New study offers revolutionary method for analyzing cell interactions in cancer microbiologystudy

A new paper from Elham Azizi’s lab and collaborators has been accepted for publication in Genome Research, marking a significant advancement in the study of dynamic single-cell interactions.

The study led by Cameron Park and Shouvik Mani introduces a tool called DIISCO, that aims to improve our understanding of how cells in the body interact over time, with a particular focus on cancer and immune cells.

Cell–cell interactions play a critical role in how our bodies respond to treatments. In the context of cancer, immune cells interact with each other and with cancer cells, determining whether a patient’s disease goes into remission or relapse.

Tracking how these interactions evolve during treatment is key to understanding why some patients respond well, while others do not. However, current tools do not account for these changes over time, limiting their ability to capture the full picture of these cellular interactions.

To address this gap, the Azizi lab developed DIISCO, a new method that analyzes how cell interactions change over time using single-cell RNA sequencing (scRNA-seq) data. Traditional methods provide a static view, but DIISCO looks at longitudinal scRNA-seq samples to track how cells respond to treatments in a dynamic, time-sensitive way.

The team built a probabilistic model using a network of Gaussian Processes to analyze how the cell composition changes over time, and these changes reflect ongoing interactions between cell types.

The model was tested on both simulated data and real data from a co-culture experiment involving T cells and lymphoma cells. DIISCO proved to be more robust and accurate than existing methods, even identifying time-varying receptor-ligand genes that other techniques missed.

The study shows that DIISCO has the potential to transform how researchers study cell-cell interactions, particularly in the tumor microenvironment. By understanding how these interactions evolve during cancer treatments, scientists can uncover important insights into why some therapies succeed or fail. The model’s success in predicting cell dynamics in both simulated and real-world data demonstrates its potential as a powerful tool for cancer research and treatment.

“Our hope is that researchers will find DIISCO to be a valuable tool for characterizing temporal cell-cell interactions on their own datasets. As scRNA-seq becomes more widespread and cost-effective, building temporal datasets is becoming increasingly common, and DIISCO is well-suited to meet this growing demand.

“The method also has broad clinical applications, particularly in understanding dynamic cellular behavior over time,” explained Cameron Park, Ph.D. student in the Department of Biomedical Engineering and lead contributor to the study.

With the rising popularity of single-cell sequencing and decreasing costs, the team hopes DIISCO will be widely adopted by researchers looking to study time-dependent interactions in their own data.

Looking ahead, the research team plans to apply DIISCO to a larger, more heterogeneous dataset from relapsed leukemia patients to further validate its utility. Efforts are also underway to expand DIISCO to incorporate spatial data, enabling more detailed insights into how cells interact within the body.

In addition, the Azizi group is exploring ways to modify the framework to accommodate other types of biological data, broadening its potential impact across various areas of biomedical research.