Interactive SCExplorer workflow

This article shows how to prepare one or more processed Seurat objects for interactive exploration with SCExplorer. Use this workflow when static figures are not enough and users need to inspect cell groups, embeddings, marker genes, or multiple datasets through a Shiny interface.

SCExplorer is best treated as a deployment layer for already processed objects. Run preprocessing, quality control, integration, annotation, and key result checks before exporting the app directory.

The code below is intentionally shown without execution during pkgdown builds. Run it in a local interactive R session when preparing a deployment artifact.

The app does not replace the analysis report. It exposes the processed objects so collaborators can inspect embeddings, metadata, and marker expression behind the static figures.

Prepare Objects

Start with named Seurat objects. The names become dataset identifiers in the exported SCExplorer directory.

library(scop)

data(pancreas_sub)
data(panc8_sub)

pancreas_sub <- standard_scop(pancreas_sub, verbose = FALSE)
panc8_sub <- integration_scop(
  srt_merge = panc8_sub,
  batch = "tech",
  integration_method = "Harmony"
)

Before export, verify that each object has the reductions, metadata labels, and feature layers expected by the app. Anything left in metadata can become visible to users, so remove temporary columns or rename them before export if they would be confusing.

SeuratObject::Reductions(pancreas_sub)
SeuratObject::Reductions(panc8_sub)

colnames(pancreas_sub@meta.data)
colnames(panc8_sub@meta.data)

Export the App Directory

PrepareSCExplorer() writes the app data bundle. Keep base_dir explicit so that the exported directory can be archived, copied, or served later.

PrepareSCExplorer(
  list(
    mouse_pancreas = pancreas_sub,
    human_pancreas = panc8_sub
  ),
  base_dir = "./SCExplorer"
)

list.files("./SCExplorer")

The exported directory is the portable artifact. It can be used locally or as a site directory for a Shiny Server deployment. The directory contains serialized data and app configuration, not a new analysis. Recreate it whenever upstream preprocessing or annotations change.

Run the App Locally

RunSCExplorer() returns a Shiny app object. Run it interactively during local inspection.

app <- RunSCExplorer(base_dir = "./SCExplorer")

if (interactive()) {
  shiny::runApp(app)
}

Decide What to Include

SCExplorer is more useful when the exported objects are curated. Before export, consider keeping:

• stable embeddings used in the manuscript or report;
• annotation columns that users should compare;
• QC columns that explain filtering decisions;
• feature layers needed for marker inspection;
• integrated reductions for multi-dataset objects.

Avoid exporting many redundant intermediate reductions or temporary metadata columns. They make the app harder to navigate and increase deployment size. For multi-object exports, use dataset names that are meaningful to readers, because those names become the first navigation layer in the app.

What to Report

A compact SCExplorer deployment note should include:

• the objects included and their dataset names;
• the preprocessing state of each object;
• the exported base_dir;
• the reductions and metadata columns intended for user inspection;
• whether the directory is meant for local use or Shiny Server deployment.

Treat SCExplorer as a companion to static reports. Static figures should still carry the key conclusions; the app should make the underlying object easier to inspect.