Scratch Assay vs. Transwell vs. Boyden Chamber — Which Cell Migration Assay is Right for Your Experiment?
For collective cell migration and wound healing research, the scratch assay is the gold standard — especially when standardized with photochemical wound creation for reproducible gap geometry. Transwell and Boyden chamber assays measure individual cell chemotaxis and invasion, making them complementary, not interchangeable. The right choice depends on whether you study sheet migration, chemotaxis, or matrix invasion.
Three assays dominate in vitro cell migration research: the scratch/wound healing assay, the Transwell migration assay, and the Boyden chamber assay. A January 2026 Nature Methods review found these assays are “often used interchangeably” — a problem the authors attribute to a lack of practical selection guidelines. This guide fills that gap with a direct, decision-ready comparison.
Method Overview: What Each Assay Actually Measures
划痕/伤口愈合测定
- Collective sheet migration — physiologically relevant
- Real-time kinetic data possible
- Compatible with standard brightfield microscopy
- Photochemical variant: <5% wound width CV
- Measures 2D migration only
- Manual scratching introduces variability
Transwell Migration Assay
- Individual cell chemotaxis measurement
- Chemoattractant gradient well-defined
- Widely published — easy literature comparison
- Endpoint only — no kinetics
- Cell fixation required for analysis
- High inter-assay variability
Boyden Chamber Assay
- Invasion through ECM matrix measurable
- Chemotaxis + haptotaxis quantifiable
- Pore size selectable per cell type
- Most technically demanding
- No live imaging possible
- Requires staining and manual counting
Head-to-Head Comparison Table
| 标准 | Scratch Assay (Photochemical) | Transwell | Boyden Chamber |
|---|---|---|---|
| Migration type | Collective sheet migration | Individual chemotaxis | Individual chemotaxis + invasion |
| 生理学相关性 | High — models wound healing in vivo | 中等 | 中等 |
| Live imaging | Yes — continuous time-lapse | No — endpoint only | No — endpoint only |
| Reproducibility (CV) | <5% (photochemical) | 15–30% | 20–35% |
| 96-well compatible | Yes — full plate automation | Yes — but manual counting | Yes — limited throughput |
| Drug screening | Excellent — kinetic IC50 curves | Good — single endpoint | Good — invasion focus |
| ECM coating | Flexible — Fibronectin, Collagen, Laminin | Membrane coating | Matrigel or ECM matrix |
| Cell number required | Low — confluent monolayer | 中等 | High — limited by pore density |
| Analysis method | AI automated gap closure | Manual staining + counting | Manual staining + counting |
| Cost per experiment | From €59/plate | €80–150/plate | €100–200/plate |
When to Use Each Method
Decision Guide — Choose Your Assay
Match your biological question to the right method
The Reproducibility Problem — Why Scratch Assays Get a Bad Reputation
The scratch assay’s main weakness has historically been reproducibility: a pipette tip dragged across a confluent monolayer produces irregular wound widths between wells and operators, with coefficient of variation (CV) regularly exceeding 30%. This makes detecting small drug effects statistically impossible.
Photochemical wound creation solves this entirely. A precision light mask defines the exact wound geometry before light exposure — same width, same position, same area in every well, regardless of operator. CV drops below 5%, bringing scratch assay reproducibility well below that of Transwell assays.
For wound healing research and drug screening: standardized scratch assay wins on every metric that matters.
Kinetic data, physiological relevance, scalability to 96-well, automated analysis, and lower cost per data point all favor the photochemical scratch assay over Transwell and Boyden alternatives — provided wound creation is standardized. The ScratchMaker plate system eliminates the one weakness that gave scratch assays a reproducibility problem.
Try the standardized scratch assay
ScratchMaker Plates — from €59. Compatible with any fluorescence microscope.
