From bacteria eaters to cancer-cell destroyers, the immune system enlists a diverse army of cells to defend the body against illness. When fighting the same disease-causing agent, however, one individual might experience nothing more than a cough that disappears in a few days, while another may have more severe symptoms like breathing difficulties.
These differences at the individual level are rooted in subtle variations at the single-cell level—a tiny world, yet responsible for so much of biology’s complexity. But how exactly these immune cells work together and what happens when they fail have remained unresolved questions.
Jinmiao Chen, a Principal Investigator at A*STAR’s Singapore Immunology Network (SIgN), is harnessing single-cell technologies and artificial intelligence (AI) to characterize the medley of activities that immune cells are engaged in. By developing novel analytic methods, her group makes sense of cellular data such as genetic information and signals used in cell–cell communication, to understand how differences at the cellular level manifest as different responses to diseases and treatments.
In this interview with A*STAR Research, Chen dives deep into the complex world of immune cells and shares how her analytical work may serve as a springboard for advancing research on precision immunotherapies that are tailored to match patients’ individual needs and immune system characteristics.
1. What are the major challenges that you aim to solve with your research?
Characterizing immune cells and cell–cell interactions is critical for understanding the mechanisms used by immune cells to promote disease progression and response to treatment. While diverse immune cell subsets have been characterized, much less is known about the interactions between subsets of immune cells, as well as the interactions of immune cells with non-immune cells like cancer tumor cells.
To examine these interactions, single-cell technology is generating large and complex datasets—creating atlases of immune cells. However, our ability to generate complex, high-dimensional data has far outstripped our ability to analyze and integrate it. Comparison between atlases has been difficult owing to the plethora of protocols used. Moreover, many studies have generated large volumes of overlapping data. There is thus an unmet need for unified analysis, integration and annotation of these datasets to reveal synergy between studies.
Single-cell technologies will become more and more accessible, cost-effective and widely applied in translational research for a broad spectrum of diseases.
2. Given your background in computer science, what led you to bioinformatics research, and in particular, single-cell analysis technologies?
During my PhD degree, I developed new AI techniques for data analysis, but my method was not widely used by biologists. I started to wonder how I could use my expertise to help biologists. When single-cell analysis first emerged, I was fascinated by its power and realized that this approach produces big and complex data, which is ideal for AI systems to process. Moreover, joining SIgN allowed me to work with different labs and learn about the beauty of immunology. Altogether, these inspired my research on combining single-cell technologies and AI for precision immunology.
3. What is the impact of the single-cell approach in immunology?
Understanding the interplay between the immune system and disease—and the huge variation in individual response due to fine-grained differences at the single-cell level—is a vital area of research. With single-cell ‘omics technologies, we have analyzed immune cells at increasing scale and resolution.
Two newly emerging technologies called single-cell multi-omics and spatial ‘omics have been transforming our understanding of the immune system. Multi-omics can identify subtle differences between cells, allowing us to dissect variable immune responses in cancer, autoimmune and infectious diseases. Spatial ‘omics technologies, which simultaneously measure gene expression, protein production and cell location, enable the reconstruction of tissue structure and cell-cell interactions. This is critical for understanding the tumor microenvironment and characterizing interactions between tumor cells and immune cells.
4. What are some of the most interesting research projects that you are pursuing right now?
By integrating existing public datasets into a comprehensive and unified atlas, our lab is generating well-annotated big data that will be useful for the development and training of AI models. Together with data-driven AI, Deep Integration of Single-Cell Omics (DISCO) provides immunologists with a reference Google map for studying the immune system in healthy and sick individuals. It currently carries data on six million cells across various tissues, from bones to the brain.
We also developed a spatial ‘omics pipeline called the Unsupervised Spatially Embedded Deep Representation of Spatial Transcriptomics (SEDR) that combines genetic data with spatial relationships, accelerating follow-up analysis and integration tasks. This method was highly accurate in retracing the development of a brain region, identifying not only the genes expressed but also where they needed to be activated.
Meanwhile, in an ongoing study on gastric cancer, we discovered cell subsets that favor tumor growth, suppress the immune response and are resistant to chemotherapy. These were associated with poor clinical outcomes for patients.
5. How would your work in single-cell analytics support advances in medicine?
Soon, single-cell technologies will become more and more accessible, cost-effective and widely applied in translational research for a broad spectrum of diseases.
Our single-cell studies contribute directly to Singapore’s national precision medicine research strategy. For example, the DISCO atlas characterizes all immune cells in healthy and diseased states, building a critical foundation for studying individual differences in immune responses. These atlases, combined with single-cell analysis of dysfunctional immune responses, will advance the development of personalized immunotherapies.
6. How do you plan to take your research forward in the next few years?
Our lab will continue to build DISCO, including detailed, zoomed-in atlases on immune development from infants to the elderly, COVID-19 and other viral diseases, autoimmune diseases and various cancers. We will also use spatial ‘omics and AI models to characterize tumor-immune cell-cell interactions in mouse breast cancer models before and after immunotherapy, as well as predict breast cancer patients’ responses to immunotherapy.