At the UCSF School of Pharmacy, Bo Huang, PhD, a professor in the Department of Pharmaceutical Chemistry, is blending biological discovery with tool-building to reimagine how scientists visualize and understand life inside cells.

His lab builds cutting-edge microscopy technologies and uses AI to interpret massive datasets, while also investigating protein condensates — dynamic assemblies that influence gene expression and can become dysregulated in cancer. Huang is currently leading an NIH-funded project to develop microscopy and computational platforms for analyzing mammalian cell libraries and facilitating easy adoption by research labs everywhere.

"I’ve always had a hard time curbing my curiosity,” Huang said. “I’m interested in too many questions, and many times these questions are of interest to me because they are too challenging to be studied by the methods we have.”

Microscopy as a discovery engine

Seeking out new methods and welcoming surprise is how Huang describes his scientific philosophy.

"Microscopy is my favorite technology because you can see things that are totally unexpected, and the advancement of AI allows us to extract knowledge from massive scale microscopy data,” he said. “Bridging the microscopy data with other domains — the perturbations, genetics, sequencing results — is giving microscopy a comeback as a primary discovery tool.”

At UCSF, the pace of discovery is fueled by a collaborative, interdisciplinary ecosystem. Huang recalled how a chance hallway conversation in 2013 resulted in a breakthrough, the ability to visualize genes in living cells using CRISPR technology that was new at the time.

A cross-disciplinary team and experiment formed within a week, thanks to immediate proximity in Byers/Genentech Hall and support from CRISPR pioneers Lei (Stanley) Qi, PhD, Luke Gilbert, PhD, and Jonathan Weissman, PhD, as well as Nobel Laureate Elizabeth Blackburn, PhD. With postdocs and experts in viral delivery pooling their knowledge, what began as a spontaneous idea became a fully developed imaging method, submitted and published within the same year.

“All the resources, the expertise, the ideas are just readily available all around us at UCSF,” Huang said. “It’s a biologist’s paradise.”

Pushing image data into predictive models

That same collaborative environment accelerated Huang’s push to interpret unprecedented volumes of microscopy data using AI, as imaging outpaced human analysis.
Huang describes working with Emaad Khwaja, PhD, then a first-year bioengineering student, who proposed that the amino acid sequence of a protein could be treated like “text” for text-to-image AI models. Huang initially doubted the idea, but his curiosity opened the door to a new way of predicting what a protein should look like under the microscope, based solely on its sequence, a model that revealed new computational possibilities.

"What amazed me is that the model could go from sequence to image or from image back to sequence,” Huang said. “It could translate between data types we usually keep separate.”
The project also demonstrated that AI could help translate raw cellular visuals into quantitative, testable biological insights, laying the groundwork for new discovery approaches in pharmaceutical science.

Condensates and new therapeutic paths

These tools and findings feed into Huang’s second major research area: biomolecular condensates, which have emerged as important regulators of gene expression and cancer biology.
In a collaboration with the lab of Trever Bivona, MD, PhD, a professor in the Department of Medicine, imaging revealed that cancer-driving fusion proteins were organizing into cellular condensates, not membrane structures, as previously thought.

"We now understand that condensates are very prevalent. Some are physiological, some are pathological, and the cell must have some quality control system for them like cells do for proteins,” Huang said. “Our focus is to figure out what this is, so we could utilize this enormous mechanism to get rid of a pathological current state, and also to help prevent a healthy current state from becoming pathological.”

A vision for the future

Advancing science and discovery at the pace of technological innovation requires navigating a constant tension, Huang said.

"The conceptual conflict to resolve is between the years of knowledge and traditional practice, versus staying on top of new methods before they are really ready to use. You can’t wait for everything to be ironed out because you’d get left behind,” he said.

As chemistry, cell biology, optics, and computation increasingly converge at UCSF, Huang’s curiosity persists. “It’s still my dream to be able to see how every single molecule in a cell is dancing around and doing its job in real time.”