Computational breakthrough could shed light on cancer’s beginnings

Biomedical engineering research has developed a computational platform that can link changes in cell environment to aberrant cellular organization

Bahram Parvin

Bahram Parvin has spent the last 10 years trying to understand how minute changes in a cell's environment might cause aberrant organization. He's looking for premalignant cellular formations that could shed light on how cancer develops.

As part of his work to answer this question, Parvin, a professor of electrical and biomedical engineering, and his team developed an award-winning computational platform that allows researchers to analyze and manipulate 3D cell culture models.

The platform is called BioSig3D, and the web-based system was a 2014 R&D 100 winner. Its potential is significant for cancer researchers, who can use the system to design experiments and then generate sophisticated 3D models of how cells respond under changes in their environment - either identifying changes that lead to premalignant states or evaluating how tumor cells respond to a given treatment.

Schematic explaining BioSig3D

Schematic explaining how BioSig3D works.

Research combines computational methods and life sciences

Parvin's research thrives at the intersection of cell biology and computational science. To analyze 3D cell culture models, which provide unique insights into colony organization not available with 2D models, Parvin and his team had to develop new computational methods to parse the rich representations.

"What attracted me to this field is the complexities of biological systems," said Parvin. "It's pushing engineers to think about new ways to develop new techniques for in silico biology, and by doing that you move forward. I think that understanding biological complexities will be a driver for the field of engineering."

Bringing the complexities of the life sciences into a computational system requires enormous interdisciplinary expertise.

"The main challenge is having the necessary talent focused on different aspects of the problem so that they can integrate their science in a more effective way," said Parvin, who joined the University in 2014 from Lawrence Berkeley National Laboratory. "It's not just about computation but it's also about being able to select the right model system, optimizing the assay, and using the right set of tools for readout and validation. Everything along the way needs to be optimized so that your assay is quantitative and repeatable."

A second strand of Parvin's research focuses on computational histopathology. Using images of tumor sections from The Cancer Genome Atlas, he and Hang Chang, a member or Parvin's lab at LBL who joined the University as an assistant professor this fall, have developed algorithms that can link the molecular organization of various tumor types to clinical data on patient outcomes, paving the way for precision medicine in cancer treatment. The work is so computationally intensive that it takes several weeks to run through a group of 1000 samples. Eventually, the team hopes to migrate to the cloud to do the processing.

Microbial imaging could help provide targeted therapies

Now, Parvin is turning his attention to the growing field of microbial imaging. Parvin and Qingsu Cheng, a postdoctoral scholar, are working on developing probes that can image and uniquely identify microbes in situ; ultimately, he hopes to better understand the diverse communities of microbes teeming around human bodies and the natural environment.

"If I were to look 5 years from now, being able to have a unique label for every microbe and being able to look at them live and monitor their growth and molecular machinery in their own natural environment, I think that's going to be exciting," said Parvin.

If Parvin achieves that, his research program will have come full circle.

"They're all interrelated," he says of his three research thrusts. "If you are successful in identifying an aberrant pathway through computational histopathology, then maybe you could study the mechanisms in a 3D cell culture model, and then if you have the probes you can actually have targeted therapy. So there are opportunities to bring these three different things together. But that's very, very long term."