Research has HIV-1 and cancer treatment applications
Individual cells in a population often show dramatically different responses to stress or other environmental challenges, even if the cells are all genetically identical.
Recent research has shown that randomness in gene expression (the process by which genetic information in our DNA is converted into a functional product, such as a protein) can lead to different responses of individual cells to the same environmental stress. In particular, what scientists refer to as rare events (e.g., very high or very low protein production in a cell, relative to the population average) can lead to the survival of a small subset of cells under stressful conditions. For example, drug exposure leads to cell death for most cancer cells in a tumor, but a rare subset of cells survive, which eventually leads to the development of drug resistance.
Associate Professor of Physics Rahul Kulkarni says that in order to effectively treat diseases such as cancer and HIV-1 viral infection, it is critical to develop quantitative stochastic models of the underlying process of gene expression.
The National Science Foundation’s Divisions of Mathematical Sciences and Molecular and Cellular Biology have awarded Kulkarni a three-year $309,998 grant for his project, “Large Deviations and Driven Processes for Stochastic Models of Gene Expression and Its Regulation.” It’s the latest grant for Kulkarni, who has been continuously funded by the NSF since his arrival on campus seven years ago.
“Given that gene expression in a single cell is a random process, can we predict the probability of rare events?” Kulkarni said. “Computationally what you would do is you would say, ‘I don’t really know, so let me simulate the process say a million times until I get that rare event.’ Then I have just one example, so you need to repeat the process many more times.
“The problem is that you want to look at a random process, but condition it on a rare event occurring. How do you do that? You can define a different process, called the driven process, for which the rare event is now a typical event,” Kulkarni said. “Imagine that you are tossing a fair coin. You toss it 1,000 times and say 90 percent of the time you get heads. That would be a rare event for a fair coin. However, for a biased coin with a 90 percent chance getting heads, that would be a typical event. So you want to figure out a process that is such that the rare event for the original process is a typical event for the driven process. So that is something that we are going to determine for gene expression models and then we look at models of HIV-1 gene expression. We are also going to look at models for gene expression in cancer and try to understand how the different molecular parameters control the probabilities of rare events.”
The grant will be used to support graduate students in UMass Boston’s Computational Sciences PhD Program, an interdisciplinary program housed in the College of Science and Mathematics. The interdisciplinary nature of the program and the expertise the students bring is key to the success of this project.
“The analysis requires tools and approaches from physics and applied mathematics, which will be integrated with teaching efforts to effectively train students and future scientists focusing on interdisciplinary research in the life sciences,” Kulkarni said.
The students will develop computational models and establish collaborations with experimental groups at the University of California at San Francisco and the University of Pennsylvania. The work is potentially life-changing. A particular focus will be modeling the phenomenon of latency in HIV-1 viral infections, which is considered the most significant obstacle to eradicating HIV-1.
“What we can do is try to see signatures of cellular regulatory mechanisms controlling the rare events and maybe a screen [could be developed] based on that, but for that to happen, we need to first develop the framework,” Kulkarni said.