Users Track

If you’re running microscopy samples and want to go from raw TIFFs to difference images without writing Python, this is your track. It covers installation, your first run using the GUI, how to read the diagnostic plots, how to zoom in on regions of interest, and what to do when things don’t look right.

• Before You Start
  • What you’ll need
  • 1. Opening a terminal
  • 2. Creating a GitHub account
  • 3. Installing Python
  • 4. Installing JupyterLab
  • You’re ready
• Installation
  • Requirements
  • Getting the code
  • Installing the dependencies
  • Setting up your raw image directory
  • Verifying the install
• Quickstart
  • Launching the GUI
  • Survey your raw channels (recommended)
  • Picking your samples and dyes
  • Picking a scale method
  • Running
  • Watching the diagnostic plots
  • Finding your results
• GUI Tour
  • Input Selection
  • Scale
  • Execution Options
  • Advanced Parameters
  • The three action buttons
  • Parameter Comparison
  • 🔍 Zoom & Inspect
  • 👁️ Blink Comparison
  • Progress bar and output
• Scale Methods
  • The problem
  • Auto (Moffat fit) — the default
  • Ratio percentile
  • Manual override
  • All three at once
  • Reading the excess heatmap
• Interpreting Results
  • Pre/post heatmap
  • Excess heatmap
  • Pre/post histograms
  • Difference histogram
  • Ratio histogram with Moffat fit
  • Difference image
  • What to do next
• Zoom & Inspect
  • Why this exists
  • Basic workflow
  • Side-by-side comparison
  • Tips
  • Using the zoom function from Python
• Memory and Performance
  • What “serial” and “parallel” mean here
  • Why parallel mode uses more memory
  • When to leave parallel mode OFF
  • When to turn parallel mode ON
  • Checking your RAM
  • Watching memory during a run
  • Pre-flight resource checks (new in 2.4)
  • Subprocess mode for low-RAM machines (new in 2.4)
  • Diagnosing memory issues (new in 2.4)
  • What exo2micro does on its own to manage memory
• Troubleshooting
  • “No raw files found” / “Raw image directory not found”
  • “Some requested (sample, dye) pairs have no raw files”
  • “Estimated peak RAM exceeds available RAM”
  • “Estimated output size exceeds free disk space”
  • “Subprocess killed (likely OOM)”
  • “Kernel dies mid-batch” or “kernel restarts during run”
  • Alignment doesn’t look right
  • Wrong channel detected
  • Scale estimate looks implausible
  • Large negative patches in the difference image
  • Banding or striping in the difference image
  • Empty or nearly-empty difference image
  • Pipeline says a checkpoint is missing
  • The GUI is slow or unresponsive
  • Filename problems