Workflows and Data Flow

Workflows and Data Flow

This page shows how the three tabs fit together and recommends an end-to-end workflow.

A detailed flow diagram for each stage is on its own page: Segmentation, Bleed-Through Correction, and FRET Analysis. The cross-tab pipeline is summarized below.

---

The pipeline at a glance

 Raw images (.tif / .czi)
        │
        ▼
┌───────────────────────────┐
│ 1. Cellpose & Manual      │  segment cells → label mask
│    Segmentation           │  refine ROIs by hand
└───────────────────────────┘
        │  segmented stacks (mask + channels)
        ├──────────────► Send to Donor / Acceptor ──┐
        │                                            ▼
        │                              ┌───────────────────────────┐
        │                              │ 2. Bleed-Through          │  fit S1/S2 (+S3/S4)
        │                              │    Correction             │  → bt_params.json
        │                              └───────────────────────────┘
        │                                            │ coefficients
        ▼                                            ▼
┌─────────────────────────────────────────────────────────────────┐
│ 3. FRET Analysis                                                  │
│    apply correction → efficiency maps → stats & figures          │
└─────────────────────────────────────────────────────────────────┘

 Segment both (membrane + whole-cell) ──► Send to Intensity ──┐
                                                              ▼
                                        ┌───────────────────────────┐
                                        │ 4. Intensity / Densitometry│  membrane vs
                                        │    (independent of FRET)   │  whole-cell,
                                        └───────────────────────────┘  CTCF, stats

---

Recommended end-to-end workflow

1. Segment your images in the Segmentation tab.

   • Load raw .tif/.czi images, set the Cellpose parameters, and run segmentation.
   • Refine with the ROI Manager where needed.
   • Forward control images to the bleed-through channels using Send to Donor / Send to Acceptor, and experimental images to FRET using Send to FRET or Batch Segment && Transfer.

2. Characterize bleed-through in the Bleed-Through Correction tab.

   • For each channel (S1, S2, and optionally S3/S4), run the analysis, choose a fitting model, and Confirm Fit.
   • Save Parameters — the coefficients are written to bt_params.json (and copied next to your input images).

3. Compute FRET in the FRET Analysis tab.

   • Verify the Bleed-Through Parameters panel shows your confirmed coefficients.
   • Choose formulas, set thresholds and optional filters, assign groups.
   • Run FRET Analysis and review the Results and Visualization.
   • Export maps, statistics, and figures.

---

How data moves between tabs

Data	Produced in	Consumed in	Mechanism
Segmented stacks (mask + channels)	Segmentation	Bleed-Through, FRET	Send to Donor/Acceptor, Send to FRET, Batch Segment && Transfer, or saved files
Membrane + whole-cell stacks (both_segmented_)	Segmentation (Segment both)	Intensity / Densitometry	Send to Intensity, batch transfer, or saved files
Bleed-through coefficients	Bleed-Through	FRET	Confirmed fits, shown in the FRET Bleed-Through Parameters panel; persisted in bt_params.json
Groups	FRET (or set during batch transfer)	FRET statistics	Group tags drive aggregate comparisons
Efficiency maps & stats	FRET	external tools	Exported as TIFF / CSV / figures

The label mask travels with the image as frame 0 of every segmented stack, so each tab always knows which pixels belong to which cell. Raw channel intensities are preserved through saving, so corrections and efficiencies are computed on the original data.

---

Sessions and reuse

• Save bt_params.json to reuse a bleed-through model across sessions or datasets; the tool offers to reload the last session's parameters on startup.
• Re-load saved segmented stacks at any time — they already contain the mask, so you can skip straight to bleed-through or FRET.

---

Performance tips

• A CUDA-capable GPU greatly accelerates Cellpose segmentation.
• For large images or many cells, enable Random Sampling in the Bleed-Through tab.
• Use Batch Segment && Transfer to process many images in one pass.
• Close other memory-heavy applications when working with large datasets.