Koseki J. Kobayashi-Kirschvink

University of Chicago
Department of Medicine
5841 S. Maryland Ave
Chicago IL, 60637

Hi! I'm Koseki Kobayashi. I recently joined the University of Chicago as an Assistant Professor in the Department of Medicine, and actively searching for motivated PhD students or postdocs to join my lab! (I'm almost finished making a cool lab website, so please check back soon!)

My lab is interested in developing optical and genomic tools to study dynamical biological processes, such as finding fundamental principles that govern how cells self-organize into complex multicellular systems (e.g., embryos and organoids), and how tumors arise from otherwise healthy tissues. One way to do this is by combining the strengths of label-free microscopy (e.g., Raman microscopy) and genomic profiles (e.g., single-cell RNA sequencing) using machine learning, which led to the development of my recent work Raman2RNA. Raman2RNA showed that genomic profiles could be predicted in live single-cells using deep neural networks, demonstrating a type of live-cell omics, allowing to track genomic changes over time of the same exact cell. I also develop novel high-speed Raman microscopes that aim to overcome the long measurement times that are typical in standard Raman microscopes. Breaking the 'time barrier' is actually important, as it can open up new applications for studying expression dynamics at scale, in vitro and in vivo.

In my spare time I enjoy playing baseball and snowboarding, and recently learned how to sail.

'Live-cell omics' with Raman2RNA

Single-cell RNA-seq and other omics assays have provided unprecedented insight into cellular properties, regulation, dynamics, and function. However, because these assays are inherently destructive, they do not allow us to track temporal changes in living cells. To overcome this limitation, we developed Raman2RNA (R2R), an experimental and computational framework that infers single-cell expression profiles in live cells using Raman microscopy and machine learning. Raman2RNA can be used to unravel principles that govern cellular dynamics, such as the emergence of cellular heterogeneity and the evolution of cell states in embryo genesis or emergence of tumors from healthy tissue.

High-throughput Raman microscopy

Raman microscopy is a powerful tool that reports on the vibration energy levels of molecules non-destructively, but traditional systems are often limited by their speed and sensitivity, severely limiting its broad usage. I have developed high-throughput Raman microscopy systems that use novel optical designs and machine learning algorithms to improve the speed and sensitivity. These systems can be used to study a wide range of biological samples, including live cells and tissues.

Spatial multiomic landscape of the human placenta

I also was involved in a project that mapped the spatial multiomic landscape of the human placenta at molecular resolution. This work involved the use of state-of-the-art spatial multiomics methods to study the structure and function of the placenta at a cellular level. We developed a new method to infere gene expression dynamics with transcriptomics and chromatin accessibility information, predicting the near-term future of cell fates from genomic profiles alone. The results of this study have important implications for our understanding of placental biology and its role in human health and disease.

Academic Employment

Education

Select Publications