Modern drug discovery has become increasingly adept at identifying promising therapeutic targets. Advances in genomics, proteomics, and AI now allow researchers to sift through vast biomedical datasets to pinpoint proteins and pathways most likely to influence disease. Yet despite these advances, most drug candidates still fail in the clinic, often because the underlying biology proves more complex than anticipated.
Soley Therapeutics believes the problem may lie in where drug discovery begins. Rather than searching for drugs against predetermined targets, Soley first screens compounds for the cellular responses they provoke, then works backwards to determine their mechanisms of action, therapeutic indications, and the patients most likely to benefit.
Our whole philosophy is to flip drug discovery 180 degrees. Instead of starting with disease and target, we're starting with the cell.
—Yerem Yeghiazarians, Soley Therapeutics
"Our whole philosophy is to flip drug discovery 180 degrees," Yerem Yeghiazarians, cofounder and CEO of Soley Therapeutics, told DDN. "Instead of starting with disease and target, we're starting with the cell."
The approach stems from more than 15 years of research at the University of California, San Francisco, where Yeghiazarians, an interventional cardiologist and stem cell researcher, joined forces with cancer biologist Kurosh Ameri to answer a deceptively simple question: why do some cells survive severe stress while others die?
Listening to cells
Cells constantly monitor their surroundings, responding not only to changes in oxygen, temperature, and mechanical forces — discoveries recognized by Nobel Prizes in 2019 and 2021 — but also to drugs. Yeghiazarians emphasized that these responses contain rich biological information about whether a compound is helping, harming, or fundamentally changing cellular behavior.
The challenge, however, is translating that biological information into something researchers can interpret.
Traditional phenotypic assays typically capture only a snapshot of a cell after it has been fixed or destroyed, while molecular approaches such as transcriptomics and proteomics can generate vast datasets that are difficult to interpret and often impractical for high-throughput drug screening. Soley instead set out to understand how a cell's state evolves over time.
To capture that information, Soley tracks how hundreds of stress-response proteins change their location, abundance, and behavior after cells encounter different compounds. These measurements are combined with live-cell imaging, self-supervised machine learning, and proprietary bioinformatics to create a multidimensional picture of cellular response.
"We have taken this cellular language and translated that into understandable human language,” Yeghiazarians said. Soley now generates around three terabytes of biological data every day through automated high-throughput screening, imaging, and analysis. Its platform currently screens approximately 100,000 compounds each week across hundreds of cell lines.
Soley has also built a database that includes response profiles from FDA-approved chemotherapies, thousands of bioactive compounds that failed during development, and known toxic molecules. The goal is to use these data to predict both the efficacy and toxicity of new compounds earlier in the discovery process, while also identifying potential off-target effects by comparing cellular response patterns across drugs.
Beyond one target at a time
The platform also reflects a broader shift in thinking about complex disease biology. For conditions driven by single genetic defects, target-based drug discovery has delivered remarkable success. But diseases such as Alzheimer's disease, Parkinson's disease, ALS, and many cancers involve interconnected biological networks rather than one clearly dominant pathway.
"Despite the progress in science, medicine, and technology, for many complex diseases, we still do not know the right target to pursue,” Yeghiazarians said. “No one knows what causes ALS, brain cancer, or pancreatic cancer. So how can you confidently say this is the one target to go after? For complex diseases, one target may not be sufficient. You may need multiple targets.”
Rather than assuming a single target is responsible, Soley's platform evaluates how the entire cell responds to a compound before working backwards to understand its mechanism of action.
This strategy allows researchers to discover molecules that naturally produce coordinated effects across multiple biological pathways. Once promising compounds are identified, they work backwards to understand how they reshape cellular state, using proprietary deconvolution methods to identify the networks of targets and mechanisms of action responsible for the observed effects. The compounds are then optimized through target-independent medicinal chemistry, while additional computational analyses help predict which diseases — and ultimately which patient populations — are most likely to benefit.
Two directions from one platform
The same platform underpins two complementary therapeutic strategies. In oncology, Soley searches for compounds that selectively increase stress in cancer cells. Cancer cells already exist under substantial metabolic and environmental stress, making them potentially more vulnerable to additional perturbations than healthy tissue. The goal is to push tumor cells beyond their survival threshold while sparing normal cells.
Outside cancer, the objective is reversed. Instead of amplifying stress, the platform identifies molecules that restore healthier, more resilient cell states in diseases characterized by chronic stress or exhaustion, including neurodegeneration and metabolic disease. Yeghiazarians likened the concept to induced pluripotent stem cell technology, but instead of reprogramming cells through genetic manipulation, Soley aims to reshape cell state using small molecules. Soley is currently applying this approach to Parkinson's disease, obesity and metabolic disease, and hair regeneration, with plans to expand into other neurodegenerative disorders.
Building a pipeline
In the past two years, the platform has discovered 10 first-in-class oncology assets, with two lead oral small molecules expected to enter Phase 1 clinical trials in 2027 for separate hematologic and solid tumor indications.
If these first clinical candidates perform as hoped, Soley’s approach could help demonstrate that drug discovery doesn’t always need to begin with a predefined molecular target. Instead, it may be much more worthwhile to listen to what cells have been telling researchers all along.
This article is part of The Spin-Off, a Drug Discovery News series that follows the journey from academic discovery to biotech company. From unexpected biological insights to first-in-class therapeutic approaches, university research often lays the foundation for the next generation of medicines. The Spin-Off follows the journey beyond the lab, examining the people, technologies, and translation challenges involved in turning scientific discoveries into biotech ventures.
