Editor's Note: This is the second in a new series of "In conversation with" long-form articles we’ll be producing on a regular basis, talking to biotech CEOs and executives about deals, cutting-edge science and biotech culture, as well as their latest updates, to build up a microcosm of the life science industry as seen from those running it. Want to learn more? Contact our editor Ben Adams.
When Alice Zhang walked away from a Ph.D. program at neurogeneticist Daniel Geschwind’s lab at the University of California, Los Angeles, to cofound Verge Genomics in 2015, she had a big vision—to create a scalable drug discovery engine that uses machine learning to rapidly develop transformative and personalized treatments across disease areas that were thought to be too complex to solve.
Where better to start than neurodegenerative diseases, where biopharma companies have struggled to make enough meaningful progress to translate into new medications, and which falls into Zhang’s expertise. Fast forward to this year, Verge has already formed two collaborations with top academic institutions in amyotrophic lateral sclerosis (ALS) and Parkinson’s, respectively, and has signed deals with two large, unidentified pharma companies in bipolar disorder and cerebral amyloid angiopathy.
“Diseases are caused by complex interactions between many genes. Most drug discoveries fail because researchers target only one gene at a time,” Zhang explained to FierceBiotech the rationale behind her company. The one-target path is indeed what’s happening with some neurodegenerative disease research. Northwestern scientists single out new mutation linked to Parkinson’s disease, New target could yield treatment for Alzheimer's, ALS and more, just digging up a few headlines from our story database.
Seeing the current or traditional drug discovery model as “a guessing game,” Zhang argued that though it worked for prior diseases, complex diseases like Alzheimer’s are simply too complex to solve by serendipity.
In comparison, Verge’s approach is to map out the hundreds of genes that are co-regulated to cause a disease, based on loads of data from DNA, RNA and protein, and then find drugs that target all the genes at once. Sounds ambitious? Verge has an explosion of genomic data, artificial intelligence and the same breakthroughs that power Google’s search engines to lean on.
To Zhang, a member of the 2017 cohort of Forbes “30 under 30,” these tools have enabled better understanding of the brain and generation increased quantity, diversity, and higher resolution data sets not possible before. And those computational predictions and enormous amount of data can in return train Verge’s model and improve its algorithms—which is the whole idea of machine learning.
“We then validate the computational predictions in our in-house human-based preclinical models in an ‘all-in-human’ drug screening platform,” Zhang said. This means, instead of relying on animal data until clinical trials, Verge uses human data from Day One.
Verge has put together a team with some of the world’s top talent in machine learning, mathematics, and neuroscientists and pharma veterans. This combined team makes it possible to “rapidly advance drugs from computer to bench to clinic,” Zhang said. Verge’s lead program, in ALS, went through data collection, target discovery and a small-molecule hit in a year.
In the last year, Verge has generated what Zhang said is the world’s largest proprietary database of patient genomic data in ALS and Parkinson’s Disease.
“At Verge, we believe that breaking down the silos that exist between industry, academia, computation, and biology is crucial to fully realizing the potential of AI in drug discovery,” Zhang said.
In September, her company announced an ALS partnership with the Motor Neuron Center at Columbia University, Massachusetts General Institute for Neurodegenerative Disease, the Keck School of Medicine of USC and the Department of Neurology at the University of Michigan Medical School. The collaboration aims to pair Verge’s patient genomic database with patient-derived induced pluripotent stem cell (iPSC) models and some state-of-the-art imaging technology to bridge the gap between findings from patient disease progression and preclinical models.
Then in November, Verge teamed up with the NIH’s National Human Genome Research Institute, the Scripps Research Institute and the Dresden University of Technology to focus on Parkinson’s. These academic institutions provide samples drawn from a large collection of Parkinson’s patients and preclinical model, while Verge coordinates the genomic profiling, analysis and data generation to determine the best preclinical models for progression into the clinic and identify novel therapeutic targets, Zhang told FierceBiotech.
Other than a $4 million seed round completed two years ago, the company is not disclosing any other funds. But Zhang said the San Francisco-based company has built a portfolio that includes six different diseases with over a dozen novel opportunities in the pipeline.