burger’s startup — named New Equilibrium Biosciences — has scored $10 million in seed cash from the VC firm to prove that by reaching into those corners, they could uncover drugs against “Holy Grail” targets in everything from cancer to neurodegenerative diseases.
Specifically, the biotech is interested in a class of proteins called intrinsically disordered proteins that doesn’t have a single folded structure. MYC, for instance, is a cancer-causing IDP.
“They break this paradigm of sequence to structure to function because it’s really sequence to multiple of different conformations which have many different functions,” she told Endpoints News.
They can also range from fully unstructured to partially structured ones containing a disordered region, something condensate players like Dewpoint are focused on.
Cracking this special class, RA Capital principal Nathaniel Brooks Horwitz said, required a new breed of company that integrates both AI and physics in drug discovery, beyond what the singularly focused pioneers — Schrödinger, Relay and Silicon on the physics side; Atomwise, Recursion and Exscientia on the AI side — are doing.
“What we’re really excited about is when a company like New Equilibrium can enable a target about which we can say the following — ‘If only we could develop a drug for this target’ — we’re confident it will be successful,” he said.
As a PhD in Pittsburgh and later postdoc at MIT, Burger would read papers highlighting their importance as drug targets. But they were so hard to target that until midway through her postdoc, scientists weren’t even sure if ligands could bind to them — they wiggle around too quickly to be even seen in experiments. Even when they did find molecules that bound weakly to the proteins, there was no real way to optimize them into drug candidates.
Immersed in the entrepreneurial environment at MIT, it seemed natural to build a startup around the computational methods she had been developing to identify the set of conformations the proteins would switch between. First, though, she took a job at XtalPi, the US-China AI biotech startup backed by Tencent and later SoftBank and Morningside.
It was during that two-year stint that she met Peter Tompa, a professor at Flanders Institute for Biotechnology who’s devoted his career to studying IDP structure and function. He was interested in starting a company together.
Two weeks later, she was out.
The first thing she did after receiving funding from Petri — an accelerator set up to tailor to the needs of young founders — was to buy quantum chemical software and start generating their own training data for a new kind of physics model.
Existing computational models, Burger said, were built on what’s already known about folded structures, which renders them the wrong fit for intrinsically disordered proteins. By using AI to learn quantum chemistry — the energies and forces on each atom that give rise to a shape at any given moment — New Equilibrium’s algorithms can “see,” in silico, thousands or even millions of possible structures over time.
“We’re redoing how simulations have been done in the past by rebuilding the underlying simulation architecture to use AI for each step instead of introducing a single — the standard calculation for each step,” she said.
The next step is to zero in on the structures that are more stable or appear more frequently, screen ligands against them in a wet lab, and then fuse the fragments binding to different structures in the same molecule. Burger can’t yet reveal how many fragments they’d need for each drug made this way, but noted that the seed funding will get them closer to preclinical candidates.
For RA Capital, the deal marks another move in their shift toward earlier stages of venture financing.
Quantum computing, Horwitz noted, is looming on the horizon with hardware starting to take shape. And the technology will ultimately make it possible to “fully model the true biophysical state of even the most complex molecular interactions.”
“Which will be the companies that are the first to use quantum computing and all the power that brings to rationally design medicines for meaningful targets?” he said. “I think New Equilibrium can be that company.”
This is a syndicated post. Read the original post at Source link .