Alladi Ramakrishnan Hall

The mechanistic view of spatial biology

Ankit Agrawal

University of Würzburg, Germany

Spatial biology is like the study of tissue architecture at a molecular and morphological level. It enables single-cell analysis and maps a cell's spatial architecture to its surroundings. Furthermore, tissue architecture is governed by the nature and integrity of cellular variation determined by intrinsic factors such as gene expression noise and extrinsic signals from the niche microenvironment. In the first project, we try to understand bone tissue growth strategy based on the morphology of the cell lineage clusters from the mouse model. As we know, both oriented cell divisions and cell rearrangements are critical for proper embryogenesis and organogenesis. But little do we know how they are responsible for the elongation of long bones. We developed a pipeline to compare the cell clone clusters' various morphological features based on the fact that cell divisions and arrangements follow certain rules. We found that the orientation of the cell growth clusters changes from embryo to postnatal bone, suggesting switching the circumferential bone growth strategy to elongated bone growth strategies. The earlier division and cell arrangement rule proposed for the elongation of the bone is no longer true now, and it looks like embryo bone follows random division and cell arrangement while postnatal bone follows some complicated rules. These findings are novel to the skeletal development biology and biomedical research community and provide new insights into the mechanisms generating appropriate tissue architecture.


In the second project, we try to understand the nature and driving forces of cell state transitions by deconvolution of cell-intrinsic and - extrinsic components of cell state. Regarding this, we are developing a computational method REEST (Regression- based NichE prEdiction in Spatial Transcriptomics), that can quantify the cell state variability explained by cell state variations of cell types within the local tissue microenvironment. The REEST pipeline performs cell type annotation of spatial data by label transfer from single-cell RNA-seq-derived clusters, predicts significant local cell type interactions, and infers covarying genes in a cell type and its local neighbors to identify covarying pathways indicative of inter-cellular crosstalk. Applying REEST to spatial liver data, it observes known and previously unknown cellular interactions, e.g., between lymphatic endothelial cells (EC), fibroblasts, and cholangiocytes, interactions, and pinpoint covarying pathways between these populations. In conclusion, REEST can infer spatial determinants of cell state variation from data generated by high-resolution imaging-based technologies, which have become increasingly available, together with single-cell RNA-seq-based cell type reference data.


Ref: Rubin, S., Agrawal, A., Stegmaier, J., Krief, S., Felsenthal, N., Svorai, J., ... & Zelzer, E. (2021). Application of 3D MAPs pipeline identifies the morphological sequence chondrocytes undergo and the regulatory role of GDF5 in this process. Nature communications, 12(1), 1-16