all the information, none of the junk | biotech • healthcare • life sciences

Verge Genomics Nabs $32M For Computation-Heavy Neuro Drug Discovery

Xconomy San Francisco — 

Two years after making her first-ever rounds of business meetings at the J.P. Morgan healthcare conference, Alice Zhang has steered her biotech startup, Verge Genomics, into a $32 million Series A financing.

Zhang (pictured) was notable then and now for her age—few biotech CEOs are under 30—and for her career path. She walked away from academia, potentially months away from completing a PhD, to start San Francisco-based Verge in 2015, and then became one of the first biotechs to graduate from the Silicon Valley tech incubator Y Combinator.

In a 2016 profile, Zhang told Xconomy that in jumping from academia to a startup, “there are a lot more unexpected things. You have to go with the flow. Learn on the fly.”

In three years, Verge could be ready to test a drug in clinical trials, Zhang says now. The company’s plan to discover and ultimately develop new drugs for neurological diseases is based on a blend of heavy computation and cutting-edge biology.

Verge collects brain samples from people who have died from neurological diseases; its first program is in amyotrophic lateral sclerosis, also known as Lou Gehrig’s disease.

Verge measures the genomic activity of those neurons alone in a dish—that is, removed from the complex brain environment. It then compares the patient neurons to neurons from healthy people of similar age and uses machine-learning algorithms to predict which genes are at the center of disease-related activity, and which potential drugs might target those genes. Dozens of genes could be switched on or off in concert in a particular disease. Verge also uses skin cells from patients to test the drugs flagged by its algorithms and identify which drugs show therapeutic promise. To do so, the company first transforms those cells into neurons using induced pluripotent stem cell technology, a method of rewinding cells back to their original stem-cell state, then pushing them down a new biological path. Its inventors won the Nobel Prize in 2012.

But cells in a dish and intense computation can only paint a limited picture of what’s happening with the complex biology of neurological disease. “A cell will never be a brain,” says Michael Sasner of the Jackson Laboratory in Bar Harbor, ME, whose expertise is mice genetically engineered to mimic human neurological diseases. “You always need in vivo [mouse] models.”

So Verge also uses mouse models— mice genetically engineered to have human disease characteristics—to test its predictions. For ALS, Verge wants to see its drugs reverse muscle dysfunction, a decline in cognition, and other symptoms.

“We think that selecting drugs this way”—that is, combining data from cells gathered from patients with mouse-model experiments—“helps de-risk some the translational barriers that have plagued neuroscience development for so long,” says Zhang.

Drugs have been tested in mice for decades. What’s new in the field is the ability to capture reams of genomic and other “omic” data from cellular processes and the proteins they’re churning out, then train software to see patterns in those data that humans would never pick out. Verge isn’t the only startup combining new biological and genomic tools with machine learning. But its focus on difficult neurological diseases makes it more unusual. Investors in the new round, many of whom were Verge’s seed funders, are obviously willing to make a big bet. DFJ (formerly known as Draper Fisher Jurvetson) was the lead, with WuXi AppTec’s corporate venture fund, ALS Investment Fund, and others taking part.

With the new cash, Verge wants to start programs in Parkinson’s disease and two others Zhang declines to name. After using its $4 million seed funding to build its biocomputational platform and open in-house stem-cell labs, the company will now focus on finding medicinal chemists—the folks who hone chemical molecules into pharmaceuticals. That’s even farther away from Zhang’s original expertise, which was writing the original machine-learning algorithms at Verge’s core. “My cofounder and I wrote a lot of code at first,” she says. “But I don’t code anymore, except to de-stress.”