Current approaches profiling the transcriptome of cells in tissues offer a rich yet immobile snapshot of biology. In collaboration with the Kosuri Lab at the Salk Institute, we are developing approaches to infer cell kinematics and tissue infiltration timelines to obtain a finer temporal view of key processes that impact T cell differentiation and function during an anti-tumoral response. This project will rely on the development of both computational tools and wet lab protocols.

Understanding how tissue immune networks operate require obtaining the position, transcriptional state and environmental context of each cell within an organ. We use spatial transcriptomics, combined with protein stains and other approaches, to survey immune cell responses in mouse and human tissues using models of infection and cancer. Ongoing efforts in the lab are focused on establishing pipelines that enable analysis of these large datasets. We employ two different spatial transcriptomics technologies: optically-read probe-based transcript capture at single-molecule resolution (Xenium) and array-barcoded probe-captured sequencing at ~2 um resolution (Visium-HD). Your project will focus on developing tools to obtain biological insights, including: custom probe set design based on training data (feature selection algorithms), advanced probe set design (sequence optimization algorithms), cell segmentation and image registration, integration of spatial transcriptomics and protein readouts (Python and R), batch correction and multi-tissue integration, custom binning and pseudo-single nucleus analysis (Visium-HD), cell typing, neighborhood enrichment analysis, spatial gene set enrichment analysis, cell networks and ligand-receptor interaction analysis. These tools will be applied to the analysis of an ongoing consortia initiative aimed at profiling every immune cell in the mouse (ImmGenMaps). In addition to developing the computational aspect of this work, you will be trained and encouraged to design and execute experiments to generate your own datasets.