A Shiny webapp for creating a Seurat object without programming expertise. Load a count matrix directory, interactively filter on metadata, view results, download Seurat object and results of differential gene expression testing.
Currently hosted on Scilifelab Serve: https://som.serve.scilifelab.se/app/som.
To install:
devtools::install_github("razofz/SSOMaker", upgrade = "never")After installation, to run it:
SeuratObjectMaker::run_SOM(port = 3838)and open http://localhost:3838 in your browser.
If this is running on a server (i.e., the matrix files you want to load are not on the same computer as you're running this), adjust the line e.g. like this:
SeuratObjectMaker::run_SOM(launch.browser = F, port = 32345, running_locally = F)You can also run this Docker container: razofz/seuratobjectmaker:0.0.2.2 (also port 3838).
Selecting directory, with data validity checks:
Filtering page, setting thresholds with sliders. Violin plots and nbr cells left after filtering are updated in realtime.
The thresholds limits are updated in real-time, press "Confirm filtering thresholds" to continue:
Showing a loader while processing the data..
Results page:
More of results, differentially expressed genes per cluster (an accordion, can open for a cluster of interest at a time (to not swamp page if many clusters)):
Having pressed the "Download Seurat Object" button, the user is prompted to save the object as a .rds file (a filename including the date is suggested by default):
Note that "debugging" values are still printed etc, the webapp is not mature yet. Code structure is also "prototype-y". But the app is functional.