A pipeline for domain detection and annotation of spatial metabolomics data using hierarchical clustering and Shiny app integration
The spatialMET pipeline consist of three steps:
-
Data input: Processing raw files with Cardinal. This step raw files from equipment (e.g., .ibd and .imzML from MALDI-MSI) and produces text files containing pixel coordinates and peak intensities by pixel.
-
Spatial domain detection: Detection of spatial domain using hierarchical clustering using a computationally-efficient C algorithm named hcdist.
-
Exploratory data analysis: A Shiny app for user-frienly, point-and-click visualization and downstream analysis of hierarchical clustering results from hcdist. Downstream analyses include differential abundance test and calculation of spatial statistics for "hotspot" detection (Moran's I and Geary's C). The app also enables manual selection and annotation of regions of interest (ROIs).
Use the spatialMET Shiny app Docker container hosted at Docker Hub
Make sure Docker Desktop is installed in the computer. Then, using the command-line type:
docker run --rm -p 3838:3838 oscareospina/spatialmet_app
Building the spatialMET Shiny app
The spatialMET Shiny app can be build from the Dockerfile included in this repository:
- Download this repository and de-compress it.
- Navigate to the repository folder using the command-line
- Build the container:
docker build --no-cache -t spatialmet_app .
- Start the container:
docker run --rm -p 3838:3838 shinyapp
- Open a browser and type in the address: http://localhost:3838/