Looking for Giotto, the spatial toolbox package?
Please click on the following link to redirect to the new dedicated site. http://spatialgiotto.rc.fas.harvard.edu


The rapid development of novel technologies that can profile hundreds to thousands of single-cell transcriptomes in situ has provided new opportunities to investigate the interactions between cells and their native microenvironment. Our group has developed a number of computational tools to facilitate comprehensive exploration of the current and upcoming spatial datasets, including Giotto, a spatial trancriptomic and proteomics analysis pipeline, and developed a tool for seqFISH and single-cell RNAseq integration (as described below). A separate website has been made for Giotto (http://spatialgiotto.rc.fas.harvard.edu), where all updates and usage information are posted.

Spatial domain inference using HMRF

Hidden Markov Random Field model for detecting spatial domains with coherent gene expression patterns.

A domain may be formed by a cluster of cells from the same cell type, but it may also consist of multiple cell types. Integrating cell-type and spatial domains analysis can provide insights into the roles of intrinsic regulatory networks and spatial environment in the maintenance of cellular states (Zhu et al. Nature Biotechnology. 2018).

Cell type mapping
SeqFISH contains only 125 genes. We built a machine learning model to predict cell types on seqFISH leveraging whole transcriptome scRNAseq dataset. By combining information from cell types and spatial domains, we show that we can systematically decompose cellular variations from a scRNAseq and seqFISH integration.

Supplementary Website    HMRF Repository    Paper (nbt.4260)

This website is created by Guo-Cheng Yuan's lab.